Unnamed: 0.1
int64 0
113k
| Unnamed: 0
float64 0
113k
⌀ | title
stringlengths 7
246
| abstract
stringlengths 6
3.31k
|
|---|---|---|---|
107,300 | 107,300 |
TorchSparse: Efficient Point Cloud Inference Engine
|
Deep learning on point clouds has received increased attention thanks to its
wide applications in AR/VR and autonomous driving. These applications require
low latency and high accuracy to provide real-time user experience and ensure
user safety. Unlike conventional dense workloads, the sparse and irregular
nature of point clouds poses severe challenges to running sparse CNNs
efficiently on the general-purpose hardware. Furthermore, existing sparse
acceleration techniques for 2D images do not translate to 3D point clouds. In
this paper, we introduce TorchSparse, a high-performance point cloud inference
engine that accelerates the sparse convolution computation on GPUs. TorchSparse
directly optimizes the two bottlenecks of sparse convolution: irregular
computation and data movement. It applies adaptive matrix multiplication
grouping to trade computation for better regularity, achieving 1.4-1.5x speedup
for matrix multiplication. It also optimizes the data movement by adopting
vectorized, quantized and fused locality-aware memory access, reducing the
memory movement cost by 2.7x. Evaluated on seven representative models across
three benchmark datasets, TorchSparse achieves 1.6x and 1.5x measured
end-to-end speedup over the state-of-the-art MinkowskiEngine and SpConv,
respectively.
|
107,301 | 107,301 |
Learning Future Object Prediction with a Spatiotemporal Detection
Transformer
|
We explore future object prediction -- a challenging problem where all
objects visible in a future video frame are to be predicted. We propose to
tackle this problem end-to-end by training a detection transformer to directly
output future objects. In order to make accurate predictions about the future,
it is necessary to capture the dynamics in the scene, both of other objects and
of the ego-camera. We extend existing detection transformers in two ways to
capture the scene dynamics. First, we experiment with three different
mechanisms that enable the model to spatiotemporally process multiple frames.
Second, we feed ego-motion information to the model via cross-attention. We
show that both of these cues substantially improve future object prediction
performance. Our final approach learns to capture the dynamics and make
predictions on par with an oracle for 100 ms prediction horizons, and
outperform baselines for longer prediction horizons.
|
107,302 | 107,302 |
Simulate Time-integrated Coarse-grained Molecular Dynamics with
Geometric Machine Learning
|
Molecular dynamics (MD) simulation is the workhorse of various scientific
domains but is limited by high computational cost. Learning-based force fields
have made major progress in accelerating ab-initio MD simulation but are still
not fast enough for many real-world applications that require long-time MD
simulation. In this paper, we adopt a different machine learning approach where
we coarse-grain a physical system using graph clustering, and model the system
evolution with a very large time-integration step using graph neural networks.
A novel score-based GNN refinement module resolves the long-standing challenge
of long-time simulation instability. Despite only trained with short MD
trajectory data, our learned simulator can generalize to unseen novel systems
and simulate for much longer than the training trajectories. Properties
requiring 10-100 ns level long-time dynamics can be accurately recovered at
several-orders-of-magnitude higher speed than classical force fields. We
demonstrate the effectiveness of our method on two realistic complex systems:
(1) single-chain coarse-grained polymers in implicit solvent; (2)
multi-component Li-ion polymer electrolyte systems.
|
107,303 | 107,303 |
Provably Efficient Kernelized Q-Learning
|
We propose and analyze a kernelized version of Q-learning. Although a kernel
space is typically infinite-dimensional, extensive study has shown that
generalization is only affected by the effective dimension of the data. We
incorporate such ideas into the Q-learning framework and derive regret bounds
for arbitrary kernels. In particular, we provide concrete bounds for linear
kernels and Gaussian RBF kernels; notably, the latter bound looks almost
identical to the former, only that the actual dimension is replaced by a
different notion of dimensionality. Finally, we test our algorithm on a suite
of classic control tasks; remarkably, under the Gaussian RBF kernel, it
achieves reasonably good performance after only 1000 environmental steps, while
its neural network counterpart, deep Q-learning, still struggles.
|
107,304 | 107,304 |
A Framework for Interactive Knowledge-Aided Machine Teaching
|
Machine Teaching (MT) is an interactive process where humans train a machine
learning model by playing the role of a teacher. The process of designing an MT
system involves decisions that can impact both efficiency of human teachers and
performance of machine learners. Previous research has proposed and evaluated
specific MT systems but there is limited discussion on a general framework for
designing them. We propose a framework for designing MT systems and also detail
a system for the text classification problem as a specific instance. Our
framework focuses on three components i.e. teaching interface, machine learner,
and knowledge base; and their relations describe how each component can benefit
the others. Our preliminary experiments show how MT systems can reduce both
human teaching time and machine learner error rate.
|
107,305 | 107,305 |
Model-free Learning of Regions of Attraction via Recurrent Sets
|
We consider the problem of learning an inner approximation of the region of
attraction (ROA) of an asymptotically stable equilibrium point without an
explicit model of the dynamics. Rather than leveraging approximate models with
bounded uncertainty to find a (robust) invariant set contained in the ROA, we
propose to learn sets that satisfy a more relaxed notion of containment known
as recurrence. We define a set to be $\tau$-recurrent (resp. $k$-recurrent) if
every trajectory that starts within the set, returns to it after at most $\tau$
seconds (resp. $k$ steps). We show that under mild assumptions a
$\tau$-recurrent set containing a stable equilibrium must be a subset of its
ROA. We then leverage this property to develop algorithms that compute inner
approximations of the ROA using counter-examples of recurrence that are
obtained by sampling finite-length trajectories. Our algorithms process samples
sequentially, which allow them to continue being executed even after an initial
offline training stage. We further provide an upper bound on the number of
counter-examples used by the algorithm, and almost sure convergence guarantees.
|
107,306 | 107,306 |
Learning how to Interact with a Complex Interface using Hierarchical
Reinforcement Learning
|
Hierarchical Reinforcement Learning (HRL) allows interactive agents to
decompose complex problems into a hierarchy of sub-tasks. Higher-level tasks
can invoke the solutions of lower-level tasks as if they were primitive
actions. In this work, we study the utility of hierarchical decompositions for
learning an appropriate way to interact with a complex interface. Specifically,
we train HRL agents that can interface with applications in a simulated Android
device. We introduce a Hierarchical Distributed Deep Reinforcement Learning
architecture that learns (1) subtasks corresponding to simple finger gestures,
and (2) how to combine these gestures to solve several Android tasks. Our
approach relies on goal conditioning and can be used more generally to convert
any base RL agent into an HRL agent. We use the AndroidEnv environment to
evaluate our approach. For the experiments, the HRL agent uses a distributed
version of the popular DQN algorithm to train different components of the
hierarchy. While the native action space is completely intractable for simple
DQN agents, our architecture can be used to establish an effective way to
interact with different tasks, significantly improving the performance of the
same DQN agent over different levels of abstraction.
|
107,307 | 107,307 |
Differentially Private Learning with Margin Guarantees
|
We present a series of new differentially private (DP) algorithms with
dimension-independent margin guarantees. For the family of linear hypotheses,
we give a pure DP learning algorithm that benefits from relative deviation
margin guarantees, as well as an efficient DP learning algorithm with margin
guarantees. We also present a new efficient DP learning algorithm with margin
guarantees for kernel-based hypotheses with shift-invariant kernels, such as
Gaussian kernels, and point out how our results can be extended to other
kernels using oblivious sketching techniques. We further give a pure DP
learning algorithm for a family of feed-forward neural networks for which we
prove margin guarantees that are independent of the input dimension.
Additionally, we describe a general label DP learning algorithm, which benefits
from relative deviation margin bounds and is applicable to a broad family of
hypothesis sets, including that of neural networks. Finally, we show how our DP
learning algorithms can be augmented in a general way to include model
selection, to select the best confidence margin parameter.
|
107,308 | 107,308 |
Facilitating automated conversion of scientific knowledge into
scientific simulation models with the Machine Assisted Generation,
Calibration, and Comparison (MAGCC) Framework
|
The Machine Assisted Generation, Comparison, and Calibration (MAGCC)
framework provides machine assistance and automation of recurrent crucial steps
and processes in the development, implementation, testing, and use of
scientific simulation models. MAGCC bridges systems for knowledge extraction
via natural language processing or extracted from existing mathematical models
and provides a comprehensive workflow encompassing the composition of
scientific models and artificial intelligence (AI) assisted code generation.
MAGCC accomplishes this through: 1) the development of a comprehensively
expressive formal knowledge representation knowledgebase, the Structured
Scientific Knowledge Representation (SSKR) that encompasses all the types of
information needed to make any simulation model, 2) the use of an artificially
intelligent logic reasoning system, the Computational Modeling Assistant (CMA),
that takes information from the SSKR and generates, in a traceable fashion,
model specifications across a range of simulation modeling methods, and 3) the
use of the CMA to generate executable code for a simulation model from those
model specifications. The MAGCC framework can be customized any scientific
domain, and future work will integrate newly developed code-generating AI
systems.
|
107,309 | 107,309 |
SoftEdge: Regularizing Graph Classification with Random Soft Edges
|
Graph data augmentation plays a vital role in regularizing Graph Neural
Networks (GNNs), which leverage information exchange along edges in graphs, in
the form of message passing, for learning. Due to their effectiveness, simple
edge and node manipulations (e.g., addition and deletion) have been widely used
in graph augmentation. In this paper, we identify a limitation in such a common
augmentation technique. That is, simple edge and node manipulations can create
graphs with an identical structure or indistinguishable structures to message
passing GNNs but of conflict labels, leading to the sample collision issue and
thus the degradation of model performance. To address this problem, we propose
SoftEdge, which assigns random weights to a portion of the edges of a given
graph to construct dynamic neighborhoods over the graph. We prove that SoftEdge
creates collision-free augmented graphs. We also show that this simple method
obtains superior accuracy to popular node and edge manipulation approaches and
notable resilience to the accuracy degradation with the GNN depth.
|
107,310 | 107,310 |
Optimizing Nitrogen Management with Deep Reinforcement Learning and Crop
Simulations
|
Nitrogen (N) management is critical to sustain soil fertility and crop
production while minimizing the negative environmental impact, but is
challenging to optimize. This paper proposes an intelligent N management system
using deep reinforcement learning (RL) and crop simulations with Decision
Support System for Agrotechnology Transfer (DSSAT). We first formulate the N
management problem as an RL problem. We then train management policies with
deep Q-network and soft actor-critic algorithms, and the Gym-DSSAT interface
that allows for daily interactions between the simulated crop environment and
RL agents. According to the experiments on the maize crop in both Iowa and
Florida in the US, our RL-trained policies outperform previous empirical
methods by achieving higher or similar yield while using less fertilizers
|
107,311 | 107,311 |
STD: A Seasonal-Trend-Dispersion Decomposition of Time Series
|
The decomposition of a time series is an essential task that helps to
understand its very nature. It facilitates the analysis and forecasting of
complex time series expressing various hidden components such as the trend,
seasonal components, cyclic components and irregular fluctuations. Therefore,
it is crucial in many fields for forecasting and decision processes. In recent
years, many methods of time series decomposition have been developed, which
extract and reveal different time series properties. Unfortunately, they
neglect a very important property, i.e. time series variance. To deal with
heteroscedasticity in time series, the method proposed in this work -- a
seasonal-trend-dispersion decomposition (STD) -- extracts the trend, seasonal
component and component related to the dispersion of the time series. We define
STD decomposition in two ways: with and without an irregular component. We show
how STD can be used for time series analysis and forecasting.
|
107,312 | 107,312 |
Interpolation of Missing Swaption Volatility Data using Gibbs Sampling
on Variational Autoencoders
|
Albeit of crucial interest for both financial practitioners and researchers,
market-implied volatility data of European swaptions often exhibit large
portions of missing quotes due to illiquidity of the various underlying
swaption instruments. In this case, standard stochastic interpolation tools
like the common SABR model often cannot be calibrated to observed implied
volatility smiles, due to data being only available for the at-the-money quote
of the respective underlying swaption. Here, we propose to infer the geometry
of the full unknown implied volatility cube by learning stochastic latent
representations of implied volatility cubes via variational autoencoders,
enabling inference about the missing volatility data conditional on the
observed data by an approximate Gibbs sampling approach. Imputed estimates of
missing quotes can afterwards be used to fit a standard stochastic volatility
model. Since training data for the employed variational autoencoder model is
usually sparsely available, we test the robustness of the approach for a model
trained on synthetic data on real market quotes and we show that SABR
interpolated volatilites calibrated to reconstructed volatility cubes with
artificially imputed missing values differ by not much more than two basis
points compared to SABR fits calibrated to the complete cube. Moreover, we show
how the imputation can be used to successfully set up delta-neutral portfolios
for hedging purposes.
|
107,313 | 107,313 |
ICDBigBird: A Contextual Embedding Model for ICD Code Classification
|
The International Classification of Diseases (ICD) system is the
international standard for classifying diseases and procedures during a
healthcare encounter and is widely used for healthcare reporting and management
purposes. Assigning correct codes for clinical procedures is important for
clinical, operational, and financial decision-making in healthcare. Contextual
word embedding models have achieved state-of-the-art results in multiple NLP
tasks. However, these models have yet to achieve state-of-the-art results in
the ICD classification task since one of their main disadvantages is that they
can only process documents that contain a small number of tokens which is
rarely the case with real patient notes. In this paper, we introduce ICDBigBird
a BigBird-based model which can integrate a Graph Convolutional Network (GCN),
that takes advantage of the relations between ICD codes in order to create
'enriched' representations of their embeddings, with a BigBird contextual model
that can process larger documents. Our experiments on a real-world clinical
dataset demonstrate the effectiveness of our BigBird-based model on the ICD
classification task as it outperforms the previous state-of-the-art models.
|
107,314 | 107,314 |
Staying the course: Locating equilibria of dynamical systems on
Riemannian manifolds defined by point-clouds
|
We introduce a method to successively locate equilibria (steady states) of
dynamical systems on Riemannian manifolds. The manifolds need not be
characterized by an atlas or by the zeros of a smooth map. Instead, they can be
defined by point-clouds and sampled as needed through an iterative process. If
the manifold is an Euclidean space, our method follows isoclines, curves along
which the direction of the vector field $X$ is constant. For a generic vector
field $X$, isoclines are smooth curves and every equilibrium is a limit point
of isoclines. We generalize the definition of isoclines to Riemannian manifolds
through the use of parallel transport: generalized isoclines are curves along
which the directions of $X$ are parallel transports of each other. As in the
Euclidean case, generalized isoclines of generic vector fields $X$ are smooth
curves that connect equilibria of $X$.
Our work is motivated by computational statistical mechanics, specifically
high dimensional (stochastic) differential equations that model the dynamics of
molecular systems. Often, these dynamics concentrate near low-dimensional
manifolds and have transitions (sadddle points with a single unstable
direction) between metastable equilibria We employ iteratively sampled data and
isoclines to locate these saddle points. Coupling a black-box sampling scheme
(e.g., Markov chain Monte Carlo) with manifold learning techniques (diffusion
maps in the case presented here), we show that our method reliably locates
equilibria of $X$.
|
107,315 | 107,315 |
A Top-Down Approach to Hierarchically Coherent Probabilistic Forecasting
|
Hierarchical forecasting is a key problem in many practical multivariate
forecasting applications - the goal is to obtain coherent predictions for a
large number of correlated time series that are arranged in a pre-specified
tree hierarchy. In this paper, we present a probabilistic top-down approach to
hierarchical forecasting that uses a novel attention-based RNN model to learn
the distribution of the proportions according to which each parent prediction
is split among its children nodes at any point in time. These probabilistic
proportions are then coupled with an independent univariate probabilistic
forecasting model (such as Prophet or STS) for the root time series. The
resulting forecasts are computed in a top-down fashion and are naturally
coherent, and also support probabilistic predictions over all time series in
the hierarchy. We provide theoretical justification for the superiority of our
top-down approach compared to traditional bottom-up hierarchical modeling.
Finally, we experiment on three public datasets and demonstrate significantly
improved probabilistic forecasts, compared to state-of-the-art probabilistic
hierarchical models.
|
107,316 | 107,316 |
CycleSense: Detecting Near Miss Incidents in Bicycle Traffic from Mobile
Motion Sensors
|
In cities worldwide, cars cause health and traffic problems which could be
partly mitigated through an increased modal share of bicycles. Many people,
however, avoid cycling due to a lack of perceived safety. For city planners,
addressing this is hard as they lack insights into where cyclists feel safe and
where they do not. To gain such insights, we have in previous work proposed the
crowdsourcing platform SimRa, which allows cyclists to record their rides and
report near miss incidents via a smartphone app. In this paper, we present
CycleSense, a combination of signal processing and Machine Learning techniques,
which partially automates the detection of near miss incidents. Using the SimRa
data set, we evaluate CycleSense by comparing it to a baseline method used by
SimRa and show that it significantly improves incident detection.
|
107,317 | 107,317 |
Lightweight Hybrid CNN-ELM Model for Multi-building and Multi-floor
Classification
|
Machine learning models have become an essential tool in current indoor
positioning solutions, given their high capabilities to extract meaningful
information from the environment. Convolutional neural networks (CNNs) are one
of the most used neural networks (NNs) due to that they are capable of learning
complex patterns from the input data. Another model used in indoor positioning
solutions is the Extreme Learning Machine (ELM), which provides an acceptable
generalization performance as well as a fast speed of learning. In this paper,
we offer a lightweight combination of CNN and ELM, which provides a quick and
accurate classification of building and floor, suitable for power and
resource-constrained devices. As a result, the proposed model is 58\% faster
than the benchmark, with a slight improvement in the classification accuracy
(by less than 1\%
|
107,318 | 107,318 |
Learning Sequential Latent Variable Models from Multimodal Time Series
Data
|
Sequential modelling of high-dimensional data is an important problem that
appears in many domains including model-based reinforcement learning and
dynamics identification for control. Latent variable models applied to
sequential data (i.e., latent dynamics models) have been shown to be a
particularly effective probabilistic approach to solve this problem, especially
when dealing with images. However, in many application areas (e.g., robotics),
information from multiple sensing modalities is available -- existing latent
dynamics methods have not yet been extended to effectively make use of such
multimodal sequential data. Multimodal sensor streams can be correlated in a
useful manner and often contain complementary information across modalities. In
this work, we present a self-supervised generative modelling framework to
jointly learn a probabilistic latent state representation of multimodal data
and the respective dynamics. Using synthetic and real-world datasets from a
multimodal robotic planar pushing task, we demonstrate that our approach leads
to significant improvements in prediction and representation quality.
Furthermore, we compare to the common learning baseline of concatenating each
modality in the latent space and show that our principled probabilistic
formulation performs better. Finally, despite being fully self-supervised, we
demonstrate that our method is nearly as effective as an existing supervised
approach that relies on ground truth labels.
|
107,319 | 107,319 |
Curriculum Learning for Goal-Oriented Semantic Communications with a
Common Language
|
Goal-oriented semantic communication will be a pillar of next-generation
wireless networks. Despite significant recent efforts in this area, most prior
works are focused on specific data types (e.g., image or audio), and they
ignore the goal and effectiveness aspects of semantic transmissions. In
contrast, in this paper, a holistic goal-oriented semantic communication
framework is proposed to enable a speaker and a listener to cooperatively
execute a set of sequential tasks in a dynamic environment. A common language
based on a hierarchical belief set is proposed to enable semantic
communications between speaker and listener. The speaker, acting as an observer
of the environment, utilizes the beliefs to transmit an initial description of
its observation (called event) to the listener. The listener is then able to
infer on the transmitted description and complete it by adding related beliefs
to the transmitted beliefs of the speaker. As such, the listener reconstructs
the observed event based on the completed description, and it then takes
appropriate action in the environment based on the reconstructed event. An
optimization problem is defined to determine the perfect and abstract
description of the events while minimizing the transmission and inference costs
with constraints on the task execution time and belief efficiency. Then, a
novel bottom-up curriculum learning (CL) framework based on reinforcement
learning is proposed to solve the optimization problem and enable the speaker
and listener to gradually identify the structure of the belief set and the
perfect and abstract description of the events. Simulation results show that
the proposed CL method outperforms traditional RL in terms of convergence time,
task execution cost and time, reliability, and belief efficiency.
|
107,320 | 107,320 |
Scale-Equivariant Unrolled Neural Networks for Data-Efficient
Accelerated MRI Reconstruction
|
Unrolled neural networks have enabled state-of-the-art reconstruction
performance and fast inference times for the accelerated magnetic resonance
imaging (MRI) reconstruction task. However, these approaches depend on
fully-sampled scans as ground truth data which is either costly or not possible
to acquire in many clinical medical imaging applications; hence, reducing
dependence on data is desirable. In this work, we propose modeling the proximal
operators of unrolled neural networks with scale-equivariant convolutional
neural networks in order to improve the data-efficiency and robustness to
drifts in scale of the images that might stem from the variability of patient
anatomies or change in field-of-view across different MRI scanners. Our
approach demonstrates strong improvements over the state-of-the-art unrolled
neural networks under the same memory constraints both with and without data
augmentations on both in-distribution and out-of-distribution scaled images
without significantly increasing the train or inference time.
|
107,321 | 107,321 |
Evolution of Transparent Explainable Rule-sets
|
Most AI systems are black boxes generating reasonable outputs for given
inputs. Some domains, however, have explainability and trustworthiness
requirements that cannot be directly met by these approaches. Various methods
have therefore been developed to interpret black-box models after training.
This paper advocates an alternative approach where the models are transparent
and explainable to begin with. This approach, EVOTER, evolves rule-sets based
on simple logical expressions. The approach is evaluated in several
prediction/classification and prescription/policy search domains with and
without a surrogate. It is shown to discover meaningful rule sets that perform
similarly to black-box models. The rules can provide insight to the domain, and
make biases hidden in the data explicit. It may also be possible to edit them
directly to remove biases and add constraints. EVOTER thus forms a promising
foundation for building trustworthy AI systems for real-world applications in
the future.
|
107,322 | 107,322 |
Hypergraph Transformer: Weakly-supervised Multi-hop Reasoning for
Knowledge-based Visual Question Answering
|
Knowledge-based visual question answering (QA) aims to answer a question
which requires visually-grounded external knowledge beyond image content
itself. Answering complex questions that require multi-hop reasoning under weak
supervision is considered as a challenging problem since i) no supervision is
given to the reasoning process and ii) high-order semantics of multi-hop
knowledge facts need to be captured. In this paper, we introduce a concept of
hypergraph to encode high-level semantics of a question and a knowledge base,
and to learn high-order associations between them. The proposed model,
Hypergraph Transformer, constructs a question hypergraph and a query-aware
knowledge hypergraph, and infers an answer by encoding inter-associations
between two hypergraphs and intra-associations in both hypergraph itself.
Extensive experiments on two knowledge-based visual QA and two knowledge-based
textual QA demonstrate the effectiveness of our method, especially for
multi-hop reasoning problem. Our source code is available at
https://github.com/yujungheo/kbvqa-public.
|
107,323 | 107,323 |
SCOPE: Safe Exploration for Dynamic Computer Systems Optimization
|
Modern computer systems need to execute under strict safety constraints
(e.g., a power limit), but doing so often conflicts with their ability to
deliver high performance (i.e. minimal latency). Prior work uses machine
learning to automatically tune hardware resources such that the system
execution meets safety constraints optimally. Such solutions monitor past
system executions to learn the system's behavior under different hardware
resource allocations before dynamically tuning resources to optimize the
application execution. However, system behavior can change significantly
between different applications and even different inputs of the same
applications. Hence, the models learned using data collected a priori are often
suboptimal and violate safety constraints when used with new applications and
inputs. To address this limitation, we introduce the concept of an execution
space, which is the cross product of hardware resources, input features, and
applications. To dynamically and safely allocate hardware resources from the
execution space, we present SCOPE, a resource manager that leverages a novel
safe exploration framework. We evaluate SCOPE's ability to deliver improved
latency while minimizing power constraint violations by dynamically configuring
hardware while running a variety of Apache Spark applications. Compared to
prior approaches that minimize power constraint violations, SCOPE consumes
comparable power while improving latency by up to 9.5X. Compared to prior
approaches that minimize latency, SCOPE achieves similar latency but reduces
power constraint violation rates by up to 45.88X, achieving almost zero safety
constraint violations across all applications.
|
107,324 | 107,324 |
Neural Contrastive Clustering: Fully Unsupervised Bias Reduction for
Sentiment Classification
|
Background: Neural networks produce biased classification results due to
correlation bias (they learn correlations between their inputs and outputs to
classify samples, even when those correlations do not represent
cause-and-effect relationships).
Objective: This study introduces a fully unsupervised method of mitigating
correlation bias, demonstrated with sentiment classification on COVID-19 social
media data.
Methods: Correlation bias in sentiment classification often arises in
conversations about controversial topics. Therefore, this study uses
adversarial learning to contrast clusters based on sentiment classification
labels, with clusters produced by unsupervised topic modeling. This discourages
the neural network from learning topic-related features that produce biased
classification results.
Results: Compared to a baseline classifier, neural contrastive clustering
approximately doubles accuracy on bias-prone sentences for human-labeled
COVID-19 social media data, without adversely affecting the classifier's
overall F1 score. Despite being a fully unsupervised approach, neural
contrastive clustering achieves a larger improvement in accuracy on bias-prone
sentences than a supervised masking approach.
Conclusions: Neural contrastive clustering reduces correlation bias in
sentiment text classification. Further research is needed to explore
generalizing this technique to other neural network architectures and
application domains.
|
107,325 | 107,325 |
Gene Function Prediction with Gene Interaction Networks: A Context Graph
Kernel Approach
|
Predicting gene functions is a challenge for biologists in the post genomic
era. Interactions among genes and their products compose networks that can be
used to infer gene functions. Most previous studies adopt a linkage assumption,
i.e., they assume that gene interactions indicate functional similarities
between connected genes. In this study, we propose to use a gene's context
graph, i.e., the gene interaction network associated with the focal gene, to
infer its functions. In a kernel-based machine-learning framework, we design a
context graph kernel to capture the information in context graphs. Our
experimental study on a testbed of p53-related genes demonstrates the advantage
of using indirect gene interactions and shows the empirical superiority of the
proposed approach over linkage-assumption-based methods, such as the algorithm
to minimize inconsistent connected genes and diffusion kernels.
|
107,326 | 107,326 |
Global Mapping of Gene/Protein Interactions in PubMed Abstracts: A
Framework and an Experiment with P53 Interactions
|
Gene/protein interactions provide critical information for a thorough
understanding of cellular processes. Recently, considerable interest and effort
has been focused on the construction and analysis of genome-wide gene networks.
The large body of biomedical literature is an important source of gene/protein
interaction information. Recent advances in text mining tools have made it
possible to automatically extract such documented interactions from free-text
literature. In this paper, we propose a comprehensive framework for
constructing and analyzing large-scale gene functional networks based on the
gene/protein interactions extracted from biomedical literature repositories
using text mining tools. Our proposed framework consists of analyses of the
network topology, network topology-gene function relationship, and temporal
network evolution to distill valuable information embedded in the gene
functional interactions in literature. We demonstrate the application of the
proposed framework using a testbed of P53-related PubMed abstracts, which shows
that literature-based P53 networks exhibit small-world and scale-free
properties. We also found that high degree genes in the literature-based
networks have a high probability of appearing in the manually curated database
and genes in the same pathway tend to form local clusters in our
literature-based networks. Temporal analysis showed that genes interacting with
many other genes tend to be involved in a large number of newly discovered
interactions.
|
107,327 | 107,327 |
Analysis of Temporal Difference Learning: Linear System Approach
|
The goal of this technical note is to introduce a new finite-time convergence
analysis of temporal difference (TD) learning based on stochastic linear system
models. TD-learning is a fundamental reinforcement learning (RL) to evaluate a
given policy by estimating the corresponding value function for a Markov
decision process. While there has been a series of successful works in
theoretical analysis of TDlearning, it was not until recently that researchers
found some guarantees on its statistical efficiency by developing finite-time
error bounds. In this paper, we propose a simple control theoretic finite-time
analysis of TD-learning, which exploits linear system models and standard
notions in linear system communities. The proposed work provides new simple
templets for RL analysis, and additional insights on TD-learning and RL based
on ideas in control theory.
|
107,328 | 107,328 |
NLP Based Anomaly Detection for Categorical Time Series
|
Identifying anomalies in large multi-dimensional time series is a crucial and
difficult task across multiple domains. Few methods exist in the literature
that address this task when some of the variables are categorical in nature. We
formalize an analogy between categorical time series and classical Natural
Language Processing and demonstrate the strength of this analogy for anomaly
detection and root cause investigation by implementing and testing three
different machine learning anomaly detection and root cause investigation
models based upon it.
|
107,329 | 107,329 |
Adversarial Estimators
|
We develop an asymptotic theory of adversarial estimators ('A-estimators').
They generalize maximum-likelihood-type estimators ('M-estimators') as their
objective is maximized by some parameters and minimized by others. This class
subsumes the continuous-updating Generalized Method of Moments, Generative
Adversarial Networks and more recent proposals in machine learning and
econometrics. In these examples, researchers state which aspects of the problem
may in principle be used for estimation, and an adversary learns how to
emphasize them optimally. We derive the convergence rates of A-estimators under
pointwise and partial identification, and the normality of functionals of their
parameters. Unknown functions may be approximated via sieves such as deep
neural networks, for which we provide simplified low-level conditions. As a
corollary, we obtain the normality of neural-net M-estimators, overcoming
technical issues previously identified by the literature. Our theory yields
novel results about a variety of A-estimators, providing intuition and formal
justification for their success in recent applications.
|
107,330 | 107,330 |
Multimodal Adaptive Distillation for Leveraging Unimodal Encoders for
Vision-Language Tasks
|
Cross-modal encoders for vision-language (VL) tasks are often pretrained with
carefully curated vision-language datasets. While these datasets reach an order
of 10 million samples, the labor cost is prohibitive to scale further.
Conversely, unimodal encoders are pretrained with simpler annotations that are
less cost-prohibitive, achieving scales of hundreds of millions to billions. As
a result, unimodal encoders have achieved state-of-art (SOTA) on many
downstream tasks. However, challenges remain when applying to VL tasks. The
pretraining data is not optimal for cross-modal architectures and requires
heavy computational resources. In addition, unimodal architectures lack
cross-modal interactions that have demonstrated significant benefits for VL
tasks. Therefore, how to best leverage pretrained unimodal encoders for VL
tasks is still an area of active research. In this work, we propose a method to
leverage unimodal vision and text encoders for VL tasks that augment existing
VL approaches while conserving computational complexity. Specifically, we
propose Multimodal Adaptive Distillation (MAD), which adaptively distills
useful knowledge from pretrained encoders to cross-modal VL encoders. Second,
to better capture nuanced impacts on VL task performance, we introduce an
evaluation protocol that includes Visual Commonsense Reasoning (VCR), Visual
Entailment (SNLI-VE), and Visual Question Answering (VQA), across a variety of
data constraints and conditions of domain shift. Experiments demonstrate that
MAD leads to consistent gains in the low-shot, domain-shifted, and
fully-supervised conditions on VCR, SNLI-VE, and VQA, achieving SOTA
performance on VCR compared to other single models pretrained with image-text
data. Finally, MAD outperforms concurrent works utilizing pretrained vision
encoder from CLIP. Code will be made available.
|
107,331 | 107,331 |
Application of Federated Learning in Building a Robust COVID-19 Chest
X-ray Classification Model
|
While developing artificial intelligence (AI)-based algorithms to solve
problems, the amount of data plays a pivotal role - large amount of data helps
the researchers and engineers to develop robust AI algorithms. In the case of
building AI-based models for problems related to medical imaging, these data
need to be transferred from the medical institutions where they were acquired
to the organizations developing the algorithms. This movement of data involves
time-consuming formalities like complying with HIPAA, GDPR, etc.There is also a
risk of patients' private data getting leaked, compromising their
confidentiality. One solution to these problems is using the Federated Learning
framework.
Federated Learning (FL) helps AI models to generalize better and create a
robust AI model by using data from different sources having different
distributions and data characteristics without moving all the data to a central
server. In our paper, we apply the FL framework for training a deep learning
model to solve a binary classification problem of predicting the presence or
absence of COVID-19. We took three different sources of data and trained
individual models on each source. Then we trained an FL model on the complete
data and compared all the model performances. We demonstrated that the FL model
performs better than the individual models. Moreover, the FL model performed at
par with the model trained on all the data combined at a central server. Thus
Federated Learning leads to generalized AI models without the cost of data
transfer and regulatory overhead.
|
107,332 | 107,332 |
Taygete at SemEval-2022 Task 4: RoBERTa based models for detecting
Patronising and Condescending Language
|
This work describes the development of different models to detect patronising
and condescending language within extracts of news articles as part of the
SemEval 2022 competition (Task-4). This work explores different models based on
the pre-trained RoBERTa language model coupled with LSTM and CNN layers. The
best models achieved 15$^{th}$ rank with an F1-score of 0.5924 for subtask-A
and 12$^{th}$ in subtask-B with a macro-F1 score of 0.3763.
|
107,333 | 107,333 |
Exploring Hidden Semantics in Neural Networks with Symbolic Regression
|
Many recent studies focus on developing mechanisms to explain the black-box
behaviors of neural networks (NNs). However, little work has been done to
extract the potential hidden semantics (mathematical representation) of a
neural network. A succinct and explicit mathematical representation of a NN
model could improve the understanding and interpretation of its behaviors. To
address this need, we propose a novel symbolic regression method for neural
works (called SRNet) to discover the mathematical expressions of a NN. SRNet
creates a Cartesian genetic programming (NNCGP) to represent the hidden
semantics of a single layer in a NN. It then leverages a multi-chromosome NNCGP
to represent hidden semantics of all layers of the NN. The method uses a
(1+$\lambda$) evolutionary strategy (called MNNCGP-ES) to extract the final
mathematical expressions of all layers in the NN. Experiments on 12 symbolic
regression benchmarks and 5 classification benchmarks show that SRNet not only
can reveal the complex relationships between each layer of a NN but also can
extract the mathematical representation of the whole NN. Compared with LIME and
MAPLE, SRNet has higher interpolation accuracy and trends to approximate the
real model on the practical dataset.
|
107,334 | 107,334 |
Multi-view Information Bottleneck Without Variational Approximation
|
By "intelligently" fusing the complementary information across different
views, multi-view learning is able to improve the performance of classification
tasks. In this work, we extend the information bottleneck principle to a
supervised multi-view learning scenario and use the recently proposed
matrix-based R{\'e}nyi's $\alpha$-order entropy functional to optimize the
resulting objective directly, without the necessity of variational
approximation or adversarial training. Empirical results in both synthetic and
real-world datasets suggest that our method enjoys improved robustness to noise
and redundant information in each view, especially given limited training
samples. Code is available at~\url{https://github.com/archy666/MEIB}.
|
107,335 | 107,335 |
End-to-end symbolic regression with transformers
|
Symbolic regression, the task of predicting the mathematical expression of a
function from the observation of its values, is a difficult task which usually
involves a two-step procedure: predicting the "skeleton" of the expression up
to the choice of numerical constants, then fitting the constants by optimizing
a non-convex loss function. The dominant approach is genetic programming, which
evolves candidates by iterating this subroutine a large number of times. Neural
networks have recently been tasked to predict the correct skeleton in a single
try, but remain much less powerful. In this paper, we challenge this two-step
procedure, and task a Transformer to directly predict the full mathematical
expression, constants included. One can subsequently refine the predicted
constants by feeding them to the non-convex optimizer as an informed
initialization. We present ablations to show that this end-to-end approach
yields better results, sometimes even without the refinement step. We evaluate
our model on problems from the SRBench benchmark and show that our model
approaches the performance of state-of-the-art genetic programming with several
orders of magnitude faster inference.
|
107,336 | 107,336 |
Fourier Imager Network (FIN): A deep neural network for hologram
reconstruction with superior external generalization
|
Deep learning-based image reconstruction methods have achieved remarkable
success in phase recovery and holographic imaging. However, the generalization
of their image reconstruction performance to new types of samples never seen by
the network remains a challenge. Here we introduce a deep learning framework,
termed Fourier Imager Network (FIN), that can perform end-to-end phase recovery
and image reconstruction from raw holograms of new types of samples, exhibiting
unprecedented success in external generalization. FIN architecture is based on
spatial Fourier transform modules that process the spatial frequencies of its
inputs using learnable filters and a global receptive field. Compared with
existing convolutional deep neural networks used for hologram reconstruction,
FIN exhibits superior generalization to new types of samples, while also being
much faster in its image inference speed, completing the hologram
reconstruction task in ~0.04 s per 1 mm^2 of the sample area. We experimentally
validated the performance of FIN by training it using human lung tissue samples
and blindly testing it on human prostate, salivary gland tissue and Pap smear
samples, proving its superior external generalization and image reconstruction
speed. Beyond holographic microscopy and quantitative phase imaging, FIN and
the underlying neural network architecture might open up various new
opportunities to design broadly generalizable deep learning models in
computational imaging and machine vision fields.
|
107,337 | 107,337 |
Sharper Utility Bounds for Differentially Private Models
|
In this paper, by introducing Generalized Bernstein condition, we propose the
first $\mathcal{O}\big(\frac{\sqrt{p}}{n\epsilon}\big)$ high probability excess
population risk bound for differentially private algorithms under the
assumptions $G$-Lipschitz, $L$-smooth, and Polyak-{\L}ojasiewicz condition,
based on gradient perturbation method. If we replace the properties
$G$-Lipschitz and $L$-smooth by $\alpha$-H{\"o}lder smoothness (which can be
used in non-smooth setting), the high probability bound comes to
$\mathcal{O}\big(n^{-\frac{\alpha}{1+2\alpha}}\big)$ w.r.t $n$, which cannot
achieve $\mathcal{O}\left(1/n\right)$ when $\alpha\in(0,1]$. To solve this
problem, we propose a variant of gradient perturbation method,
\textbf{max$\{1,g\}$-Normalized Gradient Perturbation} (m-NGP). We further show
that by normalization, the high probability excess population risk bound under
assumptions $\alpha$-H{\"o}lder smooth and Polyak-{\L}ojasiewicz condition can
achieve $\mathcal{O}\big(\frac{\sqrt{p}}{n\epsilon}\big)$, which is the first
$\mathcal{O}\left(1/n\right)$ high probability excess population risk bound
w.r.t $n$ for differentially private algorithms under non-smooth conditions.
Moreover, we evaluate the performance of the new proposed algorithm m-NGP, the
experimental results show that m-NGP improves the performance of the
differentially private model over real datasets. It demonstrates that m-NGP
improves the utility bound and the accuracy of the DP model on real datasets
simultaneously.
|
107,338 | 107,338 |
Energy-efficient and Privacy-aware Social Distance Monitoring with
Low-resolution Infrared Sensors and Adaptive Inference
|
Low-resolution infrared (IR) Sensors combined with machine learning (ML) can
be leveraged to implement privacy-preserving social distance monitoring
solutions in indoor spaces. However, the need of executing these applications
on Internet of Things (IoT) edge nodes makes energy consumption critical. In
this work, we propose an energy-efficient adaptive inference solution
consisting of the cascade of a simple wake-up trigger and a 8-bit quantized
Convolutional Neural Network (CNN), which is only invoked for
difficult-to-classify frames. Deploying such adaptive system on a IoT
Microcontroller, we show that, when processing the output of a 8x8
low-resolution IR sensor, we are able to reduce the energy consumption by
37-57% with respect to a static CNN-based approach, with an accuracy drop of
less than 2% (83% balanced accuracy).
|
107,339 | 107,339 |
Privacy-preserving Social Distance Monitoring on Microcontrollers with
Low-Resolution Infrared Sensors and CNNs
|
Low-resolution infrared (IR) array sensors offer a low-cost, low-power, and
privacy-preserving alternative to optical cameras and smartphones/wearables for
social distance monitoring in indoor spaces, permitting the recognition of
basic shapes, without revealing the personal details of individuals. In this
work, we demonstrate that an accurate detection of social distance violations
can be achieved processing the raw output of a 8x8 IR array sensor with a
small-sized Convolutional Neural Network (CNN). Furthermore, the CNN can be
executed directly on a Microcontroller (MCU)-based sensor node.
With results on a newly collected open dataset, we show that our best CNN
achieves 86.3% balanced accuracy, significantly outperforming the 61% achieved
by a state-of-the-art deterministic algorithm. Changing the architectural
parameters of the CNN, we obtain a rich Pareto set of models, spanning
70.5-86.3% accuracy and 0.18-75k parameters. Deployed on a STM32L476RG MCU,
these models have a latency of 0.73-5.33ms, with an energy consumption per
inference of 9.38-68.57{\mu}J.
|
107,340 | 107,340 |
Depth Pruning with Auxiliary Networks for TinyML
|
Pruning is a neural network optimization technique that sacrifices accuracy
in exchange for lower computational requirements. Pruning has been useful when
working with extremely constrained environments in tinyML. Unfortunately,
special hardware requirements and limited study on its effectiveness on already
compact models prevent its wider adoption. Depth pruning is a form of pruning
that requires no specialized hardware but suffers from a large accuracy
falloff. To improve this, we propose a modification that utilizes a highly
efficient auxiliary network as an effective interpreter of intermediate feature
maps. Our results show a parameter reduction of 93% on the MLPerfTiny Visual
Wakewords (VWW) task and 28% on the Keyword Spotting (KWS) task with accuracy
cost of 0.65% and 1.06% respectively. When evaluated on a Cortex-M0
microcontroller, our proposed method reduces the VWW model size by 4.7x and
latency by 1.6x while counter intuitively gaining 1% accuracy. KWS model size
on Cortex-M0 was also reduced by 1.2x and latency by 1.2x at the cost of 2.21%
accuracy.
|
107,341 | 107,341 |
Sparse and Dense Approaches for the Full-rank Retrieval of Responses for
Dialogues
|
Ranking responses for a given dialogue context is a popular benchmark in
which the setup is to re-rank the ground-truth response over a limited set of
$n$ responses, where $n$ is typically 10. The predominance of this setup in
conversation response ranking has lead to a great deal of attention to building
neural re-rankers, while the first-stage retrieval step has been overlooked.
Since the correct answer is always available in the candidate list of $n$
responses, this artificial evaluation setup assumes that there is a first-stage
retrieval step which is always able to rank the correct response in its top-$n$
list. In this paper we focus on the more realistic task of full-rank retrieval
of responses, where $n$ can be up to millions of responses. We investigate both
dialogue context and response expansion techniques for sparse retrieval, as
well as zero-shot and fine-tuned dense retrieval approaches. Our findings based
on three different information-seeking dialogue datasets reveal that a learned
response expansion technique is a solid baseline for sparse retrieval. We find
the best performing method overall to be dense retrieval with intermediate
training, i.e. a step after the language model pre-training where sentence
representations are learned, followed by fine-tuning on the target
conversational data. We also investigate the intriguing phenomena that harder
negatives sampling techniques lead to worse results for the fine-tuned dense
retrieval models. The code and datasets are available at
https://github.com/Guzpenha/transformer_rankers/tree/full_rank_retrieval_dialogues.
|
107,342 | 107,342 |
A piece-wise constant approximation for non-conjugate Gaussian Process
models
|
Gaussian Processes (GPs) are a versatile and popular method in Bayesian
Machine Learning. A common modification are Sparse Variational Gaussian
Processes (SVGPs) which are well suited to deal with large datasets. While GPs
allow to elegantly deal with Gaussian-distributed target variables in closed
form, their applicability can be extended to non-Gaussian data as well. These
extensions are usually impossible to treat in closed form and hence require
approximate solutions. This paper proposes to approximate the inverse-link
function, which is necessary when working with non-Gaussian likelihoods, by a
piece-wise constant function. It will be shown that this yields a closed form
solution for the corresponding SVGP lower bound. In addition, it is
demonstrated how the piece-wise constant function itself can be optimized,
resulting in an inverse-link function that can be learnt from the data at hand.
|
107,343 | 107,343 |
Lossy compression of matrices by black-box optimisation of mixed-integer
non-linear programming
|
In edge computing, suppressing data size is a challenge for machine learning
models that perform complex tasks such as autonomous driving, in which
computational resources (speed, memory size and power) are limited. Efficient
lossy compression of matrix data has been introduced by decomposing it into the
product of an integer and real matrices. However, its optimisation is difficult
as it requires simultaneous optimisation of an integer and real variables. In
this paper, we improve this optimisation by utilising recently developed
black-box optimisation (BBO) algorithms with an Ising solver for integer
variables. In addition, the algorithm can be used to solve mixed-integer
programming problems that are linear and non-linear in terms of real and
integer variables, respectively. The differences between the choice of Ising
solvers (simulated annealing (SA), quantum annealing (QA) and simulated
quenching (SQ)) and the strategies of the BBO algorithms (BOCS, FMQA and their
variations) are discussed for further development of the BBO techniques.
|
107,344 | 107,344 |
FAIR4Cov: Fused Audio Instance and Representation for COVID-19 Detection
|
Audio-based classification techniques on body sounds have long been studied
to support diagnostic decisions, particularly in pulmonary diseases. In
response to the urgency of the COVID-19 pandemic, a growing number of models
are developed to identify COVID-19 patients based on acoustic input. Most
models focus on cough because the dry cough is the best-known symptom of
COVID-19. However, other body sounds, such as breath and speech, have also been
revealed to correlate with COVID-19 as well. In this work, rather than relying
on a specific body sound, we propose Fused Audio Instance and Representation
for COVID-19 Detection (FAIR4Cov). It relies on constructing a joint feature
vector obtained from a plurality of body sounds in waveform and spectrogram
representation. The core component of FAIR4Cov is a self-attention fusion unit
that is trained to establish the relation of multiple body sounds and audio
representations and integrate it into a compact feature vector. We set up our
experiments on different combinations of body sounds using only waveform,
spectrogram, and a joint representation of waveform and spectrogram. Our
findings show that the use of self-attention to combine extracted features from
cough, breath, and speech sounds leads to the best performance with an Area
Under the Receiver Operating Characteristic Curve (AUC) score of 0.8658, a
sensitivity of 0.8057, and a specificity of 0.7958. This AUC is 0.0227 higher
than the one of the models trained on spectrograms only and 0.0847 higher than
the one of the models trained on waveforms only. The results demonstrate that
the combination of spectrogram with waveform representation helps to enrich the
extracted features and outperforms the models with single representation.
|
107,345 | 107,345 |
Emergent Communication for Understanding Human Language Evolution:
What's Missing?
|
Emergent communication protocols among humans and artificial neural network
agents do not yet share the same properties and show some critical mismatches
in results. We describe three important phenomena with respect to the emergence
and benefits of compositionality: ease-of-learning, generalization, and group
size effects (i.e., larger groups create more systematic languages). The latter
two are not fully replicated with neural agents, which hinders the use of
neural emergent communication for language evolution research. We argue that
one possible reason for these mismatches is that key cognitive and
communicative constraints of humans are not yet integrated. Specifically, in
humans, memory constraints and the alternation between the roles of speaker and
listener underlie the emergence of linguistic structure, yet these constraints
are typically absent in neural simulations. We suggest that introducing such
communicative and cognitive constraints would promote more linguistically
plausible behaviors with neural agents.
|
107,346 | 107,346 |
Spacing Loss for Discovering Novel Categories
|
Novel Class Discovery (NCD) is a learning paradigm, where a machine learning
model is tasked to semantically group instances from unlabeled data, by
utilizing labeled instances from a disjoint set of classes. In this work, we
first characterize existing NCD approaches into single-stage and two-stage
methods based on whether they require access to labeled and unlabeled data
together while discovering new classes. Next, we devise a simple yet powerful
loss function that enforces separability in the latent space using cues from
multi-dimensional scaling, which we refer to as Spacing Loss. Our proposed
formulation can either operate as a standalone method or can be plugged into
existing methods to enhance them. We validate the efficacy of Spacing Loss with
thorough experimental evaluation across multiple settings on CIFAR-10 and
CIFAR-100 datasets.
|
107,347 | 107,347 |
Balancing Expert Utilization in Mixture-of-Experts Layers Embedded in
CNNs
|
This work addresses the problem of unbalanced expert utilization in
sparsely-gated Mixture of Expert (MoE) layers, embedded directly into
convolutional neural networks. To enable a stable training process, we present
both soft and hard constraint-based approaches. With hard constraints, the
weights of certain experts are allowed to become zero, while soft constraints
balance the contribution of experts with an additional auxiliary loss. As a
result, soft constraints handle expert utilization better and support the
expert specialization process, hard constraints mostly maintain generalized
experts and increase the model performance for many applications. Our findings
demonstrate that even with a single dataset and end-to-end training, experts
can implicitly focus on individual sub-domains of the input space. Experts in
the proposed models with MoE embeddings implicitly focus on distinct domains,
even without suitable predefined datasets. As an example, experts trained for
CIFAR-100 image classification specialize in recognizing different domains such
as sea animals or flowers without previous data clustering. Experiments with
RetinaNet and the COCO dataset further indicate that object detection experts
can also specialize in detecting objects of distinct sizes.
|
107,348 | 107,348 |
Federated Learning via Inexact ADMM
|
One of the crucial issues in federated learning is how to develop efficient
optimization algorithms. Most of the current ones require full devices
participation and/or impose strong assumptions for convergence. Different from
the widely-used gradient descent-based algorithms, this paper develops an
inexact alternating direction method of multipliers (ADMM), which is both
computation and communication-efficient, capable of combating the stragglers'
effect, and convergent under mild conditions.
|
107,349 | 107,349 |
Modelling graph dynamics in fraud detection with "Attention"
|
At online retail platforms, detecting fraudulent accounts and transactions is
crucial to improve customer experience, minimize loss, and avoid unauthorized
transactions. Despite the variety of different models for deep learning on
graphs, few approaches have been proposed for dealing with graphs that are both
heterogeneous and dynamic. In this paper, we propose DyHGN (Dynamic
Heterogeneous Graph Neural Network) and its variants to capture both temporal
and heterogeneous information. We first construct dynamic heterogeneous graphs
from registration and transaction data from eBay. Then, we build models with
diachronic entity embedding and heterogeneous graph transformer. We also use
model explainability techniques to understand the behaviors of DyHGN-* models.
Our findings reveal that modelling graph dynamics with heterogeneous inputs
need to be conducted with "attention" depending on the data structure,
distribution, and computation cost.
|
107,350 | 107,350 |
MEKER: Memory Efficient Knowledge Embedding Representation for Link
Prediction and Question Answering
|
Knowledge Graphs (KGs) are symbolically structured storages of facts. The KG
embedding contains concise data used in NLP tasks requiring implicit
information about the real world. Furthermore, the size of KGs that may be
useful in actual NLP assignments is enormous, and creating embedding over it
has memory cost issues. We represent KG as a 3rd-order binary tensor and move
beyond the standard CP decomposition by using a data-specific generalized
version of it. The generalization of the standard CP-ALS algorithm allows
obtaining optimization gradients without a backpropagation mechanism. It
reduces the memory needed in training while providing computational benefits.
We propose a MEKER, a memory-efficient KG embedding model, which yields
SOTA-comparable performance on link prediction tasks and KG-based Question
Answering.
|
107,351 | 107,351 |
Log-based Sparse Nonnegative Matrix Factorization for Data
Representation
|
Nonnegative matrix factorization (NMF) has been widely studied in recent
years due to its effectiveness in representing nonnegative data with
parts-based representations. For NMF, a sparser solution implies better
parts-based representation.However, current NMF methods do not always generate
sparse solutions.In this paper, we propose a new NMF method with log-norm
imposed on the factor matrices to enhance the sparseness.Moreover, we propose a
novel column-wisely sparse norm, named $\ell_{2,\log}$-(pseudo) norm to enhance
the robustness of the proposed method.The $\ell_{2,\log}$-(pseudo) norm is
invariant, continuous, and differentiable.For the $\ell_{2,\log}$ regularized
shrinkage problem, we derive a closed-form solution, which can be used for
other general problems.Efficient multiplicative updating rules are developed
for the optimization, which theoretically guarantees the convergence of the
objective value sequence.Extensive experimental results confirm the
effectiveness of the proposed method, as well as the enhanced sparseness and
robustness.
|
107,352 | 107,352 |
3D pride without 2D prejudice: Bias-controlled multi-level generative
models for structure-based ligand design
|
Generative models for structure-based molecular design hold significant
promise for drug discovery, with the potential to speed up the hit-to-lead
development cycle, while improving the quality of drug candidates and reducing
costs. Data sparsity and bias are, however, two main roadblocks to the
development of 3D-aware models. Here we propose a first-in-kind training
protocol based on multi-level contrastive learning for improved bias control
and data efficiency. The framework leverages the large data resources available
for 2D generative modelling with datasets of ligand-protein complexes. The
result are hierarchical generative models that are topologically unbiased,
explainable and customizable. We show how, by deconvolving the generative
posterior into chemical, topological and structural context factors, we not
only avoid common pitfalls in the design and evaluation of generative models,
but furthermore gain detailed insight into the generative process itself. This
improved transparency significantly aids method development, besides allowing
fine-grained control over novelty vs familiarity.
|
107,353 | 107,353 |
Paramixer: Parameterizing Mixing Links in Sparse Factors Works Better
than Dot-Product Self-Attention
|
Self-Attention is a widely used building block in neural modeling to mix
long-range data elements. Most self-attention neural networks employ pairwise
dot-products to specify the attention coefficients. However, these methods
require $O(N^2)$ computing cost for sequence length $N$. Even though some
approximation methods have been introduced to relieve the quadratic cost, the
performance of the dot-product approach is still bottlenecked by the low-rank
constraint in the attention matrix factorization. In this paper, we propose a
novel scalable and effective mixing building block called Paramixer. Our method
factorizes the interaction matrix into several sparse matrices, where we
parameterize the non-zero entries by MLPs with the data elements as input. The
overall computing cost of the new building block is as low as $O(N \log N)$.
Moreover, all factorizing matrices in Paramixer are full-rank, so it does not
suffer from the low-rank bottleneck. We have tested the new method on both
synthetic and various real-world long sequential data sets and compared it with
several state-of-the-art attention networks. The experimental results show that
Paramixer has better performance in most learning tasks.
|
107,354 | 107,354 |
Generative De Novo Protein Design with Global Context
|
The linear sequence of amino acids determines protein structure and function.
Protein design, known as the inverse of protein structure prediction, aims to
obtain a novel protein sequence that will fold into the defined structure.
Recent works on computational protein design have studied designing sequences
for the desired backbone structure with local positional information and
achieved competitive performance. However, similar local environments in
different backbone structures may result in different amino acids, indicating
that protein structure's global context matters. Thus, we propose the
Global-Context Aware generative de novo protein design method (GCA), consisting
of local and global modules. While local modules focus on relationships between
neighbor amino acids, global modules explicitly capture non-local contexts.
Experimental results demonstrate that the proposed GCA method outperforms
state-of-the-arts on de novo protein design. Our code and pretrained model will
be released.
|
107,355 | 107,355 |
Unknown Face Presentation Attack Detection via Localised Learning of
Multiple Kernels
|
The paper studies face spoofing, a.k.a. presentation attack detection (PAD)
in the demanding scenarios of unknown types of attack. While earlier studies
have revealed the benefits of ensemble methods, and in particular, a multiple
kernel learning approach to the problem, one limitation of such techniques is
that they typically treat the entire observation space similarly and ignore any
variability and local structure inherent to the data. This work studies this
aspect of the face presentation attack detection problem in relation to
multiple kernel learning in a one-class setting to benefit from intrinsic local
structure in bona fide face samples. More concretely, inspired by the success
of the one-class Fisher null formalism, we formulate a convex localised
multiple kernel learning algorithm by imposing a joint matrix-norm constraint
on the collection of local kernel weights and infer locally adaptive weights
for zero-shot one-class unseen attack detection.
We present a theoretical study of the proposed localised MKL algorithm using
Rademacher complexities to characterise its generalisation capability and
demonstrate the advantages of the proposed technique over some other options.
An assessment of the proposed approach on general object image datasets
illustrates its efficacy for abnormality and novelty detection while the
results of the experiments on face PAD datasets verifies its potential in
detecting unknown/unseen face presentation attacks.
|
107,356 | 107,356 |
TASAC: a twin-actor reinforcement learning framework with stochastic
policy for batch process control
|
Due to their complex nonlinear dynamics and batch-to-batch variability, batch
processes pose a challenge for process control. Due to the absence of accurate
models and resulting plant-model mismatch, these problems become harder to
address for advanced model-based control strategies. Reinforcement Learning
(RL), wherein an agent learns the policy by directly interacting with the
environment, offers a potential alternative in this context. RL frameworks with
actor-critic architecture have recently become popular for controlling systems
where state and action spaces are continuous. It has been shown that an
ensemble of actor and critic networks further helps the agent learn better
policies due to the enhanced exploration due to simultaneous policy learning.
To this end, the current study proposes a stochastic actor-critic RL algorithm,
termed Twin Actor Soft Actor-Critic (TASAC), by incorporating an ensemble of
actors for learning, in a maximum entropy framework, for batch process control.
|
107,357 | 107,357 |
Universum-inspired Supervised Contrastive Learning
|
Mixup is an efficient data augmentation method which generates additional
samples through respective convex combinations of original data points and
labels. Although being theoretically dependent on data properties, Mixup is
reported to perform well as a regularizer and calibrator contributing reliable
robustness and generalization to neural network training. In this paper,
inspired by Universum Learning which uses out-of-class samples to assist the
target tasks, we investigate Mixup from a largely under-explored perspective -
the potential to generate in-domain samples that belong to none of the target
classes, that is, universum. We find that in the framework of supervised
contrastive learning, universum-style Mixup produces surprisingly high-quality
hard negatives, greatly relieving the need for a large batch size in
contrastive learning. With these findings, we propose Universum-inspired
Contrastive learning (UniCon), which incorporates Mixup strategy to generate
universum data as g-negatives and pushes them apart from anchor samples of the
target classes. Our approach not only improves Mixup with hard labels, but also
innovates a novel measure to generate universum data. With a linear classifier
on the learned representations, our method achieves 81.68% top-1 accuracy on
CIFAR-100, surpassing the state of art by a significant margin of 5% with a
much smaller batch size, typically, 256 in UniCon vs. 1024 in SupCon using
ResNet-50.
|
107,358 | 107,358 |
Quantum Semi-Supervised Kernel Learning
|
Quantum computing leverages quantum effects to build algorithms that are
faster then their classical variants. In machine learning, for a given model
architecture, the speed of training the model is typically determined by the
size of the training dataset. Thus, quantum machine learning methods have the
potential to facilitate learning using extremely large datasets. While the
availability of data for training machine learning models is steadily
increasing, oftentimes it is much easier to collect feature vectors that to
obtain the corresponding labels. One of the approaches for addressing this
issue is to use semi-supervised learning, which leverages not only the labeled
samples, but also unlabeled feature vectors. Here, we present a quantum machine
learning algorithm for training Semi-Supervised Kernel Support Vector Machines.
The algorithm uses recent advances in quantum sample-based Hamiltonian
simulation to extend the existing Quantum LS-SVM algorithm to handle the
semi-supervised term in the loss. Through a theoretical study of the
algorithm's computational complexity, we show that it maintains the same
speedup as the fully-supervised Quantum LS-SVM.
|
107,359 | 107,359 |
EmbedTrack -- Simultaneous Cell Segmentation and Tracking Through
Learning Offsets and Clustering Bandwidths
|
A systematic analysis of the cell behavior requires automated approaches for
cell segmentation and tracking. While deep learning has been successfully
applied for the task of cell segmentation, there are few approaches for
simultaneous cell segmentation and tracking using deep learning. Here, we
present EmbedTrack, a single convolutional neural network for simultaneous cell
segmentation and tracking which predicts easy to interpret embeddings. As
embeddings, offsets of cell pixels to their cell center and bandwidths are
learned. We benchmark our approach on nine 2D data sets from the Cell Tracking
Challenge, where our approach performs on seven out of nine data sets within
the top 3 contestants including three top 1 performances. The source code is
publicly available at https://git.scc.kit.edu/kit-loe-ge/embedtrack.
|
107,360 | 107,360 |
Hierarchical Label-wise Attention Transformer Model for Explainable ICD
Coding
|
International Classification of Diseases (ICD) coding plays an important role
in systematically classifying morbidity and mortality data. In this study, we
propose a hierarchical label-wise attention Transformer model (HiLAT) for the
explainable prediction of ICD codes from clinical documents. HiLAT firstly
fine-tunes a pretrained Transformer model to represent the tokens of clinical
documents. We subsequently employ a two-level hierarchical label-wise attention
mechanism that creates label-specific document representations. These
representations are in turn used by a feed-forward neural network to predict
whether a specific ICD code is assigned to the input clinical document of
interest. We evaluate HiLAT using hospital discharge summaries and their
corresponding ICD-9 codes from the MIMIC-III database. To investigate the
performance of different types of Transformer models, we develop
ClinicalplusXLNet, which conducts continual pretraining from XLNet-Base using
all the MIMIC-III clinical notes. The experiment results show that the F1
scores of the HiLAT+ClinicalplusXLNet outperform the previous state-of-the-art
models for the top-50 most frequent ICD-9 codes from MIMIC-III. Visualisations
of attention weights present a potential explainability tool for checking the
face validity of ICD code predictions.
|
107,361 | 107,361 |
E2E Segmenter: Joint Segmenting and Decoding for Long-Form ASR
|
Improving the performance of end-to-end ASR models on long utterances ranging
from minutes to hours in length is an ongoing challenge in speech recognition.
A common solution is to segment the audio in advance using a separate voice
activity detector (VAD) that decides segment boundary locations based purely on
acoustic speech/non-speech information. VAD segmenters, however, may be
sub-optimal for real-world speech where, e.g., a complete sentence that should
be taken as a whole may contain hesitations in the middle ("set an alarm for...
5 o'clock").
We propose to replace the VAD with an end-to-end ASR model capable of
predicting segment boundaries in a streaming fashion, allowing the segmentation
decision to be conditioned not only on better acoustic features but also on
semantic features from the decoded text with negligible extra computation. In
experiments on real world long-form audio (YouTube) with lengths of up to 30
minutes, we demonstrate 8.5% relative WER improvement and 250 ms reduction in
median end-of-segment latency compared to the VAD segmenter baseline on a
state-of-the-art Conformer RNN-T model.
|
107,362 | 107,362 |
The Boltzmann Policy Distribution: Accounting for Systematic
Suboptimality in Human Models
|
Models of human behavior for prediction and collaboration tend to fall into
two categories: ones that learn from large amounts of data via imitation
learning, and ones that assume human behavior to be noisily-optimal for some
reward function. The former are very useful, but only when it is possible to
gather a lot of human data in the target environment and distribution. The
advantage of the latter type, which includes Boltzmann rationality, is the
ability to make accurate predictions in new environments without extensive data
when humans are actually close to optimal. However, these models fail when
humans exhibit systematic suboptimality, i.e. when their deviations from
optimal behavior are not independent, but instead consistent over time. Our key
insight is that systematic suboptimality can be modeled by predicting policies,
which couple action choices over time, instead of trajectories. We introduce
the Boltzmann policy distribution (BPD), which serves as a prior over human
policies and adapts via Bayesian inference to capture systematic deviations by
observing human actions during a single episode. The BPD is difficult to
compute and represent because policies lie in a high-dimensional continuous
space, but we leverage tools from generative and sequence models to enable
efficient sampling and inference. We show that the BPD enables prediction of
human behavior and human-AI collaboration equally as well as imitation
learning-based human models while using far less data.
|
107,363 | 107,363 |
Constructing dynamic residential energy lifestyles using Latent
Dirichlet Allocation
|
The rapid expansion of Advanced Meter Infrastructure (AMI) has dramatically
altered the energy information landscape. However, our ability to use this
information to generate actionable insights about residential electricity
demand remains limited. In this research, we propose and test a new framework
for understanding residential electricity demand by using a dynamic energy
lifestyles approach that is iterative and highly extensible. To obtain energy
lifestyles, we develop a novel approach that applies Latent Dirichlet
Allocation (LDA), a method commonly used for inferring the latent topical
structure of text data, to extract a series of latent household energy
attributes. By doing so, we provide a new perspective on household electricity
consumption where each household is characterized by a mixture of energy
attributes that form the building blocks for identifying a sparse collection of
energy lifestyles. We examine this approach by running experiments on one year
of hourly smart meter data from 60,000 households and we extract six energy
attributes that describe general daily use patterns. We then use clustering
techniques to derive six distinct energy lifestyle profiles from energy
attribute proportions. Our lifestyle approach is also flexible to varying time
interval lengths, and we test our lifestyle approach seasonally (Autumn,
Winter, Spring, and Summer) to track energy lifestyle dynamics within and
across households and find that around 73% of households manifest multiple
lifestyles across a year. These energy lifestyles are then compared to
different energy use characteristics, and we discuss their practical
applications for demand response program design and lifestyle change analysis.
|
107,364 | 107,364 |
Denoising of Three-Dimensional Fast Spin Echo Magnetic Resonance Images
of Knee Joints using Spatial-Variant Noise-Relevant Residual Learning of
Convolution Neural Network
|
Two-dimensional (2D) fast spin echo (FSE) techniques play a central role in
the clinical magnetic resonance imaging (MRI) of knee joints. Moreover,
three-dimensional (3D) FSE provides high-isotropic-resolution magnetic
resonance (MR) images of knee joints, but it has a reduced signal-to-noise
ratio compared to 2D FSE. Deep-learning denoising methods are a promising
approach for denoising MR images, but they are often trained using synthetic
noise due to challenges in obtaining true noise distributions for MR images. In
this study, inherent true noise information from 2-NEX acquisition was used to
develop a deep-learning model based on residual learning of convolutional
neural network (CNN), and this model was used to suppress the noise in 3D FSE
MR images of knee joints. The proposed CNN used two-step residual learning over
parallel transporting and residual blocks and was designed to comprehensively
learn real noise features from 2-NEX training data. The results of an ablation
study validated the network design. The new method achieved improved denoising
performance of 3D FSE knee MR images compared with current state-of-the-art
methods, based on the peak signal-to-noise ratio and structural similarity
index measure. The improved image quality after denoising using the new method
was verified by radiological evaluation. A deep CNN using the inherent
spatial-varying noise information in 2-NEX acquisitions was developed. This
method showed promise for clinical MRI assessments of the knee, and has
potential applications for the assessment of other anatomical structures.
|
107,365 | 107,365 |
ParkPredict+: Multimodal Intent and Motion Prediction for Vehicles in
Parking Lots with CNN and Transformer
|
The problem of multimodal intent and trajectory prediction for human-driven
vehicles in parking lots is addressed in this paper. Using models designed with
CNN and Transformer networks, we extract temporal-spatial and contextual
information from trajectory history and local bird's eye view (BEV) semantic
images, and generate predictions about intent distribution and future
trajectory sequences. Our methods outperforms existing models in accuracy,
while allowing an arbitrary number of modes, encoding complex multi-agent
scenarios, and adapting to different parking maps. In addition, we present the
first public human driving dataset in parking lot with high resolution and rich
traffic scenarios for relevant research in this field.
|
107,366 | 107,366 |
Centralized Adversarial Learning for Robust Deep Hashing
|
Deep hashing has been extensively utilized in massive image retrieval because
of its efficiency and effectiveness. Recently, it becomes a hot issue to study
adversarial examples which poses a security challenge to deep hashing models.
However, there is still a critical bottleneck: how to find a superior and exact
semantic representative as the guide to further enhance the adversarial attack
and defense in deep hashing based retrieval. We, for the first time, attempt to
design an effective adversarial learning with the min-max paradigm to improve
the robustness of hashing networks by using the generated adversarial samples.
Specifically, we obtain the optimal solution (called center code) through a
proved Continuous Hash Center Method (CHCM), which preserves the semantic
similarity with positive samples and dissimilarity with negative samples. On
one hand, we propose the Deep Hashing Central Attack (DHCA) for efficient
attack on hashing retrieval by maximizing the Hamming distance between the hash
code of adversarial example and the center code. On the other hand, we present
the Deep Hashing Central Adversarial Training (DHCAT) to optimize the hashing
networks for defense, by minimizing the Hamming distance to the center code.
Extensive experiments on the benchmark datasets verify that our attack method
can achieve better performance than the state-of-the-arts, and our defense
algorithm can effectively mitigate the effects of adversarial perturbations.
|
107,367 | 107,367 |
On Feature Learning in Neural Networks with Global Convergence
Guarantees
|
We study the optimization of wide neural networks (NNs) via gradient flow
(GF) in setups that allow feature learning while admitting non-asymptotic
global convergence guarantees. First, for wide shallow NNs under the mean-field
scaling and with a general class of activation functions, we prove that when
the input dimension is no less than the size of the training set, the training
loss converges to zero at a linear rate under GF. Building upon this analysis,
we study a model of wide multi-layer NNs whose second-to-last layer is trained
via GF, for which we also prove a linear-rate convergence of the training loss
to zero, but regardless of the input dimension. We also show empirically that,
unlike in the Neural Tangent Kernel (NTK) regime, our multi-layer model
exhibits feature learning and can achieve better generalization performance
than its NTK counterpart.
|
107,368 | 107,368 |
MNL-Bandits under Inventory and Limited Switches Constraints
|
Optimizing the assortment of products to display to customers is a key to
increasing revenue for both offline and online retailers. To trade-off between
exploring customers' preference and exploiting customers' choices learned from
data, in this paper, by adopting the Multi-Nomial Logit (MNL) choice model to
capture customers' choices over products, we study the problem of optimizing
assortments over a planning horizon $T$ for maximizing the profit of the
retailer. To make the problem setting more practical, we consider both the
inventory constraint and the limited switches constraint, where the retailer
cannot use up the resource inventory before time $T$ and is forbidden to switch
the assortment shown to customers too many times. Such a setting suits the case
when an online retailer wants to dynamically optimize the assortment selection
for a population of customers. We develop an efficient UCB-like algorithm to
optimize the assortments while learning customers' choices from data. We prove
that our algorithm can achieve a sub-linear regret bound
$\tilde{O}\left(T^{1-\alpha/2}\right)$ if $O(T^\alpha)$ switches are allowed.
%, and our regret bound is optimal with respect to $T$. Extensive numerical
experiments show that our algorithm outperforms baselines and the gap between
our algorithm's performance and the theoretical upper bound is small.
|
107,369 | 107,369 |
A Unifying Framework for Combining Complementary Strengths of Humans and
ML toward Better Predictive Decision-Making
|
Hybrid human-ML systems are increasingly in charge of consequential decisions
in a wide range of domains. A growing body of work has advanced our
understanding of these systems by providing empirical and theoretical analyses.
However, existing empirical results are mixed, and theoretical proposals are
often incompatible with each other. Our goal in this work is to bring
much-needed organization to this field by offering a unifying framework for
understanding conditions under which combining complementary strengths of human
and ML leads to higher quality decisions than those produced by them
individually -- a state to which we refer to as human-ML complementarity. We
focus specifically on the context of human-ML predictive decision-making
systems and investigate optimal ways of combining human and ML-based predictive
decisions, accounting for the underlying causes of variation in their
judgments. Within this scope, we present two crucial contributions. First,
drawing upon prior literature in human psychology, machine learning, and
human-computer interaction, we introduce a taxonomy characterizing a wide
variety of criteria across which human and machine decision-making differ.
Building on our taxonomy, our second contribution presents a unifying
optimization-based framework for formalizing how human and ML predictive
decisions should be aggregated optimally. We show that our proposed framework
encompasses several existing models of human-ML complementarity as special
cases. Last but not least, the exploratory analysis of our framework offers a
critical piece of insight for future work in this area: the mechanism by which
we combine human-ML judgments should be informed by the underlying causes of
their diverging decisions.
|
107,370 | 107,370 |
Predicting highway lane-changing maneuvers: A benchmark analysis of
machine and ensemble learning algorithms
|
Understanding and predicting highway lane-change maneuvers is essential for
driving modeling and its automation. The development of data-based
lane-changing decision-making algorithms is nowadays in full expansion. We
compare empirically in this article different machine and ensemble learning
classification techniques to the MOBIL rule-based model using trajectory data
of European two-lane highways. The analysis relies on instantaneous
measurements of up to twenty-four spatial-temporal variables with the four
neighboring vehicles on current and adjacent lanes. Preliminary descriptive
investigations by principal component and logistic analyses allow identifying
main variables intending a driver to change lanes. We predict two types of
discretionary lane-change maneuvers: Overtaking (from slow to fast lane) and
fold-down (from fast to slow lane). The prediction accuracy is quantified using
total, lane-changing and lane-keeping errors and associated receiver operating
characteristic curves. The benchmark analysis includes logistic model, linear
discriminant, decision tree, na\"ive Bayes classifier, support vector machine,
neural network machine learning algorithms, and up to ten bagging and stacking
ensemble learning meta-heuristics. If the rule-based model provides limited
predicting accuracy, especially in case of fold-down, the data-based
algorithms, devoid of modeling bias, allow significant prediction improvements.
Cross validations show that selected neural networks and stacking algorithms
allow predicting from a single observation both fold-down and overtaking
maneuvers up to four seconds in advance with high accuracy.
|
107,371 | 107,371 |
Learning to Scaffold: Optimizing Model Explanations for Teaching
|
Modern machine learning models are opaque, and as a result there is a
burgeoning academic subfield on methods that explain these models' behavior.
However, what is the precise goal of providing such explanations, and how can
we demonstrate that explanations achieve this goal? Some research argues that
explanations should help teach a student (either human or machine) to simulate
the model being explained, and that the quality of explanations can be measured
by the simulation accuracy of students on unexplained examples. In this work,
leveraging meta-learning techniques, we extend this idea to improve the quality
of the explanations themselves, specifically by optimizing explanations such
that student models more effectively learn to simulate the original model. We
train models on three natural language processing and computer vision tasks,
and find that students trained with explanations extracted with our framework
are able to simulate the teacher significantly more effectively than ones
produced with previous methods. Through human annotations and a user study, we
further find that these learned explanations more closely align with how humans
would explain the required decisions in these tasks. Our code is available at
https://github.com/coderpat/learning-scaffold
|
107,372 | 107,372 |
Reward Reports for Reinforcement Learning
|
The desire to build good systems in the face of complex societal effects
requires a dynamic approach towards equity and access. Recent approaches to
machine learning (ML) documentation have demonstrated the promise of discursive
frameworks for deliberation about these complexities. However, these
developments have been grounded in a static ML paradigm, leaving the role of
feedback and post-deployment performance unexamined. Meanwhile, recent work in
reinforcement learning design has shown that the effects of optimization
objectives on the resultant system behavior can be wide-ranging and
unpredictable. In this paper we sketch a framework for documenting deployed
learning systems, which we call Reward Reports. Taking inspiration from various
contributions to the technical literature on reinforcement learning, we outline
Reward Reports as living documents that track updates to design choices and
assumptions behind what a particular automated system is optimizing for. They
are intended to track dynamic phenomena arising from system deployment, rather
than merely static properties of models or data. After presenting the elements
of a Reward Report, we provide three examples: DeepMind's MuZero, MovieLens,
and a hypothetical deployment of a Project Flow traffic control policy.
|
107,373 | 107,373 |
Convergence of the Riemannian Langevin Algorithm
|
We study the Riemannian Langevin Algorithm for the problem of sampling from a
distribution with density $\nu$ with respect to the natural measure on a
manifold with metric $g$. We assume that the target density satisfies a
log-Sobolev inequality with respect to the metric and prove that the manifold
generalization of the Unadjusted Langevin Algorithm converges rapidly to $\nu$
for Hessian manifolds. This allows us to reduce the problem of sampling
non-smooth (constrained) densities in ${\bf R}^n$ to sampling smooth densities
over appropriate manifolds, while needing access only to the gradient of the
log-density, and this, in turn, to sampling from the natural Brownian motion on
the manifold. Our main analytic tools are (1) an extension of self-concordance
to manifolds, and (2) a stochastic approach to bounding smoothness on
manifolds. A special case of our approach is sampling isoperimetric densities
restricted to polytopes by using the metric defined by the logarithmic barrier.
|
107,374 | 107,374 |
Memory Bounds for Continual Learning
|
Continual learning, or lifelong learning, is a formidable current challenge
to machine learning. It requires the learner to solve a sequence of $k$
different learning tasks, one after the other, while retaining its aptitude for
earlier tasks; the continual learner should scale better than the obvious
solution of developing and maintaining a separate learner for each of the $k$
tasks. We embark on a complexity-theoretic study of continual learning in the
PAC framework. We make novel uses of communication complexity to establish that
any continual learner, even an improper one, needs memory that grows linearly
with $k$, strongly suggesting that the problem is intractable. When
logarithmically many passes over the learning tasks are allowed, we provide an
algorithm based on multiplicative weights update whose memory requirement
scales well; we also establish that improper learning is necessary for such
performance. We conjecture that these results may lead to new promising
approaches to continual learning.
|
107,375 | 107,375 |
Learning for Spatial Branching: An Algorithm Selection Approach
|
The use of machine learning techniques to improve the performance of
branch-and-bound optimization algorithms is a very active area in the context
of mixed integer linear problems, but little has been done for non-linear
optimization. To bridge this gap, we develop a learning framework for spatial
branching and show its efficacy in the context of the
Reformulation-Linearization Technique for polynomial optimization problems. The
proposed learning is performed offline, based on instance-specific features and
with no computational overhead when solving new instances. Novel graph-based
features are introduced, which turn out to play an important role for the
learning. Experiments on different benchmark instances from the literature show
that the learning-based branching rule significantly outperforms the standard
rules.
|
107,376 | 107,376 |
Federated Learning Enables Big Data for Rare Cancer Boundary Detection
|
Although machine learning (ML) has shown promise in numerous domains, there
are concerns about generalizability to out-of-sample data. This is currently
addressed by centrally sharing ample, and importantly diverse, data from
multiple sites. However, such centralization is challenging to scale (or even
not feasible) due to various limitations. Federated ML (FL) provides an
alternative to train accurate and generalizable ML models, by only sharing
numerical model updates. Here we present findings from the largest FL study
to-date, involving data from 71 healthcare institutions across 6 continents, to
generate an automatic tumor boundary detector for the rare disease of
glioblastoma, utilizing the largest dataset of such patients ever used in the
literature (25,256 MRI scans from 6,314 patients). We demonstrate a 33%
improvement over a publicly trained model to delineate the surgically
targetable tumor, and 23% improvement over the tumor's entire extent. We
anticipate our study to: 1) enable more studies in healthcare informed by large
and diverse data, ensuring meaningful results for rare diseases and
underrepresented populations, 2) facilitate further quantitative analyses for
glioblastoma via performance optimization of our consensus model for eventual
public release, and 3) demonstrate the effectiveness of FL at such scale and
task complexity as a paradigm shift for multi-site collaborations, alleviating
the need for data sharing.
|
107,377 | 107,377 |
How Sampling Impacts the Robustness of Stochastic Neural Networks
|
Stochastic neural networks (SNNs) are random functions and predictions are
gained by averaging over multiple realizations of this random function.
Consequently, an adversarial attack is calculated based on one set of samples
and applied to the prediction defined by another set of samples. In this paper
we analyze robustness in this setting by deriving a sufficient condition for
the given prediction process to be robust against the calculated attack. This
allows us to identify the factors that lead to an increased robustness of SNNs
and helps to explain the impact of the variance and the amount of samples.
Among other things, our theoretical analysis gives insights into (i) why
increasing the amount of samples drawn for the estimation of adversarial
examples increases the attack's strength, (ii) why decreasing sample size
during inference hardly influences the robustness, and (iii) why a higher
prediction variance between realizations relates to a higher robustness. We
verify the validity of our theoretical findings by an extensive empirical
analysis.
|
107,378 | 107,378 |
Detecting early signs of depression in the conversational domain: The
role of transfer learning in low-resource scenarios
|
The high prevalence of depression in society has given rise to the need for
new digital tools to assist in its early detection. To this end, existing
research has mainly focused on detecting depression in the domain of social
media, where there is a sufficient amount of data. However, with the rise of
conversational agents like Siri or Alexa, the conversational domain is becoming
more critical. Unfortunately, there is a lack of data in the conversational
domain. We perform a study focusing on domain adaptation from social media to
the conversational domain. Our approach mainly exploits the linguistic
information preserved in the vector representation of text. We describe
transfer learning techniques to classify users who suffer from early signs of
depression with high recall. We achieve state-of-the-art results on a commonly
used conversational dataset, and we highlight how the method can easily be used
in conversational agents. We publicly release all source code.
|
107,379 | 107,379 |
Learning-to-Rank at the Speed of Sampling: Plackett-Luce Gradient
Estimation With Minimal Computational Complexity
|
Plackett-Luce gradient estimation enables the optimization of stochastic
ranking models within feasible time constraints through sampling techniques.
Unfortunately, the computational complexity of existing methods does not scale
well with the length of the rankings, i.e. the ranking cutoff, nor with the
item collection size. In this paper, we introduce the novel PL-Rank-3 algorithm
that performs unbiased gradient estimation with a computational complexity
comparable to the best sorting algorithms. As a result, our novel
learning-to-rank method is applicable in any scenario where standard sorting is
feasible in reasonable time. Our experimental results indicate large gains in
the time required for optimization, without any loss in performance. For the
field, our contribution could potentially allow state-of-the-art
learning-to-rank methods to be applied to much larger scales than previously
feasible.
|
107,380 | 107,380 |
Compressibility: Power of PCA in Clustering Problems Beyond
Dimensionality Reduction
|
In this paper we take a step towards understanding the impact of principle
component analysis (PCA) in the context of unsupervised clustering beyond a
dimensionality reduction tool. We explore another property of PCA in vector
clustering problems, which we call compressibility. This phenomenon shows that
PCA significantly reduces the distance of data points belonging to the same
clusters, while reducing inter-cluster distances relatively mildly. This gap
explains many empirical observations found in practice. For example, in
single-cell RNA-sequencing analysis, which is an application of vector
clustering in biology, it has been observed that applying PCA on datasets
significantly improves the accuracy of classical clustering algorithms such as
K-means.
We study this compression gap in both theory and practice. On the theoretical
side, we analyze PCA in a fairly general probabilistic setup, which we call the
random vector model. In terms of practice, we verify the compressibility of PCA
on multiple single-cell RNA-seq datasets.
|
107,381 | 107,381 |
Generative sampling in tractography using autoencoders (GESTA)
|
Current tractography methods use the local orientation information to
propagate streamlines from seed locations. Many such seeds provide streamlines
that stop prematurely or fail to map the true pathways because some white
matter bundles are "harder-to-track" than others. This results in tractography
reconstructions with poor white and gray matter spatial coverage. In this work,
we propose a generative, autoencoder-based method, named GESTA (Generative
Sampling in Tractography using Autoencoders), that produces streamlines with
better spatial coverage. Compared to other deep learning methods, our
autoencoder-based framework is not constrained by any prior or a fixed set of
bundles. GESTA produces new and complete streamlines for any white matter
bundle. GESTA is shown to be effective on both synthetic and human brain in
vivo data. Our streamline evaluation framework ensures that the streamlines
produced by GESTA are anatomically plausible and fit well to the local
diffusion signal. The streamline evaluation criteria assess anatomy (white
matter coverage), local orientation alignment (direction), geometry features of
streamlines, and optionally, gray matter connectivity. The GESTA framework
offers considerable gains in bundle coverage using a reduced set of seeding
streamlines with a 1.5x improvement for the "Fiber Cup", and 6x for the ISMRM
2015 Tractography Challenge datasets. Similarly, it provides a 4x white matter
volume increase on the BIL&GIN callosal homotopic dataset. It also successfully
generates new streamlines in poorly populated bundles, such as the fornix and
other hard-to-track bundles, on in vivo data. GESTA is thus the first deep
tractography generative method that can improve white matter reconstruction of
hard-to-track bundles.
|
107,382 | 107,382 |
Comparative Study of Machine Learning Test Case Prioritization for
Continuous Integration Testing
|
There is a growing body of research indicating the potential of machine
learning to tackle complex software testing challenges. One such challenge
pertains to continuous integration testing, which is highly time-constrained,
and generates a large amount of data coming from iterative code commits and
test runs. In such a setting, we can use plentiful test data for training
machine learning predictors to identify test cases able to speed up the
detection of regression bugs introduced during code integration. However,
different machine learning models can have different fault prediction
performance depending on the context and the parameters of continuous
integration testing, for example variable time budget available for continuous
integration cycles, or the size of test execution history used for learning to
prioritize failing test cases. Existing studies on test case prioritization
rarely study both of these factors, which are essential for the continuous
integration practice. In this study we perform a comprehensive comparison of
the fault prediction performance of machine learning approaches that have shown
the best performance on test case prioritization tasks in the literature. We
evaluate the accuracy of the classifiers in predicting fault-detecting tests
for different values of the continuous integration time budget and with
different length of test history used for training the classifiers. In
evaluation, we use real-world industrial datasets from a continuous integration
practice. The results show that different machine learning models have
different performance for different size of test history used for model
training and for different time budget available for test case execution. Our
results imply that machine learning approaches for test prioritization in
continuous integration testing should be carefully configured to achieve
optimal performance.
|
107,383 | 107,383 |
Label a Herd in Minutes: Individual Holstein-Friesian Cattle
Identification
|
We describe a practically evaluated approach for training visual cattle ID
systems for a whole farm requiring only ten minutes of labelling effort. In
particular, for the task of automatic identification of individual
Holstein-Friesians in real-world farm CCTV, we show that self-supervision,
metric learning, cluster analysis, and active learning can complement each
other to significantly reduce the annotation requirements usually needed to
train cattle identification frameworks. Evaluating the approach on the test
portion of the publicly available Cows2021 dataset, for training we use 23,350
frames across 435 single individual tracklets generated by automated oriented
cattle detection and tracking in operational farm footage. Self-supervised
metric learning is first employed to initialise a candidate identity space
where each tracklet is considered a distinct entity. Grouping entities into
equivalence classes representing cattle identities is then performed by
automated merging via cluster analysis and active learning. Critically, we
identify the inflection point at which automated choices cannot replicate
improvements based on human intervention to reduce annotation to a minimum.
Experimental results show that cluster analysis and a few minutes of labelling
after automated self-supervision can improve the test identification accuracy
of 153 identities to 92.44% (ARI=0.93) from the 74.9% (ARI=0.754) obtained by
self-supervision only. These promising results indicate that a tailored
combination of human and machine reasoning in visual cattle ID pipelines can be
highly effective whilst requiring only minimal labelling effort. We provide all
key source code and network weights with this paper for easy result
reproduction.
|
107,384 | 107,384 |
Error-in-variables modelling for operator learning
|
Deep operator learning has emerged as a promising tool for reduced-order
modelling and PDE model discovery. Leveraging the expressive power of deep
neural networks, especially in high dimensions, such methods learn the mapping
between functional state variables. While proposed methods have assumed noise
only in the dependent variables, experimental and numerical data for operator
learning typically exhibit noise in the independent variables as well, since
both variables represent signals that are subject to measurement error. In
regression on scalar data, failure to account for noisy independent variables
can lead to biased parameter estimates. With noisy independent variables,
linear models fitted via ordinary least squares (OLS) will show attenuation
bias, wherein the slope will be underestimated. In this work, we derive an
analogue of attenuation bias for linear operator regression with white noise in
both the independent and dependent variables. In the nonlinear setting, we
computationally demonstrate underprediction of the action of the Burgers
operator in the presence of noise in the independent variable. We propose
error-in-variables (EiV) models for two operator regression methods,
MOR-Physics and DeepONet, and demonstrate that these new models reduce bias in
the presence of noisy independent variables for a variety of operator learning
problems. Considering the Burgers operator in 1D and 2D, we demonstrate that
EiV operator learning robustly recovers operators in high-noise regimes that
defeat OLS operator learning. We also introduce an EiV model for time-evolving
PDE discovery and show that OLS and EiV perform similarly in learning the
Kuramoto-Sivashinsky evolution operator from corrupted data, suggesting that
the effect of bias in OLS operator learning depends on the regularity of the
target operator.
|
107,385 | 107,385 |
Revealing Occlusions with 4D Neural Fields
|
For computer vision systems to operate in dynamic situations, they need to be
able to represent and reason about object permanence. We introduce a framework
for learning to estimate 4D visual representations from monocular RGB-D, which
is able to persist objects, even once they become obstructed by occlusions.
Unlike traditional video representations, we encode point clouds into a
continuous representation, which permits the model to attend across the
spatiotemporal context to resolve occlusions. On two large video datasets that
we release along with this paper, our experiments show that the representation
is able to successfully reveal occlusions for several tasks, without any
architectural changes. Visualizations show that the attention mechanism
automatically learns to follow occluded objects. Since our approach can be
trained end-to-end and is easily adaptable, we believe it will be useful for
handling occlusions in many video understanding tasks. Data, code, and models
are available at https://occlusions.cs.columbia.edu/.
|
107,386 | 107,386 |
SegDiscover: Visual Concept Discovery via Unsupervised Semantic
Segmentation
|
Visual concept discovery has long been deemed important to improve
interpretability of neural networks, because a bank of semantically meaningful
concepts would provide us with a starting point for building machine learning
models that exhibit intelligible reasoning process. Previous methods have
disadvantages: either they rely on labelled support sets that incorporate human
biases for objects that are "useful," or they fail to identify multiple
concepts that occur within a single image. We reframe the concept discovery
task as an unsupervised semantic segmentation problem, and present SegDiscover,
a novel framework that discovers semantically meaningful visual concepts from
imagery datasets with complex scenes without supervision. Our method contains
three important pieces: generating concept primitives from raw images,
discovering concepts by clustering in the latent space of a self-supervised
pretrained encoder, and concept refinement via neural network smoothing.
Experimental results provide evidence that our method can discover multiple
concepts within a single image and outperforms state-of-the-art unsupervised
methods on complex datasets such as Cityscapes and COCO-Stuff. Our method can
be further used as a neural network explanation tool by comparing results
obtained by different encoders.
|
107,387 | 107,387 |
A Tale of Two Models: Constructing Evasive Attacks on Edge Models
|
Full-precision deep learning models are typically too large or costly to
deploy on edge devices. To accommodate to the limited hardware resources,
models are adapted to the edge using various edge-adaptation techniques, such
as quantization and pruning. While such techniques may have a negligible impact
on top-line accuracy, the adapted models exhibit subtle differences in output
compared to the original model from which they are derived. In this paper, we
introduce a new evasive attack, DIVA, that exploits these differences in edge
adaptation, by adding adversarial noise to input data that maximizes the output
difference between the original and adapted model. Such an attack is
particularly dangerous, because the malicious input will trick the adapted
model running on the edge, but will be virtually undetectable by the original
model, which typically serves as the authoritative model version, used for
validation, debugging and retraining. We compare DIVA to a state-of-the-art
attack, PGD, and show that DIVA is only 1.7-3.6% worse on attacking the adapted
model but 1.9-4.2 times more likely not to be detected by the the original
model under a whitebox and semi-blackbox setting, compared to PGD.
|
107,388 | 107,388 |
Counterfactual Learning To Rank for Utility-Maximizing Query
Autocompletion
|
Conventional methods for query autocompletion aim to predict which completed
query a user will select from a list. A shortcoming of this approach is that
users often do not know which query will provide the best retrieval performance
on the current information retrieval system, meaning that any query
autocompletion methods trained to mimic user behavior can lead to suboptimal
query suggestions. To overcome this limitation, we propose a new approach that
explicitly optimizes the query suggestions for downstream retrieval
performance. We formulate this as a problem of ranking a set of rankings, where
each query suggestion is represented by the downstream item ranking it
produces. We then present a learning method that ranks query suggestions by the
quality of their item rankings. The algorithm is based on a counterfactual
learning approach that is able to leverage feedback on the items (e.g., clicks,
purchases) to evaluate query suggestions through an unbiased estimator, thus
avoiding the assumption that users write or select optimal queries. We
establish theoretical support for the proposed approach and provide
learning-theoretic guarantees. We also present empirical results on publicly
available datasets, and demonstrate real-world applicability using data from an
online shopping store.
|
107,389 | 107,389 |
A Multi-level Alignment Training Scheme for Video-and-Language Grounding
|
To solve video-and-language grounding tasks, the key is for the network to
understand the connection between the two modalities. For a pair of video and
language description, their semantic relation is reflected by their encodings'
similarity. A good multi-modality encoder should be able to well capture both
inputs' semantics and encode them in the shared feature space where embedding
distance gets properly translated into their semantic similarity. In this work,
we focused on this semantic connection between video and language, and
developed a multi-level alignment training scheme to directly shape the
encoding process. Global and segment levels of video-language alignment pairs
were designed, based on the information similarity ranging from high-level
context to fine-grained semantics. The contrastive loss was used to contrast
the encodings' similarities between the positive and negative alignment pairs,
and to ensure the network is trained in such a way that similar information is
encoded closely in the shared feature space while information of different
semantics is kept apart. Our multi-level alignment training can be applied to
various video-and-language grounding tasks. Together with the task-specific
training loss, our framework achieved comparable performance to previous
state-of-the-arts on multiple video QA and retrieval datasets.
|
107,390 | 107,390 |
Balancing Fairness and Accuracy in Sentiment Detection using Multiple
Black Box Models
|
Sentiment detection is an important building block for multiple information
retrieval tasks such as product recommendation, cyberbullying detection, and
misinformation detection. Unsurprisingly, multiple commercial APIs, each with
different levels of accuracy and fairness, are now available for sentiment
detection. While combining inputs from multiple modalities or black-box models
for increasing accuracy is commonly studied in multimedia computing literature,
there has been little work on combining different modalities for increasing
fairness of the resulting decision. In this work, we audit multiple commercial
sentiment detection APIs for the gender bias in two actor news headlines
settings and report on the level of bias observed. Next, we propose a "Flexible
Fair Regression" approach, which ensures satisfactory accuracy and fairness by
jointly learning from multiple black-box models. The results pave way for fair
yet accurate sentiment detectors for multiple applications.
|
107,391 | 107,391 |
Evaluation of Multi-Scale Multiple Instance Learning to Improve Thyroid
Cancer Classification
|
Thyroid cancer is currently the fifth most common malignancy diagnosed in
women. Since differentiation of cancer sub-types is important for treatment and
current, manual methods are time consuming and subjective, automatic
computer-aided differentiation of cancer types is crucial. Manual
differentiation of thyroid cancer is based on tissue sections, analysed by
pathologists using histological features. Due to the enormous size of gigapixel
whole slide images, holistic classification using deep learning methods is not
feasible. Patch based multiple instance learning approaches, combined with
aggregations such as bag-of-words, is a common approach. This work's
contribution is to extend a patch based state-of-the-art method by generating
and combining feature vectors of three different patch resolutions and
analysing three distinct ways of combining them. The results showed
improvements in one of the three multi-scale approaches, while the others led
to decreased scores. This provides motivation for analysis and discussion of
the individual approaches.
|
107,392 | 107,392 |
Visual Attention Emerges from Recurrent Sparse Reconstruction
|
Visual attention helps achieve robust perception under noise, corruption, and
distribution shifts in human vision, which are areas where modern neural
networks still fall short. We present VARS, Visual Attention from Recurrent
Sparse reconstruction, a new attention formulation built on two prominent
features of the human visual attention mechanism: recurrency and sparsity.
Related features are grouped together via recurrent connections between
neurons, with salient objects emerging via sparse regularization. VARS adopts
an attractor network with recurrent connections that converges toward a stable
pattern over time. Network layers are represented as ordinary differential
equations (ODEs), formulating attention as a recurrent attractor network that
equivalently optimizes the sparse reconstruction of input using a dictionary of
"templates" encoding underlying patterns of data. We show that self-attention
is a special case of VARS with a single-step optimization and no sparsity
constraint. VARS can be readily used as a replacement for self-attention in
popular vision transformers, consistently improving their robustness across
various benchmarks. Code is released on GitHub (https://github.com/bfshi/VARS).
|
107,393 | 107,393 |
Statistical inference of travelers' route choice preferences with
system-level data
|
Traditional network models encapsulate travel behavior among all
origin-destination pairs based on a simplified and generic utility function.
Typically, the utility function consists of travel time solely and its
coefficients are equated to estimates obtained from stated preference data.
While this modeling strategy is reasonable, the inherent sampling bias in
individual-level data may be further amplified over network flow aggregation,
leading to inaccurate flow estimates. This data must be collected from surveys
or travel diaries, which may be labor intensive, costly and limited to a small
time period. To address these limitations, this study extends classical
bi-level formulations to estimate travelers' utility functions with multiple
attributes using system-level data. We formulate a methodology grounded on
non-linear least squares to statistically infer travelers' utility function in
the network context using traffic counts, traffic speeds, traffic incidents and
sociodemographic information, among other attributes. The analysis of the
mathematical properties of the optimization problem and of its pseudo-convexity
motivate the use of normalized gradient descent. We also develop a hypothesis
test framework to examine statistical properties of the utility function
coefficients and to perform attributes selection. Experiments on synthetic data
show that the coefficients are consistently recovered and that hypothesis tests
are a reliable statistic to identify which attributes are determinants of
travelers' route choices. Besides, a series of Monte-Carlo experiments suggest
that statistical inference is robust to noise in the Origin-Destination matrix
and in the traffic counts, and to various levels of sensor coverage. The
methodology is also deployed at a large scale using real-world multi-source
data in Fresno, CA collected before and during the COVID-19 outbreak.
|
107,394 | 107,394 |
CLIP-Dissect: Automatic Description of Neuron Representations in Deep
Vision Networks
|
In this paper, we propose CLIP-Dissect, a new technique to automatically
describe the function of individual hidden neurons inside vision networks.
CLIP-Dissect leverages recent advances in multimodal vision/language models to
label internal neurons with open-ended concepts without the need for any
labeled data or human examples, which are required for existing tools to
succeed. We show that CLIP-Dissect provides more accurate descriptions than
existing methods for neurons where the ground-truth is available as well as
qualitatively good descriptions for hidden layer neurons. In addition, our
method is very flexible: it is model agnostic, can easily handle new concepts
and can be extended to take advantage of better multimodal models in the
future. Finally CLIP-Dissect is computationally efficient and labels all
neurons of a layer in a large vision model in tens of minutes.
|
107,395 | 107,395 |
Spherical Rotation Dimension Reduction with Geometric Loss Functions
|
Modern datasets witness high-dimensionality and nontrivial geometries of
spaces they live in. It would be helpful in data analysis to reduce the
dimensionality while retaining the geometric structure of the dataset.
Motivated by this observation, we propose a general dimension reduction method
by incorporating geometric information. Our Spherical Rotation Component
Analysis (SRCA) is a dimension reduction method that uses spheres or
ellipsoids, to approximate a low-dimensional manifold. This method not only
generalizes the Spherical Component Analysis (SPCA) method in terms of theories
and algorithms and presents a comprehensive comparison of our method, as an
optimization problem with theoretical guarantee and also as a structural
preserving low-rank representation of data. Results relative to
state-of-the-art competitors show considerable gains in ability to accurately
approximate the subspace with fewer components and better structural
preserving. In addition, we have pointed out that this method is a specific
incarnation of a grander idea of using a geometrically induced loss function in
dimension reduction tasks.
|
107,396 | 107,396 |
Discovering Intrinsic Reward with Contrastive Random Walk
|
The aim of this paper is to demonstrate the efficacy of using Contrastive
Random Walk as a curiosity method to achieve faster convergence to the optimal
policy.Contrastive Random Walk defines the transition matrix of a random walk
with the help of neural networks. It learns a meaningful state representation
with a closed loop. The loss of Contrastive Random Walk serves as an intrinsic
reward and is added to the environment reward. Our method works well in
non-tabular sparse reward scenarios, in the sense that our method receives the
highest reward within the same iterations compared to other methods. Meanwhile,
Contrastive Random Walk is more robust. The performance doesn't change much
with different random initialization of environments. We also find that
adaptive restart and appropriate temperature are crucial to the performance of
Contrastive Random Walk.
|
107,397 | 107,397 |
All-optical graph representation learning using integrated diffractive
photonic computing units
|
Photonic neural networks perform brain-inspired computations using photons
instead of electrons that can achieve substantially improved computing
performance. However, existing architectures can only handle data with regular
structures, e.g., images or videos, but fail to generalize to graph-structured
data beyond Euclidean space, e.g., social networks or document co-citation
networks. Here, we propose an all-optical graph representation learning
architecture, termed diffractive graph neural network (DGNN), based on the
integrated diffractive photonic computing units (DPUs) to address this
limitation. Specifically, DGNN optically encodes node attributes into strip
optical waveguides, which are transformed by DPUs and aggregated by on-chip
optical couplers to extract their feature representations. Each DPU comprises
successive passive layers of metalines to modulate the electromagnetic optical
field via diffraction, where the metaline structures are learnable parameters
shared across graph nodes. DGNN captures complex dependencies among the node
neighborhoods and eliminates the nonlinear transition functions during the
light-speed optical message passing over graph structures. We demonstrate the
use of DGNN extracted features for node and graph-level classification tasks
with benchmark databases and achieve superior performance. Our work opens up a
new direction for designing application-specific integrated photonic circuits
for high-efficiency processing of large-scale graph data structures using deep
learning.
|
107,398 | 107,398 |
Smoothed Online Combinatorial Optimization Using Imperfect Predictions
|
Smoothed online combinatorial optimization considers a learner who repeatedly
chooses a combinatorial decision to minimize an unknown changing cost function
with a penalty on switching decisions in consecutive rounds. We study smoothed
online combinatorial optimization problems when an imperfect predictive model
is available, where the model can forecast the future cost functions with
uncertainty. We show that using predictions to plan for a finite time horizon
leads to regret dependent on the total predictive uncertainty and an additional
switching cost. This observation suggests choosing a suitable planning window
to balance between uncertainty and switching cost, which leads to an online
algorithm with guarantees on the upper and lower bounds of the cumulative
regret. Lastly, we provide an iterative algorithm to approximately solve the
planning problem in real-time. Empirically, our algorithm shows a significant
improvement in cumulative regret compared to other baselines in synthetic
online distributed streaming problems.
|
107,399 | 107,399 |
Distributed Dynamic Safe Screening Algorithms for Sparse Regularization
|
Distributed optimization has been widely used as one of the most efficient
approaches for model training with massive samples. However, large-scale
learning problems with both massive samples and high-dimensional features
widely exist in the era of big data. Safe screening is a popular technique to
speed up high-dimensional models by discarding the inactive features with zero
coefficients. Nevertheless, existing safe screening methods are limited to the
sequential setting. In this paper, we propose a new distributed dynamic safe
screening (DDSS) method for sparsity regularized models and apply it on
shared-memory and distributed-memory architecture respectively, which can
achieve significant speedup without any loss of accuracy by simultaneously
enjoying the sparsity of the model and dataset. To the best of our knowledge,
this is the first work of distributed safe dynamic screening method.
Theoretically, we prove that the proposed method achieves the linear
convergence rate with lower overall complexity and can eliminate almost all the
inactive features in a finite number of iterations almost surely. Finally,
extensive experimental results on benchmark datasets confirm the superiority of
our proposed method.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.