doi
stringlengths 9
10
| chunk-id
stringlengths 1
4
| chunk
stringlengths 1
1.57k
| id
stringlengths 9
10
| title
stringlengths 10
127
| summary
stringlengths 581
1.92k
| source
stringlengths 30
31
| authors
sequence | categories
sequence | comment
stringclasses 28
values | journal_ref
stringclasses 1
value | primary_category
stringclasses 6
values | published
stringlengths 8
8
| updated
stringlengths 8
8
| references
list |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1806.01261 | 168 | CxGaZigWg89NzV+TpXhTOC8Nso0ppRN6RiHlkoao/bzYs85ubBKSKJE2ScNKdTfEzmNtZ7FgU3G1Ez0qrcQ//OGmYlu/JzLNDMo2fKjKBPEJGRxNAm5QmbEzBLKFLe7EjahijJjq6nZErzVk9dJ77LluS3v/qrRbpZ1VOEMzqEJHlxDG+6gA11gEMIzvMKbw50X5935WEYrTjlzCn/gfP4AA6WMYA==</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 169 | sha1_base64="6WjtAQy1eEki3DeLmUkkI9Sk/Os=">AAAB53icbVDLSsNAFL2pr1pfVZduBovQVUlE0GXBjcsK9gFtkMnkph06mYSZiVBCf8CNC0Xc+kvu/BunaRbaemDgcM65zL0nSAXXxnW/ncrG5tb2TnW3trd/cHhUPz7p6SRTDLssEYkaBFSj4BK7hhuBg1QhjQOB/WB6u/D7T6g0T+SDmaXox3QsecQZNVbqPNYbbsstQNaJV5IGlLD5r1GYsCxGaZigWg89NzV+TpXhTOC8Nso0ppRN6RiHlkoao | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 170 | CxGaZigWg89NzV+TpXhTOC8Nso0ppRN6RiHlkoao/bzYs85ubBKSKJE2ScNKdTfEzmNtZ7FgU3G1Ez0qrcQ//OGmYlu/JzLNDMo2fKjKBPEJGRxNAm5QmbEzBLKFLe7EjahijJjq6nZErzVk9dJ77LluS3v/qrRbpZ1VOEMzqEJHlxDG+6gA11gEMIzvMKbw50X5935WEYrTjlzCn/gfP4AA6WMYA==</latexit> | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 171 | (c) Global update | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 172 | Figure 3: Updates in a GN block. Blue indicates the element that is being updated, and black
indicates other elements which are involved in the update (note that the pre-update value of the
blue element is also used in the update). See Equation 1 for details on the notation.
3. vis applied to each node i, to compute an updated node attribute, v0
i. In our running
example, vmay compute something analogous to the updated position, velocity, and kinetic
energy of each ball. The set of resulting per-node outputs is, V0=fv0
igi=1:Nv.
4.e!uis applied to E0, and aggregates all edge updates, into e0, which will then be used in the
next step's global update. In our running example, e!umay compute the summed forces
(which should be zero, in this case, due to Newton's third law) and the springs' potential
energies.
5.v!uis applied to V0, and aggregates all node updates, into v0, which will then be used in
the next step's global update. In our running example, v!umight compute the total kinetic
energy of the system. | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 173 | the next step's global update. In our running example, v!umight compute the total kinetic
energy of the system.
6. uis applied once per graph, and computes an update for the global attribute, u0. In our
running example, umight compute something analogous to the net forces and total energy
of the physical system.
Note, though we assume this sequence of steps here, the order is not strictly enforced: it is possible
to reverse the update functions to proceed from global, to per-node, to per-edge updates, for example.
Kearnes et al. (2016) computes edge updates from nodes in a similar manner.
3.2.4 Relational inductive biases in graph networks
Our GN framework imposes several strong relational inductive biases when used as components in
a learning process. First, graphs can express arbitrary relationships among entities, which means
the GN's input determines how representations interact and are isolated, rather than those choices
being determined by the xed architecture. For example, the assumption that two entities have a
relationship|and thus should interact|is expressed by an edge between the entities' corresponding
nodes. Similarly, the absence of an edge expresses the assumption that the nodes have no relationship
and should not in
uence each other directly. | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 174 | nodes. Similarly, the absence of an edge expresses the assumption that the nodes have no relationship
and should not in
uence each other directly.
Second, graphs represent entities and their relations as sets, which are invariant to permutations.
This means GNs are invariant to the order of these elements6, which is often desirable. For example,
the objects in a scene do not have a natural ordering (see Sec. 2.2).
Third, a GN's per-edge and per-node functions are reused across all edges and nodes, respectively.
This means GNs automatically support a form of combinatorial generalization (see Section 5.1):
because graphs are composed of edges, nodes, and global features, a single GN can operate on
graphs of di erent sizes (numbers of edges and nodes) and shapes (edge connectivity).
6Note, an ordering can be imposed by encoding the indices in the node or edge attributes, or via the edges
themselves (e.g. by encoding a chain or partial ordering).
13
4 Design principles for graph network architectures
The GN framework can be used to implement a wide variety of architectures, in accordance with
the design principles listed above in Section 3.2, which also correspond to the sub-sections (4.1, | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 175 | the design principles listed above in Section 3.2, which also correspond to the sub-sections (4.1,
4.2, and 4.3) below. In general, the framework is agnostic to speci c attribute representations and
functional forms. Here, however, we focus mainly on deep learning architectures, which allow GNs
to act as learnable graph-to-graph function approximators.
4.1 Flexible representations
Graph networks support highly
exible graph representations in two ways: rst, in terms of the
representation of the attributes; and second, in terms of the structure of the graph itself.
4.1.1 Attributes
The global, node, and edge attributes of a GN block can use arbitrary representational formats. In
deep learning implementations, real-valued vectors and tensors are most common. However, other
data structures such as sequences, sets, or even graphs could also be used.
The requirements of the problem will often determine what representations should be used for
the attributes. For example, when the input data is an image, the attributes might be represented
as tensors of image patches; however, when the input data is a text document, the attributes might
be sequences of words corresponding to sentences.
For each GN block within a broader architecture, the edge and node outputs typically correspond | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 176 | be sequences of words corresponding to sentences.
For each GN block within a broader architecture, the edge and node outputs typically correspond
to lists of vectors or tensors, one per edge or node, and the global outputs correspond to a single
vector or tensor. This allows a GN's output to be passed to other deep learning building blocks
such as MLPs, CNNs, and RNNs.
The output of a GN block can also be tailored to the demands of the task. In particular,
Anedge-focused GN uses the edges as output, for example to make decisions about interactions
among entities (Kipf et al., 2018; Hamrick et al., 2018).
Anode-focused GN uses the nodes as output, for example to reason about physical systems
(Battaglia et al., 2016; Chang et al., 2017; Wang et al., 2018b; Sanchez-Gonzalez et al., 2018).
Agraph-focused GN uses the globals as output, for example to predict the potential energy of
a physical system (Battaglia et al., 2016), the properties of a molecule (Gilmer et al., 2017),
or answers to questions about a visual scene (Santoro et al., 2017).
The nodes, edges, and global outputs can also be mixed-and-matched depending on the task. For | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 177 | or answers to questions about a visual scene (Santoro et al., 2017).
The nodes, edges, and global outputs can also be mixed-and-matched depending on the task. For
example, Hamrick et al. (2018) used both the output edge and global attributes to compute a policy
over actions.
4.1.2 Graph structure
When de ning how the input data will be represented as a graph, there are generally two scenarios:
rst, the input explicitly speci es the relational structure; and second, the relational structure must
be inferred or assumed. These are not hard distinctions, but extremes along a continuum.
Examples of data with more explicitly speci ed entities and relations include knowledge graphs,
social networks, parse trees, optimization problems, chemical graphs, road networks, and physical
systems with known interactions. Figures 2a-d illustrate how such data can be expressed as graphs.
Examples of data where the relational structure is not made explicit, and must be inferred or
assumed, include visual scenes, text corpora, programming language source code, and multi-agent
14
systems. In these types of settings, the data may be formatted as a set of entities without relations,
or even just a vector or tensor (e.g., an image). If the entities are not speci ed explicitly, they might | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 178 | or even just a vector or tensor (e.g., an image). If the entities are not speci ed explicitly, they might
be assumed, for instance, by treating each word in a sentence (Vaswani et al., 2017) or each local
feature vector in a CNN's output feature map, as a node (Watters et al., 2017; Santoro et al., 2017;
Wang et al., 2018c) (Figures 2e-f). Or, it might be possible to use a separate learned mechanism to
infer entities from an unstructured signal (Luong et al., 2015; Mnih et al., 2014; Eslami et al., 2016;
van Steenkiste et al., 2018). If relations are not available, the simplest approach is to instantiate all
possible directed edges between entities (Figure 2f). This can be prohibitive for large numbers of
entities, however, because the number of possible edges grows quadratically with the number of
nodes. Thus developing more sophisticated ways of inferring sparse structure from unstructured
data (Kipf et al., 2018) is an important future direction.
4.2 Con gurable within-block structure
The structure and functions within a GN block can be con gured in di erent ways, which o ers | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 179 | 4.2 Con gurable within-block structure
The structure and functions within a GN block can be con gured in di erent ways, which o ers
exibility in what information is made available as inputs to its functions, as well as how output edge,
node, and global updates are produced. In particular, each in Equation 1 must be implemented
with some function, f, wheref's argument signature determines what information it requires as
input; in Figure 4, the incoming arrows to each depict whether u,V, andEare taken as inputs.
Hamrick et al. (2018) and Sanchez-Gonzalez et al. (2018) used the full GN block shown in Figure 4a.
Their implementations used neural networks (denoted NNe,NNv, and NNubelow, to indicate that
they are di erent functions with di erent parameters). Their implementations used elementwise
summation, but averages and max/min could also be used,
e(ek;vrk;vsk;u):=fe(ek;vrk;vsk;u) = NNe([ek;vrk;vsk;u]) (2)
v
e0
i;vi;u:=fv
e0
i;vi;u | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 180 | v
e0
i;vi;u:=fv
e0
i;vi;u
= NNv
[ e0
i;vi;u]
u
e0; v0;u:=fu
e0; v0;u
= NNu
[ e0; v0;u]
e!v
E0
i:= =X
fk:rk=ige0
k
v!u
V0:= =X
iv0
i
e!u
E0:= =X
ke0
k
where [ x;y;z] indicates vector/tensor concatenation. For vector attributes, a MLP is often used for
, while for tensors such as image feature maps, CNNs may be more suitable.
The functions can also use RNNs, which requires an additional hidden state as input and
output. Figure 4b shows a very simple version of a GN block with RNNs as functions: there is no
message-passing in this formulation, and this type of block might be used for recurrent smoothing of | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 181 | message-passing in this formulation, and this type of block might be used for recurrent smoothing of
some dynamic graph states. Of course, RNNs as functions could also be used in a full GN block
(Figure 4a).
A variety of other architectures can be expressed in the GN framework, often as di erent
function choices and within-block con gurations. The remaining sub-sections explore how a GN's
within-block structure can be con gured in di erent ways, with examples of published work which
uses such con gurations. See the Appendix for details.
15
(a) Full GN block
u0,u0hid | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 182 | <latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 183 | sha1_base64="W2hu4ghYojV0SmOcBbcdBDITOSo=">AAACDnicbVC7TsMwFHV4lvIKMLJYVBUMqEoQEgwMlVgYi0QfUhtFjuu0Vm0nsh2kKsoXsPArLAwgxMrMxt/gpBlKy5EsHZ9zr+69J4gZVdpxfqyV1bX1jc3KVnV7Z3dv3z447KgokZi0ccQi2QuQIowK0tZUM9KLJUE8YKQbTG5zv/tIpKKReNDTmHgcjQQNKUbaSL5dH3Ckx0GYJtnpOZz7+AWXPB3TYebbNafhFIDLxC1JDZRo+fb3YBjhhBOhM | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 184 | AWXPB3TYebbNafhFIDLxC1JDZRo+fb3YBjhhBOhMUNK9V0n1l6KpKaYkaw6SBSJEZ6gEekbKhAnykuLczJYN8oQhpE0T2hYqPMdKeJKTXlgKvMd1aKXi/95/USH115KRZxoIvBsUJgwqCOYZwOHVBKs2dQQhCU1u0I8RhJhbRKsmhDcxZOXSeei4ToN9/6y1rwp46iAY3ACzoALrkAT3IEWaAMMnsALeAPv1rP1an1Yn7PSFavsOQJ/YH39AtnQnJg=</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 185 | sha1_base64="W2hu4ghYojV0SmOcBbcdBDITOSo=">AAACDnicbVC7TsMwFHV4lvIKMLJYVBUMqEoQEgwMlVgYi0QfUhtFjuu0Vm0nsh2kKsoXsPArLAwgxMrMxt/gpBlKy5EsHZ9zr+69J4gZVdpxfqyV1bX1jc3KVnV7Z3dv3z447KgokZi0ccQi2QuQIowK0tZUM9KLJUE8YKQbTG5zv/tIpKKReNDTmHgcjQQNKUbaSL5dH3Ckx0GYJtnpOZz7+AWXPB3TYebbNafhFIDLxC1JDZRo+fb3YBjhhBOhM | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 186 | AWXPB3TYebbNafhFIDLxC1JDZRo+fb3YBjhhBOhMUNK9V0n1l6KpKaYkaw6SBSJEZ6gEekbKhAnykuLczJYN8oQhpE0T2hYqPMdKeJKTXlgKvMd1aKXi/95/USH115KRZxoIvBsUJgwqCOYZwOHVBKs2dQQhCU1u0I8RhJhbRKsmhDcxZOXSeei4ToN9/6y1rwp46iAY3ACzoALrkAT3IEWaAMMnsALeAPv1rP1an1Yn7PSFavsOQJ/YH39AtnQnJg=</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 187 | sha1_base64="W2hu4ghYojV0SmOcBbcdBDITOSo=">AAACDnicbVC7TsMwFHV4lvIKMLJYVBUMqEoQEgwMlVgYi0QfUhtFjuu0Vm0nsh2kKsoXsPArLAwgxMrMxt/gpBlKy5EsHZ9zr+69J4gZVdpxfqyV1bX1jc3KVnV7Z3dv3z447KgokZi0ccQi2QuQIowK0tZUM9KLJUE8YKQbTG5zv/tIpKKReNDTmHgcjQQNKUbaSL5dH3Ckx0GYJtnpOZz7+AWXPB3TYebbNafhFIDLxC1JDZRo+fb3YBjhhBOhM | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 188 | AWXPB3TYebbNafhFIDLxC1JDZRo+fb3YBjhhBOhMUNK9V0n1l6KpKaYkaw6SBSJEZ6gEekbKhAnykuLczJYN8oQhpE0T2hYqPMdKeJKTXlgKvMd1aKXi/95/USH115KRZxoIvBsUJgwqCOYZwOHVBKs2dQQhCU1u0I8RhJhbRKsmhDcxZOXSeei4ToN9/6y1rwp46iAY3ACzoALrkAT3IEWaAMMnsALeAPv1rP1an1Yn7PSFavsOQJ/YH39AtnQnJg=</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 189 | sha1_base64="W2hu4ghYojV0SmOcBbcdBDITOSo=">AAACDnicbVC7TsMwFHV4lvIKMLJYVBUMqEoQEgwMlVgYi0QfUhtFjuu0Vm0nsh2kKsoXsPArLAwgxMrMxt/gpBlKy5EsHZ9zr+69J4gZVdpxfqyV1bX1jc3KVnV7Z3dv3z447KgokZi0ccQi2QuQIowK0tZUM9KLJUE8YKQbTG5zv/tIpKKReNDTmHgcjQQNKUbaSL5dH3Ckx0GYJtnpOZz7+AWXPB3TYebbNafhFIDLxC1JDZRo+fb3YBjhhBOhM | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 190 | AWXPB3TYebbNafhFIDLxC1JDZRo+fb3YBjhhBOhMUNK9V0n1l6KpKaYkaw6SBSJEZ6gEekbKhAnykuLczJYN8oQhpE0T2hYqPMdKeJKTXlgKvMd1aKXi/95/USH115KRZxoIvBsUJgwqCOYZwOHVBKs2dQQhCU1u0I8RhJhbRKsmhDcxZOXSeei4ToN9/6y1rwp46iAY3ACzoALrkAT3IEWaAMMnsALeAPv1rP1an1Yn7PSFavsOQJ/YH39AtnQnJg=</latexit>E0,E0hid | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 191 | <latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 192 | sha1_base64="ZYHbDTDbKJNm66GNrt+dUMK6558=">AAAB/HicbVDLSgMxFM3UV62v0S7dBIvUhZQZEXThoiAFlxXsA9phyGQybWiSGZKMUIb6K25cKOLWD3Hn35hpZ6GtBwKHc+7lnpwgYVRpx/m2SmvrG5tb5e3Kzu7e/oF9eNRVcSox6eCYxbIfIEUYFaSjqWakn0iCeMBIL5jc5n7vkUhFY/GgpwnxOBoJGlGMtJF8u9qqn8NW3R9ypMeSZ2Mazny75jScOeAqcQtSAwXavv01DGOcciI0Zkipgesk2 | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 193 | ny75jScOeAqcQtSAwXavv01DGOcciI0Zkipgesk2suQ1BQzMqsMU0UShCdoRAaGCsSJ8rJ5+Bk8NUoIo1iaJzScq783MsSVmvLATOYZ1bKXi/95g1RH115GRZJqIvDiUJQyqGOYNwFDKgnWbGoIwpKarBCPkURYm74qpgR3+curpHvRcJ2Ge39Za94UdZTBMTgBZ8AFV6AJ7kAbdAAGU/AMXsGb9WS9WO/Wx2K0ZBU7VfAH1ucPbBGT+A==</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 194 | sha1_base64="ZYHbDTDbKJNm66GNrt+dUMK6558=">AAAB/HicbVDLSgMxFM3UV62v0S7dBIvUhZQZEXThoiAFlxXsA9phyGQybWiSGZKMUIb6K25cKOLWD3Hn35hpZ6GtBwKHc+7lnpwgYVRpx/m2SmvrG5tb5e3Kzu7e/oF9eNRVcSox6eCYxbIfIEUYFaSjqWakn0iCeMBIL5jc5n7vkUhFY/GgpwnxOBoJGlGMtJF8u9qqn8NW3R9ypMeSZ2Mazny75jScOeAqcQtSAwXavv01DGOcciI0Zkipgesk2 | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 195 | ny75jScOeAqcQtSAwXavv01DGOcciI0Zkipgesk2suQ1BQzMqsMU0UShCdoRAaGCsSJ8rJ5+Bk8NUoIo1iaJzScq783MsSVmvLATOYZ1bKXi/95g1RH115GRZJqIvDiUJQyqGOYNwFDKgnWbGoIwpKarBCPkURYm74qpgR3+curpHvRcJ2Ge39Za94UdZTBMTgBZ8AFV6AJ7kAbdAAGU/AMXsGb9WS9WO/Wx2K0ZBU7VfAH1ucPbBGT+A==</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 196 | sha1_base64="ZYHbDTDbKJNm66GNrt+dUMK6558=">AAAB/HicbVDLSgMxFM3UV62v0S7dBIvUhZQZEXThoiAFlxXsA9phyGQybWiSGZKMUIb6K25cKOLWD3Hn35hpZ6GtBwKHc+7lnpwgYVRpx/m2SmvrG5tb5e3Kzu7e/oF9eNRVcSox6eCYxbIfIEUYFaSjqWakn0iCeMBIL5jc5n7vkUhFY/GgpwnxOBoJGlGMtJF8u9qqn8NW3R9ypMeSZ2Mazny75jScOeAqcQtSAwXavv01DGOcciI0Zkipgesk2 | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 197 | ny75jScOeAqcQtSAwXavv01DGOcciI0Zkipgesk2suQ1BQzMqsMU0UShCdoRAaGCsSJ8rJ5+Bk8NUoIo1iaJzScq783MsSVmvLATOYZ1bKXi/95g1RH115GRZJqIvDiUJQyqGOYNwFDKgnWbGoIwpKarBCPkURYm74qpgR3+curpHvRcJ2Ge39Za94UdZTBMTgBZ8AFV6AJ7kAbdAAGU/AMXsGb9WS9WO/Wx2K0ZBU7VfAH1ucPbBGT+A==</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 198 | sha1_base64="ZYHbDTDbKJNm66GNrt+dUMK6558=">AAAB/HicbVDLSgMxFM3UV62v0S7dBIvUhZQZEXThoiAFlxXsA9phyGQybWiSGZKMUIb6K25cKOLWD3Hn35hpZ6GtBwKHc+7lnpwgYVRpx/m2SmvrG5tb5e3Kzu7e/oF9eNRVcSox6eCYxbIfIEUYFaSjqWakn0iCeMBIL5jc5n7vkUhFY/GgpwnxOBoJGlGMtJF8u9qqn8NW3R9ypMeSZ2Mazny75jScOeAqcQtSAwXavv01DGOcciI0Zkipgesk2 | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 199 | ny75jScOeAqcQtSAwXavv01DGOcciI0Zkipgesk2suQ1BQzMqsMU0UShCdoRAaGCsSJ8rJ5+Bk8NUoIo1iaJzScq783MsSVmvLATOYZ1bKXi/95g1RH115GRZJqIvDiUJQyqGOYNwFDKgnWbGoIwpKarBCPkURYm74qpgR3+curpHvRcJ2Ge39Za94UdZTBMTgBZ8AFV6AJ7kAbdAAGU/AMXsGb9WS9WO/Wx2K0ZBU7VfAH1ucPbBGT+A==</latexit>V0,V0hid | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 200 | <latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 201 | sha1_base64="WhBBytL7AOLCOPFKqkdbfTZ2xq4=">AAAB/HicbVBPS8MwHE3nvzn/VXf0EhwyDzJaEfTgYeDF4wTXDbZS0jTdwpK0JKkwyvwqXjwo4tUP4s1vY7r1oJsPAo/3fj9+Ly9MGVXacb6tytr6xuZWdbu2s7u3f2AfHnkqySQmXZywRPZDpAijgnQ11Yz0U0kQDxnphZPbwu89EqloIh70NCU+RyNBY4qRNlJg173mOfSawZAjPZY8H9NoFtgNp+XMAVeJW5IGKNEJ7K9hlOCME6ExQ0oNXCfVf | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 202 | tgNp+XMAVeJW5IGKNEJ7K9hlOCME6ExQ0oNXCfVfo6kppiRWW2YKZIiPEEjMjBUIE6Un8/Dz+CpUSIYJ9I8oeFc/b2RI67UlIdmssiolr1C/M8bZDq+9nMq0kwTgReH4oxBncCiCRhRSbBmU0MQltRkhXiMJMLa9FUzJbjLX14l3kXLdVru/WWjfVPWUQXH4AScARdcgTa4Ax3QBRhMwTN4BW/Wk/VivVsfi9GKVe7UwR9Ynz+huZQa</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 203 | sha1_base64="WhBBytL7AOLCOPFKqkdbfTZ2xq4=">AAAB/HicbVBPS8MwHE3nvzn/VXf0EhwyDzJaEfTgYeDF4wTXDbZS0jTdwpK0JKkwyvwqXjwo4tUP4s1vY7r1oJsPAo/3fj9+Ly9MGVXacb6tytr6xuZWdbu2s7u3f2AfHnkqySQmXZywRPZDpAijgnQ11Yz0U0kQDxnphZPbwu89EqloIh70NCU+RyNBY4qRNlJg173mOfSawZAjPZY8H9NoFtgNp+XMAVeJW5IGKNEJ7K9hlOCME6ExQ0oNXCfVf | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 204 | tgNp+XMAVeJW5IGKNEJ7K9hlOCME6ExQ0oNXCfVfo6kppiRWW2YKZIiPEEjMjBUIE6Un8/Dz+CpUSIYJ9I8oeFc/b2RI67UlIdmssiolr1C/M8bZDq+9nMq0kwTgReH4oxBncCiCRhRSbBmU0MQltRkhXiMJMLa9FUzJbjLX14l3kXLdVru/WWjfVPWUQXH4AScARdcgTa4Ax3QBRhMwTN4BW/Wk/VivVsfi9GKVe7UwR9Ynz+huZQa</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 205 | sha1_base64="WhBBytL7AOLCOPFKqkdbfTZ2xq4=">AAAB/HicbVBPS8MwHE3nvzn/VXf0EhwyDzJaEfTgYeDF4wTXDbZS0jTdwpK0JKkwyvwqXjwo4tUP4s1vY7r1oJsPAo/3fj9+Ly9MGVXacb6tytr6xuZWdbu2s7u3f2AfHnkqySQmXZywRPZDpAijgnQ11Yz0U0kQDxnphZPbwu89EqloIh70NCU+RyNBY4qRNlJg173mOfSawZAjPZY8H9NoFtgNp+XMAVeJW5IGKNEJ7K9hlOCME6ExQ0oNXCfVf | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 206 | tgNp+XMAVeJW5IGKNEJ7K9hlOCME6ExQ0oNXCfVfo6kppiRWW2YKZIiPEEjMjBUIE6Un8/Dz+CpUSIYJ9I8oeFc/b2RI67UlIdmssiolr1C/M8bZDq+9nMq0kwTgReH4oxBncCiCRhRSbBmU0MQltRkhXiMJMLa9FUzJbjLX14l3kXLdVru/WWjfVPWUQXH4AScARdcgTa4Ax3QBRhMwTN4BW/Wk/VivVsfi9GKVe7UwR9Ynz+huZQa</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 207 | sha1_base64="WhBBytL7AOLCOPFKqkdbfTZ2xq4=">AAAB/HicbVBPS8MwHE3nvzn/VXf0EhwyDzJaEfTgYeDF4wTXDbZS0jTdwpK0JKkwyvwqXjwo4tUP4s1vY7r1oJsPAo/3fj9+Ly9MGVXacb6tytr6xuZWdbu2s7u3f2AfHnkqySQmXZywRPZDpAijgnQ11Yz0U0kQDxnphZPbwu89EqloIh70NCU+RyNBY4qRNlJg173mOfSawZAjPZY8H9NoFtgNp+XMAVeJW5IGKNEJ7K9hlOCME6ExQ0oNXCfVf | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 208 | tgNp+XMAVeJW5IGKNEJ7K9hlOCME6ExQ0oNXCfVfo6kppiRWW2YKZIiPEEjMjBUIE6Un8/Dz+CpUSIYJ9I8oeFc/b2RI67UlIdmssiolr1C/M8bZDq+9nMq0kwTgReH4oxBncCiCRhRSbBmU0MQltRkhXiMJMLa9FUzJbjLX14l3kXLdVru/WWjfVPWUQXH4AScARdcgTa4Ax3QBRhMwTN4BW/Wk/VivVsfi9GKVe7UwR9Ynz+huZQa</latexit>E,Ehid | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 209 | <latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 210 | sha1_base64="DYZek5SmevKS8py25dx0aIqUbBY=">AAAB+nicbVDLSsNAFL3xWesr1aWbwSK4kJKIoAsXBSm4rGAf0IYwmUzboTNJmJkoJfZT3LhQxK1f4s6/cdJmoa0HBg7n3Ms9c4KEM6Ud59taWV1b39gsbZW3d3b39u3KQVvFqSS0RWIey26AFeUsoi3NNKfdRFIsAk47wfgm9zsPVCoWR/d6klBP4GHEBoxgbSTfrjTOUMPvC6xHUmQjFk59u+rUnBnQMnELUoUCTd/+6ocxSQWNNOFYqZ7rJNrLs | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 211 | +rUnBnQMnELUoUCTd/+6ocxSQWNNOFYqZ7rJNrLsNSMcDot91NFE0zGeEh7hkZYUOVls+hTdGKUEA1iaV6k0Uz9vZFhodREBGYyz6gWvVz8z+ulenDlZSxKUk0jMj80SDnSMcp7QCGTlGg+MQQTyUxWREZYYqJNW2VTgrv45WXSPq+5Ts29u6jWr4s6SnAEx3AKLlxCHW6hCS0g8AjP8Apv1pP1Yr1bH/PRFavYOYQ/sD5/AKXVk5Y=</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 212 | sha1_base64="DYZek5SmevKS8py25dx0aIqUbBY=">AAAB+nicbVDLSsNAFL3xWesr1aWbwSK4kJKIoAsXBSm4rGAf0IYwmUzboTNJmJkoJfZT3LhQxK1f4s6/cdJmoa0HBg7n3Ms9c4KEM6Ud59taWV1b39gsbZW3d3b39u3KQVvFqSS0RWIey26AFeUsoi3NNKfdRFIsAk47wfgm9zsPVCoWR/d6klBP4GHEBoxgbSTfrjTOUMPvC6xHUmQjFk59u+rUnBnQMnELUoUCTd/+6ocxSQWNNOFYqZ7rJNrLs | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 213 | +rUnBnQMnELUoUCTd/+6ocxSQWNNOFYqZ7rJNrLsNSMcDot91NFE0zGeEh7hkZYUOVls+hTdGKUEA1iaV6k0Uz9vZFhodREBGYyz6gWvVz8z+ulenDlZSxKUk0jMj80SDnSMcp7QCGTlGg+MQQTyUxWREZYYqJNW2VTgrv45WXSPq+5Ts29u6jWr4s6SnAEx3AKLlxCHW6hCS0g8AjP8Apv1pP1Yr1bH/PRFavYOYQ/sD5/AKXVk5Y=</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 214 | sha1_base64="DYZek5SmevKS8py25dx0aIqUbBY=">AAAB+nicbVDLSsNAFL3xWesr1aWbwSK4kJKIoAsXBSm4rGAf0IYwmUzboTNJmJkoJfZT3LhQxK1f4s6/cdJmoa0HBg7n3Ms9c4KEM6Ud59taWV1b39gsbZW3d3b39u3KQVvFqSS0RWIey26AFeUsoi3NNKfdRFIsAk47wfgm9zsPVCoWR/d6klBP4GHEBoxgbSTfrjTOUMPvC6xHUmQjFk59u+rUnBnQMnELUoUCTd/+6ocxSQWNNOFYqZ7rJNrLs | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 215 | +rUnBnQMnELUoUCTd/+6ocxSQWNNOFYqZ7rJNrLsNSMcDot91NFE0zGeEh7hkZYUOVls+hTdGKUEA1iaV6k0Uz9vZFhodREBGYyz6gWvVz8z+ulenDlZSxKUk0jMj80SDnSMcp7QCGTlGg+MQQTyUxWREZYYqJNW2VTgrv45WXSPq+5Ts29u6jWr4s6SnAEx3AKLlxCHW6hCS0g8AjP8Apv1pP1Yr1bH/PRFavYOYQ/sD5/AKXVk5Y=</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 216 | sha1_base64="DYZek5SmevKS8py25dx0aIqUbBY=">AAAB+nicbVDLSsNAFL3xWesr1aWbwSK4kJKIoAsXBSm4rGAf0IYwmUzboTNJmJkoJfZT3LhQxK1f4s6/cdJmoa0HBg7n3Ms9c4KEM6Ud59taWV1b39gsbZW3d3b39u3KQVvFqSS0RWIey26AFeUsoi3NNKfdRFIsAk47wfgm9zsPVCoWR/d6klBP4GHEBoxgbSTfrjTOUMPvC6xHUmQjFk59u+rUnBnQMnELUoUCTd/+6ocxSQWNNOFYqZ7rJNrLs | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 217 | +rUnBnQMnELUoUCTd/+6ocxSQWNNOFYqZ7rJNrLsNSMcDot91NFE0zGeEh7hkZYUOVls+hTdGKUEA1iaV6k0Uz9vZFhodREBGYyz6gWvVz8z+ulenDlZSxKUk0jMj80SDnSMcp7QCGTlGg+MQQTyUxWREZYYqJNW2VTgrv45WXSPq+5Ts29u6jWr4s6SnAEx3AKLlxCHW6hCS0g8AjP8Apv1pP1Yr1bH/PRFavYOYQ/sD5/AKXVk5Y=</latexit>V, | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 218 | Vhid | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 219 | <latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 220 | sha1_base64="brwuxF74R6OEOykh378as/RBfzE=">AAAB+nicbVBNS8NAFHzxs9avVI9eFovgQUoigh48FLx4rGDTQhvCZrNtl+4mYXejlNif4sWDIl79Jd78N27aHLR1YGGYeY83O2HKmdKO822trK6tb2xWtqrbO7t7+3btwFNJJgltk4QnshtiRTmLaVszzWk3lRSLkNNOOL4p/M4DlYol8b2epNQXeBizASNYGymwa94Z8oK+wHokRT5i0TSw607DmQEtE7ckdSjRCuyvfpSQTNBYE46V6rlOqv0cS | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 221 | 07DmQEtE7ckdSjRCuyvfpSQTNBYE46V6rlOqv0cS80Ip9NqP1M0xWSMh7RnaIwFVX4+iz5FJ0aJ0CCR5sUazdTfGzkWSk1EaCaLjGrRK8T/vF6mB1d+zuI00zQm80ODjCOdoKIHFDFJieYTQzCRzGRFZIQlJtq0VTUluItfXibeecN1Gu7dRb15XdZRgSM4hlNw4RKacAstaAOBR3iGV3iznqwX6936mI+uWOXOIfyB9fkD20qTuA==</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 222 | sha1_base64="brwuxF74R6OEOykh378as/RBfzE=">AAAB+nicbVBNS8NAFHzxs9avVI9eFovgQUoigh48FLx4rGDTQhvCZrNtl+4mYXejlNif4sWDIl79Jd78N27aHLR1YGGYeY83O2HKmdKO822trK6tb2xWtqrbO7t7+3btwFNJJgltk4QnshtiRTmLaVszzWk3lRSLkNNOOL4p/M4DlYol8b2epNQXeBizASNYGymwa94Z8oK+wHokRT5i0TSw607DmQEtE7ckdSjRCuyvfpSQTNBYE46V6rlOqv0cS | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 223 | 07DmQEtE7ckdSjRCuyvfpSQTNBYE46V6rlOqv0cS80Ip9NqP1M0xWSMh7RnaIwFVX4+iz5FJ0aJ0CCR5sUazdTfGzkWSk1EaCaLjGrRK8T/vF6mB1d+zuI00zQm80ODjCOdoKIHFDFJieYTQzCRzGRFZIQlJtq0VTUluItfXibeecN1Gu7dRb15XdZRgSM4hlNw4RKacAstaAOBR3iGV3iznqwX6936mI+uWOXOIfyB9fkD20qTuA==</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 224 | sha1_base64="brwuxF74R6OEOykh378as/RBfzE=">AAAB+nicbVBNS8NAFHzxs9avVI9eFovgQUoigh48FLx4rGDTQhvCZrNtl+4mYXejlNif4sWDIl79Jd78N27aHLR1YGGYeY83O2HKmdKO822trK6tb2xWtqrbO7t7+3btwFNJJgltk4QnshtiRTmLaVszzWk3lRSLkNNOOL4p/M4DlYol8b2epNQXeBizASNYGymwa94Z8oK+wHokRT5i0TSw607DmQEtE7ckdSjRCuyvfpSQTNBYE46V6rlOqv0cS | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 225 | 07DmQEtE7ckdSjRCuyvfpSQTNBYE46V6rlOqv0cS80Ip9NqP1M0xWSMh7RnaIwFVX4+iz5FJ0aJ0CCR5sUazdTfGzkWSk1EaCaLjGrRK8T/vF6mB1d+zuI00zQm80ODjCOdoKIHFDFJieYTQzCRzGRFZIQlJtq0VTUluItfXibeecN1Gu7dRb15XdZRgSM4hlNw4RKacAstaAOBR3iGV3iznqwX6936mI+uWOXOIfyB9fkD20qTuA==</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 226 | sha1_base64="brwuxF74R6OEOykh378as/RBfzE=">AAAB+nicbVBNS8NAFHzxs9avVI9eFovgQUoigh48FLx4rGDTQhvCZrNtl+4mYXejlNif4sWDIl79Jd78N27aHLR1YGGYeY83O2HKmdKO822trK6tb2xWtqrbO7t7+3btwFNJJgltk4QnshtiRTmLaVszzWk3lRSLkNNOOL4p/M4DlYol8b2epNQXeBizASNYGymwa94Z8oK+wHokRT5i0TSw607DmQEtE7ckdSjRCuyvfpSQTNBYE46V6rlOqv0cS | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 227 | 07DmQEtE7ckdSjRCuyvfpSQTNBYE46V6rlOqv0cS80Ip9NqP1M0xWSMh7RnaIwFVX4+iz5FJ0aJ0CCR5sUazdTfGzkWSk1EaCaLjGrRK8T/vF6mB1d+zuI00zQm80ODjCOdoKIHFDFJieYTQzCRzGRFZIQlJtq0VTUluItfXibeecN1Gu7dRb15XdZRgSM4hlNw4RKacAstaAOBR3iGV3iznqwX6936mI+uWOXOIfyB9fkD20qTuA==</latexit>u,uhid | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 228 | <latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 229 | sha1_base64="UO6spgXZarocGO1zBKohmEwCj+c=">AAACDHicbZDNSsNAFIVv6l+tf1WXbgaL4EJKIoIuXBTcuKxgW6ENZTKZtENnkjAzEUrIA7jxVdy4UMStD+DOt3GSBtTWAwMf597L3Hu8mDOlbfvLqiwtr6yuVddrG5tb2zv13b2uihJJaIdEPJJ3HlaUs5B2NNOc3sWSYuFx2vMmV3m9d0+lYlF4q6cxdQUehSxgBGtjDeuNgcB67AVpkp2gHx4WKEU6Zn5muuymXQgtglNCA0q1h/XPgR+RRNBQE | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 230 | x4WKEU6Zn5muuymXQgtglNCA0q1h/XPgR+RRNBQE46V6jt2rN0US80Ip1ltkCgaYzLBI9o3GGJBlZsWx2ToyDg+CiJpXqhR4f6eSLFQaio805nvqOZruflfrZ/o4MJNWRgnmoZk9lGQcKQjlCeDfCYp0XxqABPJzK6IjLHERJv8aiYEZ/7kReieNh276dycNVqXZRxVOIBDOAYHzqEF19CGDhB4gCd4gVfr0Xq23qz3WWvFKmf24Y+sj28GGpw2</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 231 | sha1_base64="UO6spgXZarocGO1zBKohmEwCj+c=">AAACDHicbZDNSsNAFIVv6l+tf1WXbgaL4EJKIoIuXBTcuKxgW6ENZTKZtENnkjAzEUrIA7jxVdy4UMStD+DOt3GSBtTWAwMf597L3Hu8mDOlbfvLqiwtr6yuVddrG5tb2zv13b2uihJJaIdEPJJ3HlaUs5B2NNOc3sWSYuFx2vMmV3m9d0+lYlF4q6cxdQUehSxgBGtjDeuNgcB67AVpkp2gHx4WKEU6Zn5muuymXQgtglNCA0q1h/XPgR+RRNBQE | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 232 | x4WKEU6Zn5muuymXQgtglNCA0q1h/XPgR+RRNBQE46V6jt2rN0US80Ip1ltkCgaYzLBI9o3GGJBlZsWx2ToyDg+CiJpXqhR4f6eSLFQaio805nvqOZruflfrZ/o4MJNWRgnmoZk9lGQcKQjlCeDfCYp0XxqABPJzK6IjLHERJv8aiYEZ/7kReieNh276dycNVqXZRxVOIBDOAYHzqEF19CGDhB4gCd4gVfr0Xq23qz3WWvFKmf24Y+sj28GGpw2</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 233 | sha1_base64="UO6spgXZarocGO1zBKohmEwCj+c=">AAACDHicbZDNSsNAFIVv6l+tf1WXbgaL4EJKIoIuXBTcuKxgW6ENZTKZtENnkjAzEUrIA7jxVdy4UMStD+DOt3GSBtTWAwMf597L3Hu8mDOlbfvLqiwtr6yuVddrG5tb2zv13b2uihJJaIdEPJJ3HlaUs5B2NNOc3sWSYuFx2vMmV3m9d0+lYlF4q6cxdQUehSxgBGtjDeuNgcB67AVpkp2gHx4WKEU6Zn5muuymXQgtglNCA0q1h/XPgR+RRNBQE | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 234 | x4WKEU6Zn5muuymXQgtglNCA0q1h/XPgR+RRNBQE46V6jt2rN0US80Ip1ltkCgaYzLBI9o3GGJBlZsWx2ToyDg+CiJpXqhR4f6eSLFQaio805nvqOZruflfrZ/o4MJNWRgnmoZk9lGQcKQjlCeDfCYp0XxqABPJzK6IjLHERJv8aiYEZ/7kReieNh276dycNVqXZRxVOIBDOAYHzqEF19CGDhB4gCd4gVfr0Xq23qz3WWvFKmf24Y+sj28GGpw2</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 235 | sha1_base64="UO6spgXZarocGO1zBKohmEwCj+c=">AAACDHicbZDNSsNAFIVv6l+tf1WXbgaL4EJKIoIuXBTcuKxgW6ENZTKZtENnkjAzEUrIA7jxVdy4UMStD+DOt3GSBtTWAwMf597L3Hu8mDOlbfvLqiwtr6yuVddrG5tb2zv13b2uihJJaIdEPJJ3HlaUs5B2NNOc3sWSYuFx2vMmV3m9d0+lYlF4q6cxdQUehSxgBGtjDeuNgcB67AVpkp2gHx4WKEU6Zn5muuymXQgtglNCA0q1h/XPgR+RRNBQE | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 236 | x4WKEU6Zn5muuymXQgtglNCA0q1h/XPgR+RRNBQE46V6jt2rN0US80Ip1ltkCgaYzLBI9o3GGJBlZsWx2ToyDg+CiJpXqhR4f6eSLFQaio805nvqOZruflfrZ/o4MJNWRgnmoZk9lGQcKQjlCeDfCYp0XxqABPJzK6IjLHERJv8aiYEZ/7kReieNh276dycNVqXZRxVOIBDOAYHzqEF19CGDhB4gCd4gVfr0Xq23qz3WWvFKmf24Y+sj28GGpw2</latexit> | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 237 | Edge blockNode blockGlobal block u | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 238 | <latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 239 | sha1_base64="znt8hwWv6wryqwCugrweUa+jkM8=">AAAB7XicbVA9SwNBEJ3zM8avqGBjsxgEq3Bno4VFwMYygrkEkjPubfaSNXu7y+6eEo78BxsLRWz9P3b+GzcfhSY+GHi8N8PMvFhxZqzvf3tLyyura+uFjeLm1vbObmlvPzQy04TWieRSN2NsKGeC1i2znDaVpjiNOW3Eg6ux33ik2jApbu1Q0SjFPcESRrB1UthWfXaXdUplv+JPgBZJMCPl6mH4dA8AtU7pq92VJEupsIRjY1qBr2yUY20Z4XRUb | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 240 | mH4dA8AtU7pq92VJEupsIRjY1qBr2yUY20Z4XRUbGeGKkwGuEdbjgqcUhPlk2tH6MQpXZRI7UpYNFF/T+Q4NWaYxq4zxbZv5r2x+J/XymxyEeVMqMxSQaaLkowjK9H4ddRlmhLLh45gopm7FZE+1phYF1DRhRDMv7xIwrNK4FeCG5fGJUxRgCM4hlMI4ByqcA01qAOBB3iGV3jzpPfivXsf09YlbzZzAH/gff4AqE6Qnw==</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 241 | sha1_base64="Nc0DXje6uYB+/0fHlXL99yCq0no=">AAAB7XicbVC7SgNBFL0bXzG+ooKNzWAQrMKujRYWARvLCGYTSNYwO5lNxszODjOzyrLkH2wsFNHS/7HzA/wPJ49CEw9cOJxzL/feE0rOtHHdL6ewtLyyulZcL21sbm3vlHf3fJ2kitAGSXiiWiHWlDNBG4YZTltSURyHnDbD4eXYb95TpVkibkwmaRDjvmARI9hYye/IAbtNu+WKW3UnQIvEm5FK7cB/kNn3e71b/uz0EpLGVBjCsdZtz5UmyLEyj | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 242 | 5FK7cB/kNn3e71b/uz0EpLGVBjCsdZtz5UmyLEyjHA6KnVSTSUmQ9ynbUsFjqkO8sm1I3RslR6KEmVLGDRRf0/kONY6i0PbGWMz0PPeWPzPa6cmOg9yJmRqqCDTRVHKkUnQ+HXUY4oSwzNLMFHM3orIACtMjA2oZEPw5l9eJP5p1XOr3rVN4wKmKMIhHMEJeHAGNbiCOjSAwB08wjO8OInz5Lw6b9PWgjOb2Yc/cD5+AGXTkqw=</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 243 | sha1_base64="Nc0DXje6uYB+/0fHlXL99yCq0no=">AAAB7XicbVC7SgNBFL0bXzG+ooKNzWAQrMKujRYWARvLCGYTSNYwO5lNxszODjOzyrLkH2wsFNHS/7HzA/wPJ49CEw9cOJxzL/feE0rOtHHdL6ewtLyyulZcL21sbm3vlHf3fJ2kitAGSXiiWiHWlDNBG4YZTltSURyHnDbD4eXYb95TpVkibkwmaRDjvmARI9hYye/IAbtNu+WKW3UnQIvEm5FK7cB/kNn3e71b/uz0EpLGVBjCsdZtz5UmyLEyj | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 244 | 5FK7cB/kNn3e71b/uz0EpLGVBjCsdZtz5UmyLEyjHA6KnVSTSUmQ9ynbUsFjqkO8sm1I3RslR6KEmVLGDRRf0/kONY6i0PbGWMz0PPeWPzPa6cmOg9yJmRqqCDTRVHKkUnQ+HXUY4oSwzNLMFHM3orIACtMjA2oZEPw5l9eJP5p1XOr3rVN4wKmKMIhHMEJeHAGNbiCOjSAwB08wjO8OInz5Lw6b9PWgjOb2Yc/cD5+AGXTkqw=</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 245 | sha1_base64="S5XnA5iYIAgqxiI+i0ptSwAiKP4=">AAAB7XicbVA9SwNBEJ2LXzF+RS1tFoNgFe5stLAI2FhGMB+QnGFvM5es2ds9dveEEPIfbCwUsfX/2Plv3CRXaOKDgcd7M8zMi1LBjfX9b6+wtr6xuVXcLu3s7u0flA+PmkZlmmGDKaF0O6IGBZfYsNwKbKcaaRIJbEWjm5nfekJtuJL3dpximNCB5DFn1Dqp2U2H/CHrlSt+1Z+DrJIgJxXIUe+Vv7p9xbIEpWWCGtMJ/NSGE6otZwKnpW5mMKVsR | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 246 | e+Vv7p9xbIEpWWCGtMJ/NSGE6otZwKnpW5mMKVsRAfYcVTSBE04mV87JWdO6ZNYaVfSkrn6e2JCE2PGSeQ6E2qHZtmbif95nczGV+GEyzSzKNliUZwJYhWZvU76XCOzYuwIZZq7WwkbUk2ZdQGVXAjB8surpHlRDfxqcOdXatd5HEU4gVM4hwAuoQa3UIcGMHiEZ3iFN095L96797FoLXj5zDH8gff5A53Djxw=</latexit> | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 247 | <latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 248 | sha1_base64="m8MJ1M94ujO0d0COo5n2Dsol6rc=">AAAB53icbVC7SgNBFL0bXzG+opY2g0GwCrs2phAM2FhGMA9IFpmdvZsMmZ1dZmaFsKS0sbFQxNZP8Rfs/AZ/wsmj0MQDFw7nnMt9BKng2rjul1NYWV1b3yhulra2d3b3yvsHLZ1kimGTJSJRnYBqFFxi03AjsJMqpHEgsB0MryZ++x6V5om8NaMU/Zj2JY84o8ZKjbtyxa26U5Bl4s1J5fLj+4EAgM1/9sKEZTFKwwTVuuu5qfFzqgxnAselXqYxp | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 249 | 4EAgM1/9sKEZTFKwwTVuuu5qfFzqgxnAselXqYxpWxI+9i1VNIYtZ9P9xyTE6uEJEqULWnIVP3dkdNY61Ec2GRMzUAvehPxP6+bmajm51ymmUHJZoOiTBCTkMnRJOQKmREjSyhT3O5K2IAqyox9Tck+wVs8eZm0zqqeW/Vu3Er9AmYowhEcwyl4cA51uIYGNIFBCI/wDC8Od56cV+dtFi04855D+APn/Qd/sI8A</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 250 | sha1_base64="q1zM5ZsCNMJAtNUZI97B98ATlaA=">AAAB53icbVC7SgNBFL3rM8ZX1FKQwSBYhV0bLQQDNpYJmAckQWZn7yZDZmeXmVkhLCltbCwUsfUvbP0FO79BP8LJo9DEAxcO55zLffiJ4Nq47qezsLi0vLKaW8uvb2xubRd2dus6ThXDGotFrJo+1Si4xJrhRmAzUUgjX2DD71+O/MYtKs1jeW0GCXYi2pU85IwaK1VuCkW35I5B5ok3JcWL96+7g7fqt81/tIOYpRFKwwTVuuW5ielkVBnOBA7z7 | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 251 | 6+7g7fqt81/tIOYpRFKwwTVuuW5ielkVBnOBA7z7VRjQlmfdrFlqaQR6k423nNIjqwSkDBWtqQhY/V3R0YjrQeRb5MRNT09643E/7xWasKzTsZlkhqUbDIoTAUxMRkdTQKukBkxsIQyxe2uhPWooszY1+TtE7zZk+dJ/aTkuSWv6hbL5zBBDvbhEI7Bg1MowxVUoAYMAriHR3hyuPPgPDsvk+iCM+3Zgz9wXn8AGkeQ8w==</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 252 | sha1_base64="q1zM5ZsCNMJAtNUZI97B98ATlaA=">AAAB53icbVC7SgNBFL3rM8ZX1FKQwSBYhV0bLQQDNpYJmAckQWZn7yZDZmeXmVkhLCltbCwUsfUvbP0FO79BP8LJo9DEAxcO55zLffiJ4Nq47qezsLi0vLKaW8uvb2xubRd2dus6ThXDGotFrJo+1Si4xJrhRmAzUUgjX2DD71+O/MYtKs1jeW0GCXYi2pU85IwaK1VuCkW35I5B5ok3JcWL96+7g7fqt81/tIOYpRFKwwTVuuW5ielkVBnOBA7z7 | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 253 | 6+7g7fqt81/tIOYpRFKwwTVuuW5ielkVBnOBA7z7VRjQlmfdrFlqaQR6k423nNIjqwSkDBWtqQhY/V3R0YjrQeRb5MRNT09643E/7xWasKzTsZlkhqUbDIoTAUxMRkdTQKukBkxsIQyxe2uhPWooszY1+TtE7zZk+dJ/aTkuSWv6hbL5zBBDvbhEI7Bg1MowxVUoAYMAriHR3hyuPPgPDsvk+iCM+3Zgz9wXn8AGkeQ8w==</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 254 | sha1_base64="ioxb3woZF1oAlTScqds23PrgiMY=">AAAB53icbVC7SgNBFL3rM8ZX1NJmMAhWYdZGC4uAjWUE84BkkdnZ2WTI7Owyc1cIS37AxkIRW3/Jzr9xkmyhiQcGDuecy9x7wkxJi5R+e2vrG5tb25Wd6u7e/sFh7ei4Y9PccNHmqUpNL2RWKKlFGyUq0cuMYEmoRDcc38787pMwVqb6ASeZCBI21DKWnKGTWo+1Om3QOcgq8UtShxIu/zWIUp4nQiNXzNq+TzMMCmZQciWm1UFuRcb4mA1F31HNE | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 255 | p4nQiNXzNq+TzMMCmZQciWm1UFuRcb4mA1F31HNEmGDYr7nlJw7JSJxatzTSObq74mCJdZOktAlE4Yju+zNxP+8fo7xdVBIneUoNF98FOeKYEpmR5NIGsFRTRxh3Ei3K+EjZhhHV03VleAvn7xKOpcNnzb8e1pv3pR1VOAUzuACfLiCJtxBC9rAIYJneIU3T3ov3rv3sYiueeXMCfyB9/kDCGmMcA==</latexit> v | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 256 | <latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 257 | sha1_base64="HCiXjOq04H3f4Ed7vqyiRfd+2dI=">AAAB7XicbVA9SwNBEJ2LXzF+nQo2NotBsAp3NlpYBGwsI5hLIDnj3mYvWbO3e+zuRcKR/2BjoYit/8fOf+Pmo9DEBwOP92aYmRelnGnjed9OYWV1bX2juFna2t7Z3XP3DwItM0VonUguVTPCmnImaN0ww2kzVRQnEaeNaHA98RtDqjST4s6MUhomuCdYzAg2VgraaZ/dDztu2at4U6Bl4s9JuXoUPD0AQK3jfrW7kmQJFYZwrHXL91IT5lgZRjgdl | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 258 | XoUPD0AQK3jfrW7kmQJFYZwrHXL91IT5lgZRjgdl9qZpikmA9yjLUsFTqgO8+m1Y3RqlS6KpbIlDJqqvydynGg9SiLbmWDT14veRPzPa2UmvgxzJtLMUEFmi+KMIyPR5HXUZYoSw0eWYKKYvRWRPlaYGBtQyYbgL768TILziu9V/FubxhXMUIRjOIEz8OECqnADNagDgUd4hld4c6Tz4rw7H7PWgjOfOYQ/cD5/AKnSkKA=</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 259 | sha1_base64="/voSHBGyFE5xYPVKPYM/GTIarLQ=">AAAB7XicbVC7SgNBFL3rM8ZXVLCxWQyCVdi10cIiYGMZwWwCyRpmJ7PJmNmZYWY2siz5BxsLRbT0f+z8AP/DyaPQxAMXDufcy733RJJRbTzvy1laXlldWy9sFDe3tnd2S3v7gRapwqSOBROqGSFNGOWkbqhhpCkVQUnESCMaXI39xpAoTQW/NZkkYYJ6nMYUI2OloC379G7YKZW9ijeBu0j8GSlXD4MHmX2/1zqlz3ZX4DQh3GCGtG75njRhjpShm | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 260 | SlXD4MHmX2/1zqlz3ZX4DQh3GCGtG75njRhjpShmJFRsZ1qIhEeoB5pWcpRQnSYT64duSdW6bqxULa4cSfq74kcJVpnSWQ7E2T6et4bi/95rdTEF2FOuUwN4Xi6KE6Za4Q7ft3tUkWwYZklCCtqb3VxHymEjQ2oaEPw519eJMFZxfcq/o1N4xKmKMARHMMp+HAOVbiGGtQBwz08wjO8OMJ5cl6dt2nrkjObOYA/cD5+AGdXkq0=</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 261 | sha1_base64="/voSHBGyFE5xYPVKPYM/GTIarLQ=">AAAB7XicbVC7SgNBFL3rM8ZXVLCxWQyCVdi10cIiYGMZwWwCyRpmJ7PJmNmZYWY2siz5BxsLRbT0f+z8AP/DyaPQxAMXDufcy733RJJRbTzvy1laXlldWy9sFDe3tnd2S3v7gRapwqSOBROqGSFNGOWkbqhhpCkVQUnESCMaXI39xpAoTQW/NZkkYYJ6nMYUI2OloC379G7YKZW9ijeBu0j8GSlXD4MHmX2/1zqlz3ZX4DQh3GCGtG75njRhjpShm | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 262 | SlXD4MHmX2/1zqlz3ZX4DQh3GCGtG75njRhjpShmJFRsZ1qIhEeoB5pWcpRQnSYT64duSdW6bqxULa4cSfq74kcJVpnSWQ7E2T6et4bi/95rdTEF2FOuUwN4Xi6KE6Za4Q7ft3tUkWwYZklCCtqb3VxHymEjQ2oaEPw519eJMFZxfcq/o1N4xKmKMARHMMp+HAOVbiGGtQBwz08wjO8OMJ5cl6dt2nrkjObOYA/cD5+AGdXkq0=</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 263 | sha1_base64="Fc8T4ygtia14k1z/CDji4ezWDqY=">AAAB7XicbVA9SwNBEJ3zM8avqKXNYhCswp2NFhYBG8sI5gOSM+xtNsmavd1jdy4QjvwHGwtFbP0/dv4bN8kVmvhg4PHeDDPzokQKi77/7a2tb2xubRd2irt7+weHpaPjhtWpYbzOtNSmFVHLpVC8jgIlbyWG0ziSvBmNbmd+c8yNFVo94CThYUwHSvQFo+ikRicZisdxt1T2K/4cZJUEOSlDjlq39NXpaZbGXCGT1Np24CcYZtSgYJJPi53U8oSyE | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 264 | lq39NXpaZbGXCGT1Np24CcYZtSgYJJPi53U8oSyER3wtqOKxtyG2fzaKTl3So/0tXGlkMzV3xMZja2dxJHrjCkO7bI3E//z2in2r8NMqCRFrthiUT+VBDWZvU56wnCGcuIIZUa4WwkbUkMZuoCKLoRg+eVV0risBH4luPfL1Zs8jgKcwhlcQABXUIU7qEEdGDzBM7zCm6e9F+/d+1i0rnn5zAn8gff5A59Hjx0=</latexit> e | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 265 | <latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 266 | sha1_base64="gRKFy+QFytmwqWy0cvo5FmmPz8I=">AAAB7XicbVA9SwNBEJ3zM8avqGBjsxgEq3Bno4VFwMYygrkEkjPubeaSNXu3x+6eEo78BxsLRWz9P3b+GzcfhSY+GHi8N8PMvDAVXBvX/XaWlldW19YLG8XNre2d3dLevq9lphjWmRRSNUOqUfAE64Ybgc1UIY1DgY1wcDX2G4+oNJfJrRmmGMS0l/CIM2qs5LfTPr/DTqnsVtwJyCLxZqRcPfSf7gGg1il9tbuSZTEmhgmqdctzUxPkVBnOBI6K7 | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
1806.01261 | 267 | fSf7gGg1il9tbuSZTEmhgmqdctzUxPkVBnOBI6K7UxjStmA9rBlaUJj1EE+uXZETqzSJZFUthJDJurviZzGWg/j0HbG1PT1vDcW//NamYkugpwnaWYwYdNFUSaIkWT8OulyhcyIoSWUKW5vJaxPFWXGBlS0IXjzLy8S/6ziuRXvxqZxCVMU4AiO4RQ8OIcqXEMN6sDgAZ7hFd4c6bw4787HtHXJmc0cwB84nz+QDpCP</latexit><latexit | 1806.01261 | Relational inductive biases, deep learning, and graph networks | Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice. | http://arxiv.org/pdf/1806.01261 | [
"Peter W. Battaglia",
"Jessica B. Hamrick",
"Victor Bapst",
"Alvaro Sanchez-Gonzalez",
"Vinicius Zambaldi",
"Mateusz Malinowski",
"Andrea Tacchetti",
"David Raposo",
"Adam Santoro",
"Ryan Faulkner",
"Caglar Gulcehre",
"Francis Song",
"Andrew Ballard",
"Justin Gilmer",
"George Dahl",
"Ashish Vaswani",
"Kelsey Allen",
"Charles Nash",
"Victoria Langston",
"Chris Dyer",
"Nicolas Heess",
"Daan Wierstra",
"Pushmeet Kohli",
"Matt Botvinick",
"Oriol Vinyals",
"Yujia Li",
"Razvan Pascanu"
] | [
"cs.LG",
"cs.AI",
"stat.ML"
] | null | null | cs.LG | 20180604 | 20181017 | [] |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.