Unnamed: 0.1
int64 0
113k
| Unnamed: 0
float64 0
113k
⌀ | title
stringlengths 7
246
| abstract
stringlengths 6
3.31k
|
|---|---|---|---|
112,500 | 112,500 | Considering users' behaviours in improving the responses of an
information base | In this paper, our aim is to propose a model that helps in the efficient use
of an information system by users, within the organization represented by the
IS, in order to resolve their decisional problems. In other words we want to
aid the user within an organization in obtaining the information that
corresponds to his needs (informational needs that result from his decisional
problems). This type of information system is what we refer to as economic
intelligence system because of its support for economic intelligence processes
of the organisation. Our assumption is that every EI process begins with the
identification of the decisional problem which is translated into an
informational need. This need is then translated into one or many information
search problems (ISP). We also assumed that an ISP is expressed in terms of the
user's expectations and that these expectations determine the activities or the
behaviors of the user, when he/she uses an IS. The model we are proposing is
used for the conception of the IS so that the process of retrieving of
solution(s) or the responses given by the system to an ISP is based on these
behaviours and correspond to the needs of the user.
|
112,501 | 112,501 | Low-complexity modular policies: learning to play Pac-Man and a new
framework beyond MDPs | In this paper we propose a method that learns to play Pac-Man. We define a
set of high-level observation and action modules. Actions are temporally
extended, and multiple action modules may be in effect concurrently. A decision
of the agent is represented as a rule-based policy. For learning, we apply the
cross-entropy method, a recent global optimization algorithm. The learned
policies reached better score than the hand-crafted policy, and neared the
score of average human players. We argue that learning is successful mainly
because (i) the policy space includes the combination of individual actions and
thus it is sufficiently rich, (ii) the search is biased towards low-complexity
policies and low complexity solutions can be found quickly if they exist. Based
on these principles, we formulate a new theoretical framework, which can be
found in the Appendix as supporting material.
|
112,502 | 112,502 | Evolving controllers for simulated car racing | This paper describes the evolution of controllers for racing a simulated
radio-controlled car around a track, modelled on a real physical track. Five
different controller architectures were compared, based on neural networks,
force fields and action sequences. The controllers use either egocentric (first
person), Newtonian (third person) or no information about the state of the car
(open-loop controller). The only controller that was able to evolve good racing
behaviour was based on a neural network acting on egocentric inputs.
|
112,503 | 112,503 | Hedging predictions in machine learning | Recent advances in machine learning make it possible to design efficient
prediction algorithms for data sets with huge numbers of parameters. This paper
describes a new technique for "hedging" the predictions output by many such
algorithms, including support vector machines, kernel ridge regression, kernel
nearest neighbours, and by many other state-of-the-art methods. The hedged
predictions for the labels of new objects include quantitative measures of
their own accuracy and reliability. These measures are provably valid under the
assumption of randomness, traditional in machine learning: the objects and
their labels are assumed to be generated independently from the same
probability distribution. In particular, it becomes possible to control (up to
statistical fluctuations) the number of erroneous predictions by selecting a
suitable confidence level. Validity being achieved automatically, the remaining
goal of hedged prediction is efficiency: taking full account of the new
objects' features and other available information to produce as accurate
predictions as possible. This can be done successfully using the powerful
machinery of modern machine learning.
|
112,504 | 112,504 | A Relational Approach to Functional Decomposition of Logic Circuits | Functional decomposition of logic circuits has profound influence on all
quality aspects of the cost-effective implementation of modern digital systems.
In this paper, a relational approach to the decomposition of logic circuits is
proposed. This approach is parallel to the normalization of relational
databases, they are governed by the same concepts of functional dependency (FD)
and multi-valued dependency (MVD). It is manifest that the functional
decomposition of switching function actually exploits the same idea and serves
a similar purpose as database normalization. Partitions play an important role
in the decomposition. The interdependency of two partitions can be represented
by a bipartite graph. We demonstrate that both FD and MVD can be represented by
bipartite graphs with specific topological properties, which are delineated by
partitions of minterms. It follows that our algorithms are procedures of
constructing those specific bipartite graphs of interest to meet the
information-lossless criteria of functional decomposition.
|
112,505 | 112,505 | CSCR:Computer Supported Collaborative Research | It is suggested that a new area of CSCR (Computer Supported Collaborative
Research) is distinguished from CSCW and CSCL and that the demarcation between
the three areas could do with greater clarification and prescription.
|
112,506 | 112,506 | How Random is a Coin Toss? Bayesian Inference and the Symbolic Dynamics
of Deterministic Chaos | Symbolic dynamics has proven to be an invaluable tool in analyzing the
mechanisms that lead to unpredictability and random behavior in nonlinear
dynamical systems. Surprisingly, a discrete partition of continuous state space
can produce a coarse-grained description of the behavior that accurately
describes the invariant properties of an underlying chaotic attractor. In
particular, measures of the rate of information production--the topological and
metric entropy rates--can be estimated from the outputs of Markov or generating
partitions. Here we develop Bayesian inference for k-th order Markov chains as
a method to finding generating partitions and estimating entropy rates from
finite samples of discretized data produced by coarse-grained dynamical
systems.
|
112,507 | 112,507 | Very Sparse Stable Random Projections, Estimators and Tail Bounds for
Stable Random Projections | This paper will focus on three different aspects in improving the current
practice of stable random projections.
Firstly, we propose {\em very sparse stable random projections} to
significantly reduce the processing and storage cost, by replacing the
$\alpha$-stable distribution with a mixture of a symmetric $\alpha$-Pareto
distribution (with probability $\beta$, $0<\beta\leq1$) and a point mass at the
origin (with a probability $1-\beta$). This leads to a significant
$\frac{1}{\beta}$-fold speedup for small $\beta$.
Secondly, we provide an improved estimator for recovering the original
$l_\alpha$ norms from the projected data. The standard estimator is based on
the (absolute) sample median, while we suggest using the geometric mean. The
geometric mean estimator we propose is strictly unbiased and is easier to
study. Moreover, the geometric mean estimator is more accurate, especially
non-asymptotically.
Thirdly, we provide an adequate answer to the basic question of how many
projections (samples) are needed for achieving some pre-specified level of
accuracy. \cite{Proc:Indyk_FOCS00,Article:Indyk_TKDE03} did not provide a
criterion that can be used in practice. The geometric mean estimator we propose
allows us to derive sharp tail bounds which can be expressed in exponential
forms with constants explicitly given.
|
112,508 | 112,508 | Functional Bregman Divergence and Bayesian Estimation of Distributions | A class of distortions termed functional Bregman divergences is defined,
which includes squared error and relative entropy. A functional Bregman
divergence acts on functions or distributions, and generalizes the standard
Bregman divergence for vectors and a previous pointwise Bregman divergence that
was defined for functions. A recently published result showed that the mean
minimizes the expected Bregman divergence. The new functional definition
enables the extension of this result to the continuous case to show that the
mean minimizes the expected functional Bregman divergence over a set of
functions or distributions. It is shown how this theorem applies to the
Bayesian estimation of distributions. Estimation of the uniform distribution
from independent and identically drawn samples is used as a case study.
|
112,509 | 112,509 | Low-rank matrix factorization with attributes | We develop a new collaborative filtering (CF) method that combines both
previously known users' preferences, i.e. standard CF, as well as product/user
attributes, i.e. classical function approximation, to predict a given user's
interest in a particular product. Our method is a generalized low rank matrix
completion problem, where we learn a function whose inputs are pairs of vectors
-- the standard low rank matrix completion problem being a special case where
the inputs to the function are the row and column indices of the matrix. We
solve this generalized matrix completion problem using tensor product kernels
for which we also formally generalize standard kernel properties. Benchmark
experiments on movie ratings show the advantages of our generalized matrix
completion method over the standard matrix completion one with no information
about movies or people, as well as over standard multi-task or single task
learning methods.
|
112,510 | 112,510 | A Unified View of TD Algorithms; Introducing Full-Gradient TD and
Equi-Gradient Descent TD | This paper addresses the issue of policy evaluation in Markov Decision
Processes, using linear function approximation. It provides a unified view of
algorithms such as TD(lambda), LSTD(lambda), iLSTD, residual-gradient TD. It is
asserted that they all consist in minimizing a gradient function and differ by
the form of this function and their means of minimizing it. Two new schemes are
introduced in that framework: Full-gradient TD which uses a generalization of
the principle introduced in iLSTD, and EGD TD, which reduces the gradient by
successive equi-gradient descents. These three algorithms form a new
intermediate family with the interesting property of making much better use of
the samples than TD while keeping a gradient descent scheme, which is useful
for complexity issues and optimistic policy iteration.
|
112,511 | 112,511 | A Novel Bayesian Classifier using Copula Functions | A useful method for representing Bayesian classifiers is through
\emph{discriminant functions}. Here, using copula functions, we propose a new
model for discriminants. This model provides a rich and generalized class of
decision boundaries. These decision boundaries significantly boost the
classification accuracy especially for high dimensional feature spaces. We
strengthen our analysis through simulation results.
|
112,512 | 112,512 | Player co-modelling in a strategy board game: discovering how to play
fast | In this paper we experiment with a 2-player strategy board game where playing
models are evolved using reinforcement learning and neural networks. The models
are evolved to speed up automatic game development based on human involvement
at varying levels of sophistication and density when compared to fully
autonomous playing. The experimental results suggest a clear and measurable
association between the ability to win games and the ability to do that fast,
while at the same time demonstrating that there is a minimum level of human
involvement beyond which no learning really occurs.
|
112,513 | 112,513 | Loop corrections for approximate inference | We propose a method for improving approximate inference methods that corrects
for the influence of loops in the graphical model. The method is applicable to
arbitrary factor graphs, provided that the size of the Markov blankets is not
too large. It is an alternative implementation of an idea introduced recently
by Montanari and Rizzo (2005). In its simplest form, which amounts to the
assumption that no loops are present, the method reduces to the minimal Cluster
Variation Method approximation (which uses maximal factors as outer clusters).
On the other hand, using estimates of the effect of loops (obtained by some
approximate inference algorithm) and applying the Loop Correcting (LC) method
usually gives significantly better results than applying the approximate
inference algorithm directly without loop corrections. Indeed, we often observe
that the loop corrected error is approximately the square of the error of the
approximate inference method used to estimate the effect of loops. We compare
different variants of the Loop Correcting method with other approximate
inference methods on a variety of graphical models, including "real world"
networks, and conclude that the LC approach generally obtains the most accurate
results.
|
112,514 | 112,514 | Approximation of the Two-Part MDL Code | Approximation of the optimal two-part MDL code for given data, through
successive monotonically length-decreasing two-part MDL codes, has the
following properties: (i) computation of each step may take arbitrarily long;
(ii) we may not know when we reach the optimum, or whether we will reach the
optimum at all; (iii) the sequence of models generated may not monotonically
improve the goodness of fit; but (iv) the model associated with the optimum has
(almost) the best goodness of fit. To express the practically interesting
goodness of fit of individual models for individual data sets we have to rely
on Kolmogorov complexity.
|
112,515 | 112,515 | Using state space differential geometry for nonlinear blind source
separation | Given a time series of multicomponent measurements of an evolving stimulus,
nonlinear blind source separation (BSS) seeks to find a "source" time series,
comprised of statistically independent combinations of the measured components.
In this paper, we seek a source time series with local velocity cross
correlations that vanish everywhere in stimulus state space. However, in an
earlier paper the local velocity correlation matrix was shown to constitute a
metric on state space. Therefore, nonlinear BSS maps onto a problem of
differential geometry: given the metric observed in the measurement coordinate
system, find another coordinate system in which the metric is diagonal
everywhere. We show how to determine if the observed data are separable in this
way, and, if they are, we show how to construct the required transformation to
the source coordinate system, which is essentially unique except for an unknown
rotation that can be found by applying the methods of linear BSS. Thus, the
proposed technique solves nonlinear BSS in many situations or, at least,
reduces it to linear BSS, without the use of probabilistic, parametric, or
iterative procedures. This paper also describes a generalization of this
methodology that performs nonlinear independent subspace separation. In every
case, the resulting decomposition of the observed data is an intrinsic property
of the stimulus' evolution in the sense that it does not depend on the way the
observer chooses to view it (e.g., the choice of the observing machine's
sensors). In other words, the decomposition is a property of the evolution of
the "real" stimulus that is "out there" broadcasting energy to the observer.
The technique is illustrated with analytic and numerical examples.
|
112,516 | 112,516 | Statistical Mechanics of On-line Learning when a Moving Teacher Goes
around an Unlearnable True Teacher | In the framework of on-line learning, a learning machine might move around a
teacher due to the differences in structures or output functions between the
teacher and the learning machine. In this paper we analyze the generalization
performance of a new student supervised by a moving machine. A model composed
of a fixed true teacher, a moving teacher, and a student is treated
theoretically using statistical mechanics, where the true teacher is a
nonmonotonic perceptron and the others are simple perceptrons. Calculating the
generalization errors numerically, we show that the generalization errors of a
student can temporarily become smaller than that of a moving teacher, even if
the student only uses examples from the moving teacher. However, the
generalization error of the student eventually becomes the same value with that
of the moving teacher. This behavior is qualitatively different from that of a
linear model.
|
112,517 | 112,517 | Time Series Forecasting: Obtaining Long Term Trends with Self-Organizing
Maps | Kohonen self-organisation maps are a well know classification tool, commonly
used in a wide variety of problems, but with limited applications in time
series forecasting context. In this paper, we propose a forecasting method
specifically designed for multi-dimensional long-term trends prediction, with a
double application of the Kohonen algorithm. Practical applications of the
method are also presented.
|
112,518 | 112,518 | A Delta Debugger for ILP Query Execution | Because query execution is the most crucial part of Inductive Logic
Programming (ILP) algorithms, a lot of effort is invested in developing faster
execution mechanisms. These execution mechanisms typically have a low-level
implementation, making them hard to debug. Moreover, other factors such as the
complexity of the problems handled by ILP algorithms and size of the code base
of ILP data mining systems make debugging at this level a very difficult job.
In this work, we present the trace-based debugging approach currently used in
the development of new execution mechanisms in hipP, the engine underlying the
ACE Data Mining system. This debugger uses the delta debugging algorithm to
automatically reduce the total time needed to expose bugs in ILP execution,
thus making manual debugging step much lighter.
|
112,519 | 112,519 | Algorithmic Complexity Bounds on Future Prediction Errors | We bound the future loss when predicting any (computably) stochastic sequence
online. Solomonoff finitely bounded the total deviation of his universal
predictor $M$ from the true distribution $mu$ by the algorithmic complexity of
$mu$. Here we assume we are at a time $t>1$ and already observed $x=x_1...x_t$.
We bound the future prediction performance on $x_{t+1}x_{t+2}...$ by a new
variant of algorithmic complexity of $mu$ given $x$, plus the complexity of the
randomness deficiency of $x$. The new complexity is monotone in its condition
in the sense that this complexity can only decrease if the condition is
prolonged. We also briefly discuss potential generalizations to Bayesian model
classes and to classification problems.
|
112,520 | 112,520 | Universal Algorithmic Intelligence: A mathematical top->down approach | Sequential decision theory formally solves the problem of rational agents in
uncertain worlds if the true environmental prior probability distribution is
known. Solomonoff's theory of universal induction formally solves the problem
of sequence prediction for unknown prior distribution. We combine both ideas
and get a parameter-free theory of universal Artificial Intelligence. We give
strong arguments that the resulting AIXI model is the most intelligent unbiased
agent possible. We outline how the AIXI model can formally solve a number of
problem classes, including sequence prediction, strategic games, function
minimization, reinforcement and supervised learning. The major drawback of the
AIXI model is that it is uncomputable. To overcome this problem, we construct a
modified algorithm AIXItl that is still effectively more intelligent than any
other time t and length l bounded agent. The computation time of AIXItl is of
the order t x 2^l. The discussion includes formal definitions of intelligence
order relations, the horizon problem and relations of the AIXI theory to other
AI approaches.
|
112,521 | 112,521 | Support and Quantile Tubes | This correspondence studies an estimator of the conditional support of a
distribution underlying a set of i.i.d. observations. The relation with mutual
information is shown via an extension of Fano's theorem in combination with a
generalization bound based on a compression argument. Extensions to estimating
the conditional quantile interval, and statistical guarantees on the minimal
convex hull are given.
|
112,522 | 112,522 | Bandit Algorithms for Tree Search | Bandit based methods for tree search have recently gained popularity when
applied to huge trees, e.g. in the game of go (Gelly et al., 2006). The UCT
algorithm (Kocsis and Szepesvari, 2006), a tree search method based on Upper
Confidence Bounds (UCB) (Auer et al., 2002), is believed to adapt locally to
the effective smoothness of the tree. However, we show that UCT is too
``optimistic'' in some cases, leading to a regret O(exp(exp(D))) where D is the
depth of the tree. We propose alternative bandit algorithms for tree search.
First, a modification of UCT using a confidence sequence that scales
exponentially with the horizon depth is proven to have a regret O(2^D
\sqrt{n}), but does not adapt to possible smoothness in the tree. We then
analyze Flat-UCB performed on the leaves and provide a finite regret bound with
high probability. Then, we introduce a UCB-based Bandit Algorithm for Smooth
Trees which takes into account actual smoothness of the rewards for performing
efficient ``cuts'' of sub-optimal branches with high confidence. Finally, we
present an incremental tree search version which applies when the full tree is
too big (possibly infinite) to be entirely represented and show that with high
probability, essentially only the optimal branches is indefinitely developed.
We illustrate these methods on a global optimization problem of a Lipschitz
function, given noisy data.
|
112,523 | 112,523 | Intrinsic dimension of a dataset: what properties does one expect? | We propose an axiomatic approach to the concept of an intrinsic dimension of
a dataset, based on a viewpoint of geometry of high-dimensional structures. Our
first axiom postulates that high values of dimension be indicative of the
presence of the curse of dimensionality (in a certain precise mathematical
sense). The second axiom requires the dimension to depend smoothly on a
distance between datasets (so that the dimension of a dataset and that of an
approximating principal manifold would be close to each other). The third axiom
is a normalization condition: the dimension of the Euclidean $n$-sphere $\s^n$
is $\Theta(n)$. We give an example of a dimension function satisfying our
axioms, even though it is in general computationally unfeasible, and discuss a
computationally cheap function satisfying most but not all of our axioms (the
``intrinsic dimensionality'' of Ch\'avez et al.)
|
112,524 | 112,524 | Structure induction by lossless graph compression | This work is motivated by the necessity to automate the discovery of
structure in vast and evergrowing collection of relational data commonly
represented as graphs, for example genomic networks. A novel algorithm, dubbed
Graphitour, for structure induction by lossless graph compression is presented
and illustrated by a clear and broadly known case of nested structure in a DNA
molecule. This work extends to graphs some well established approaches to
grammatical inference previously applied only to strings. The bottom-up graph
compression problem is related to the maximum cardinality (non-bipartite)
maximum cardinality matching problem. The algorithm accepts a variety of graph
types including directed graphs and graphs with labeled nodes and arcs. The
resulting structure could be used for representation and classification of
graphs.
|
112,525 | 112,525 | Reinforcement Learning for Adaptive Routing | Reinforcement learning means learning a policy--a mapping of observations
into actions--based on feedback from the environment. The learning can be
viewed as browsing a set of policies while evaluating them by trial through
interaction with the environment. We present an application of gradient ascent
algorithm for reinforcement learning to a complex domain of packet routing in
network communication and compare the performance of this algorithm to other
routing methods on a benchmark problem.
|
112,526 | 112,526 | Similarity-Based Models of Word Cooccurrence Probabilities | In many applications of natural language processing (NLP) it is necessary to
determine the likelihood of a given word combination. For example, a speech
recognizer may need to determine which of the two word combinations ``eat a
peach'' and ``eat a beach'' is more likely. Statistical NLP methods determine
the likelihood of a word combination from its frequency in a training corpus.
However, the nature of language is such that many word combinations are
infrequent and do not occur in any given corpus. In this work we propose a
method for estimating the probability of such previously unseen word
combinations using available information on ``most similar'' words.
We describe probabilistic word association models based on distributional
word similarity, and apply them to two tasks, language modeling and pseudo-word
disambiguation. In the language modeling task, a similarity-based model is used
to improve probability estimates for unseen bigrams in a back-off language
model. The similarity-based method yields a 20% perplexity improvement in the
prediction of unseen bigrams and statistically significant reductions in
speech-recognition error.
We also compare four similarity-based estimation methods against back-off and
maximum-likelihood estimation methods on a pseudo-word sense disambiguation
task in which we controlled for both unigram and bigram frequency to avoid
giving too much weight to easy-to-disambiguate high-frequency configurations.
The similarity-based methods perform up to 40% better on this particular task.
|
112,527 | 112,527 | Evolution of Neural Networks to Play the Game of Dots-and-Boxes | Dots-and-Boxes is a child's game which remains analytically unsolved. We
implement and evolve artificial neural networks to play this game, evaluating
them against simple heuristic players. Our networks do not evaluate or predict
the final outcome of the game, but rather recommend moves at each stage.
Superior generalisation of play by co-evolved populations is found, and a
comparison made with networks trained by back-propagation using simple
heuristics as an oracle.
|
112,528 | 112,528 | Practical algorithms for on-line sampling | One of the core applications of machine learning to knowledge discovery
consists on building a function (a hypothesis) from a given amount of data (for
instance a decision tree or a neural network) such that we can use it
afterwards to predict new instances of the data. In this paper, we focus on a
particular situation where we assume that the hypothesis we want to use for
prediction is very simple, and thus, the hypotheses class is of feasible size.
We study the problem of how to determine which of the hypotheses in the class
is almost the best one. We present two on-line sampling algorithms for
selecting hypotheses, give theoretical bounds on the number of necessary
examples, and analize them exprimentally. We compare them with the simple batch
sampling approach commonly used and show that in most of the situations our
algorithms use much fewer number of examples.
|
112,529 | 112,529 | A Winnow-Based Approach to Context-Sensitive Spelling Correction | A large class of machine-learning problems in natural language require the
characterization of linguistic context. Two characteristic properties of such
problems are that their feature space is of very high dimensionality, and their
target concepts refer to only a small subset of the features in the space.
Under such conditions, multiplicative weight-update algorithms such as Winnow
have been shown to have exceptionally good theoretical properties. We present
an algorithm combining variants of Winnow and weighted-majority voting, and
apply it to a problem in the aforementioned class: context-sensitive spelling
correction. This is the task of fixing spelling errors that happen to result in
valid words, such as substituting "to" for "too", "casual" for "causal", etc.
We evaluate our algorithm, WinSpell, by comparing it against BaySpell, a
statistics-based method representing the state of the art for this task. We
find: (1) When run with a full (unpruned) set of features, WinSpell achieves
accuracies significantly higher than BaySpell was able to achieve in either the
pruned or unpruned condition; (2) When compared with other systems in the
literature, WinSpell exhibits the highest performance; (3) The primary reason
that WinSpell outperforms BaySpell is that WinSpell learns a better linear
separator; (4) When run on a test set drawn from a different corpus than the
training set was drawn from, WinSpell is better able than BaySpell to adapt,
using a strategy we will present that combines supervised learning on the
training set with unsupervised learning on the (noisy) test set.
|
112,530 | 112,530 | Machine Learning of Generic and User-Focused Summarization | A key problem in text summarization is finding a salience function which
determines what information in the source should be included in the summary.
This paper describes the use of machine learning on a training corpus of
documents and their abstracts to discover salience functions which describe
what combination of features is optimal for a given summarization task. The
method addresses both "generic" and user-focused summaries.
|
112,531 | 112,531 | Learning to Resolve Natural Language Ambiguities: A Unified Approach | We analyze a few of the commonly used statistics based and machine learning
algorithms for natural language disambiguation tasks and observe that they can
be re-cast as learning linear separators in the feature space. Each of the
methods makes a priori assumptions, which it employs, given the data, when
searching for its hypothesis. Nevertheless, as we show, it searches a space
that is as rich as the space of all linear separators. We use this to build an
argument for a data driven approach which merely searches for a good linear
separator in the feature space, without further assumptions on the domain or a
specific problem.
We present such an approach - a sparse network of linear separators,
utilizing the Winnow learning algorithm - and show how to use it in a variety
of ambiguity resolution problems. The learning approach presented is
attribute-efficient and, therefore, appropriate for domains having very large
number of attributes.
In particular, we present an extensive experimental comparison of our
approach with other methods on several well studied lexical disambiguation
tasks such as context-sensitive spelling correction, prepositional phrase
attachment and part of speech tagging. In all cases we show that our approach
either outperforms other methods tried for these tasks or performs comparably
to the best.
|
112,532 | 112,532 | Forgetting Exceptions is Harmful in Language Learning | We show that in language learning, contrary to received wisdom, keeping
exceptional training instances in memory can be beneficial for generalization
accuracy. We investigate this phenomenon empirically on a selection of
benchmark natural language processing tasks: grapheme-to-phoneme conversion,
part-of-speech tagging, prepositional-phrase attachment, and base noun phrase
chunking. In a first series of experiments we combine memory-based learning
with training set editing techniques, in which instances are edited based on
their typicality and class prediction strength. Results show that editing
exceptional instances (with low typicality or low class prediction strength)
tends to harm generalization accuracy. In a second series of experiments we
compare memory-based learning and decision-tree learning methods on the same
selection of tasks, and find that decision-tree learning often performs worse
than memory-based learning. Moreover, the decrease in performance can be linked
to the degree of abstraction from exceptions (i.e., pruning or eagerness). We
provide explanations for both results in terms of the properties of the natural
language processing tasks and the learning algorithms.
|
112,533 | 112,533 | TDLeaf(lambda): Combining Temporal Difference Learning with Game-Tree
Search | In this paper we present TDLeaf(lambda), a variation on the TD(lambda)
algorithm that enables it to be used in conjunction with minimax search. We
present some experiments in both chess and backgammon which demonstrate its
utility and provide comparisons with TD(lambda) and another less radical
variant, TD-directed(lambda). In particular, our chess program, ``KnightCap,''
used TDLeaf(lambda) to learn its evaluation function while playing on the Free
Internet Chess Server (FICS, fics.onenet.net). It improved from a 1650 rating
to a 2100 rating in just 308 games. We discuss some of the reasons for this
success and the relationship between our results and Tesauro's results in
backgammon.
|
112,534 | 112,534 | KnightCap: A chess program that learns by combining TD(lambda) with
game-tree search | In this paper we present TDLeaf(lambda), a variation on the TD(lambda)
algorithm that enables it to be used in conjunction with game-tree search. We
present some experiments in which our chess program ``KnightCap'' used
TDLeaf(lambda) to learn its evaluation function while playing on the Free
Internet Chess Server (FICS, fics.onenet.net). The main success we report is
that KnightCap improved from a 1650 rating to a 2150 rating in just 308 games
and 3 days of play. As a reference, a rating of 1650 corresponds to about level
B human play (on a scale from E (1000) to A (1800)), while 2150 is human master
level. We discuss some of the reasons for this success, principle among them
being the use of on-line, rather than self-play.
|
112,535 | 112,535 | Minimum Description Length Induction, Bayesianism, and Kolmogorov
Complexity | The relationship between the Bayesian approach and the minimum description
length approach is established. We sharpen and clarify the general modeling
principles MDL and MML, abstracted as the ideal MDL principle and defined from
Bayes's rule by means of Kolmogorov complexity. The basic condition under which
the ideal principle should be applied is encapsulated as the Fundamental
Inequality, which in broad terms states that the principle is valid when the
data are random, relative to every contemplated hypothesis and also these
hypotheses are random relative to the (universal) prior. Basically, the ideal
principle states that the prior probability associated with the hypothesis
should be given by the algorithmic universal probability, and the sum of the
log universal probability of the model plus the log of the probability of the
data given the model should be minimized. If we restrict the model class to the
finite sets then application of the ideal principle turns into Kolmogorov's
minimal sufficient statistic. In general we show that data compression is
almost always the best strategy, both in hypothesis identification and
prediction.
|
112,536 | 112,536 | A Discipline of Evolutionary Programming | Genetic fitness optimization using small populations or small population
updates across generations generally suffers from randomly diverging
evolutions. We propose a notion of highly probable fitness optimization through
feasible evolutionary computing runs on small size populations. Based on
rapidly mixing Markov chains, the approach pertains to most types of
evolutionary genetic algorithms, genetic programming and the like. We establish
that for systems having associated rapidly mixing Markov chains and appropriate
stationary distributions the new method finds optimal programs (individuals)
with probability almost 1. To make the method useful would require a structured
design methodology where the development of the program and the guarantee of
the rapidly mixing property go hand in hand. We analyze a simple example to
show that the method is implementable. More significant examples require
theoretical advances, for example with respect to the Metropolis filter.
|
112,537 | 112,537 | Probabilistic Inductive Inference:a Survey | Inductive inference is a recursion-theoretic theory of learning, first
developed by E. M. Gold (1967). This paper surveys developments in
probabilistic inductive inference. We mainly focus on finite inference of
recursive functions, since this simple paradigm has produced the most
interesting (and most complex) results.
|
112,538 | 112,538 | Using Collective Intelligence to Route Internet Traffic | A COllective INtelligence (COIN) is a set of interacting reinforcement
learning (RL) algorithms designed in an automated fashion so that their
collective behavior optimizes a global utility function. We summarize the
theory of COINs, then present experiments using that theory to design COINs to
control internet traffic routing. These experiments indicate that COINs
outperform all previously investigated RL-based, shortest path routing
algorithms.
|
112,539 | 112,539 | General Principles of Learning-Based Multi-Agent Systems | We consider the problem of how to design large decentralized multi-agent
systems (MAS's) in an automated fashion, with little or no hand-tuning. Our
approach has each agent run a reinforcement learning algorithm. This converts
the problem into one of how to automatically set/update the reward functions
for each of the agents so that the global goal is achieved. In particular we do
not want the agents to ``work at cross-purposes'' as far as the global goal is
concerned. We use the term artificial COllective INtelligence (COIN) to refer
to systems that embody solutions to this problem. In this paper we present a
summary of a mathematical framework for COINs. We then investigate the
real-world applicability of the core concepts of that framework via two
computer experiments: we show that our COINs perform near optimally in a
difficult variant of Arthur's bar problem (and in particular avoid the tragedy
of the commons for that problem), and we also illustrate optimal performance
for our COINs in the leader-follower problem.
|
112,540 | 112,540 | An Efficient, Probabilistically Sound Algorithm for Segmentation and
Word Discovery | This paper presents a model-based, unsupervised algorithm for recovering word
boundaries in a natural-language text from which they have been deleted. The
algorithm is derived from a probability model of the source that generated the
text. The fundamental structure of the model is specified abstractly so that
the detailed component models of phonology, word-order, and word frequency can
be replaced in a modular fashion. The model yields a language-independent,
prior probability distribution on all possible sequences of all possible words
over a given alphabet, based on the assumption that the input was generated by
concatenating words from a fixed but unknown lexicon. The model is unusual in
that it treats the generation of a complete corpus, regardless of length, as a
single event in the probability space. Accordingly, the algorithm does not
estimate a probability distribution on words; instead, it attempts to calculate
the prior probabilities of various word sequences that could underlie the
observed text. Experiments on phonemic transcripts of spontaneous speech by
parents to young children suggest that this algorithm is more effective than
other proposed algorithms, at least when utterance boundaries are given and the
text includes a substantial number of short utterances.
Keywords: Bayesian grammar induction, probability models, minimum description
length (MDL), unsupervised learning, cognitive modeling, language acquisition,
segmentation
|
112,541 | 112,541 | Inducing a Semantically Annotated Lexicon via EM-Based Clustering | We present a technique for automatic induction of slot annotations for
subcategorization frames, based on induction of hidden classes in the EM
framework of statistical estimation. The models are empirically evalutated by a
general decision test. Induction of slot labeling for subcategorization frames
is accomplished by a further application of EM, and applied experimentally on
frame observations derived from parsing large corpora. We outline an
interpretation of the learned representations as theoretical-linguistic
decompositional lexical entries.
|
112,542 | 112,542 | Inside-Outside Estimation of a Lexicalized PCFG for German | The paper describes an extensive experiment in inside-outside estimation of a
lexicalized probabilistic context free grammar for German verb-final clauses.
Grammar and formalism features which make the experiment feasible are
described. Successive models are evaluated on precision and recall of phrase
markup.
|
112,543 | 112,543 | Statistical Inference and Probabilistic Modelling for Constraint-Based
NLP | We present a probabilistic model for constraint-based grammars and a method
for estimating the parameters of such models from incomplete, i.e., unparsed
data. Whereas methods exist to estimate the parameters of probabilistic
context-free grammars from incomplete data (Baum 1970), so far for
probabilistic grammars involving context-dependencies only parameter estimation
techniques from complete, i.e., fully parsed data have been presented (Abney
1997). However, complete-data estimation requires labor-intensive, error-prone,
and grammar-specific hand-annotating of large language corpora. We present a
log-linear probability model for constraint logic programming, and a general
algorithm to estimate the parameters of such models from incomplete data by
extending the estimation algorithm of Della-Pietra, Della-Pietra, and Lafferty
(1997) to incomplete data settings.
|
112,544 | 112,544 | Ensembles of Radial Basis Function Networks for Spectroscopic Detection
of Cervical Pre-Cancer | The mortality related to cervical cancer can be substantially reduced through
early detection and treatment. However, current detection techniques, such as
Pap smear and colposcopy, fail to achieve a concurrently high sensitivity and
specificity.
In vivo fluorescence spectroscopy is a technique which quickly,
non-invasively and quantitatively probes the biochemical and morphological
changes that occur in pre-cancerous tissue.
A multivariate statistical algorithm was used to extract clinically useful
information from tissue spectra acquired from 361 cervical sites from 95
patients at 337, 380 and 460 nm excitation wavelengths. The multivariate
statistical analysis was also employed to reduce the number of fluorescence
excitation-emission wavelength pairs required to discriminate healthy tissue
samples from pre-cancerous tissue samples. The use of connectionist methods
such as multi layered perceptrons, radial basis function networks, and
ensembles of such networks was investigated. RBF ensemble algorithms based on
fluorescence spectra potentially provide automated, and near real-time
implementation of pre-cancer detection in the hands of non-experts. The results
are more reliable, direct and accurate than those achieved by either human
experts or multivariate statistical algorithms.
|
112,545 | 112,545 | Linear and Order Statistics Combiners for Pattern Classification | Several researchers have experimentally shown that substantial improvements
can be obtained in difficult pattern recognition problems by combining or
integrating the outputs of multiple classifiers. This chapter provides an
analytical framework to quantify the improvements in classification results due
to combining. The results apply to both linear combiners and order statistics
combiners. We first show that to a first order approximation, the error rate
obtained over and above the Bayes error rate, is directly proportional to the
variance of the actual decision boundaries around the Bayes optimum boundary.
Combining classifiers in output space reduces this variance, and hence reduces
the "added" error. If N unbiased classifiers are combined by simple averaging,
the added error rate can be reduced by a factor of N if the individual errors
in approximating the decision boundaries are uncorrelated. Expressions are then
derived for linear combiners which are biased or correlated, and the effect of
output correlations on ensemble performance is quantified. For order statistics
based non-linear combiners, we derive expressions that indicate how much the
median, the maximum and in general the ith order statistic can improve
classifier performance. The analysis presented here facilitates the
understanding of the relationships among error rates, classifier boundary
distributions, and combining in output space. Experimental results on several
public domain data sets are provided to illustrate the benefits of combining
and to support the analytical results.
|
112,546 | 112,546 | Robust Combining of Disparate Classifiers through Order Statistics | Integrating the outputs of multiple classifiers via combiners or
meta-learners has led to substantial improvements in several difficult pattern
recognition problems. In the typical setting investigated till now, each
classifier is trained on data taken or resampled from a common data set, or
(almost) randomly selected subsets thereof, and thus experiences similar
quality of training data. However, in certain situations where data is acquired
and analyzed on-line at several geographically distributed locations, the
quality of data may vary substantially, leading to large discrepancies in
performance of individual classifiers. In this article we introduce and
investigate a family of classifiers based on order statistics, for robust
handling of such cases. Based on a mathematical modeling of how the decision
boundaries are affected by order statistic combiners, we derive expressions for
the reductions in error expected when such combiners are used. We show
analytically that the selection of the median, the maximum and in general, the
$i^{th}$ order statistic improves classification performance. Furthermore, we
introduce the trim and spread combiners, both based on linear combinations of
the ordered classifier outputs, and show that they are quite beneficial in
presence of outliers or uneven classifier performance. Experimental results on
several public domain data sets corroborate these findings.
|
112,547 | 112,547 | Hierarchical Reinforcement Learning with the MAXQ Value Function
Decomposition | This paper presents the MAXQ approach to hierarchical reinforcement learning
based on decomposing the target Markov decision process (MDP) into a hierarchy
of smaller MDPs and decomposing the value function of the target MDP into an
additive combination of the value functions of the smaller MDPs. The paper
defines the MAXQ hierarchy, proves formal results on its representational
power, and establishes five conditions for the safe use of state abstractions.
The paper presents an online model-free learning algorithm, MAXQ-Q, and proves
that it converges wih probability 1 to a kind of locally-optimal policy known
as a recursively optimal policy, even in the presence of the five kinds of
state abstraction. The paper evaluates the MAXQ representation and MAXQ-Q
through a series of experiments in three domains and shows experimentally that
MAXQ-Q (with state abstractions) converges to a recursively optimal policy much
faster than flat Q learning. The fact that MAXQ learns a representation of the
value function has an important benefit: it makes it possible to compute and
execute an improved, non-hierarchical policy via a procedure similar to the
policy improvement step of policy iteration. The paper demonstrates the
effectiveness of this non-hierarchical execution experimentally. Finally, the
paper concludes with a comparison to related work and a discussion of the
design tradeoffs in hierarchical reinforcement learning.
|
112,548 | 112,548 | State Abstraction in MAXQ Hierarchical Reinforcement Learning | Many researchers have explored methods for hierarchical reinforcement
learning (RL) with temporal abstractions, in which abstract actions are defined
that can perform many primitive actions before terminating. However, little is
known about learning with state abstractions, in which aspects of the state
space are ignored. In previous work, we developed the MAXQ method for
hierarchical RL. In this paper, we define five conditions under which state
abstraction can be combined with the MAXQ value function decomposition. We
prove that the MAXQ-Q learning algorithm converges under these conditions and
show experimentally that state abstraction is important for the successful
application of MAXQ-Q learning.
|
112,549 | 112,549 | Cascaded Grammatical Relation Assignment | In this paper we discuss cascaded Memory-Based grammatical relations
assignment. In the first stages of the cascade, we find chunks of several types
(NP,VP,ADJP,ADVP,PP) and label them with their adverbial function (e.g. local,
temporal). In the last stage, we assign grammatical relations to pairs of
chunks. We studied the effect of adding several levels to this cascaded
classifier and we found that even the less performing chunkers enhanced the
performance of the relation finder.
|
112,550 | 112,550 | Memory-Based Shallow Parsing | We present a memory-based learning (MBL) approach to shallow parsing in which
POS tagging, chunking, and identification of syntactic relations are formulated
as memory-based modules. The experiments reported in this paper show
competitive results, the F-value for the Wall Street Journal (WSJ) treebank is:
93.8% for NP chunking, 94.7% for VP chunking, 77.1% for subject detection and
79.0% for object detection.
|
112,551 | 112,551 | Automatically Selecting Useful Phrases for Dialogue Act Tagging | We present an empirical investigation of various ways to automatically
identify phrases in a tagged corpus that are useful for dialogue act tagging.
We found that a new method (which measures a phrase's deviation from an
optimally-predictive phrase), enhanced with a lexical filtering mechanism,
produces significantly better cues than manually-selected cue phrases, the
exhaustive set of phrases in a training corpus, and phrases chosen by
traditional metrics, like mutual information and information gain.
|
112,552 | 112,552 | MAP Lexicon is useful for segmentation and word discovery in
child-directed speech | Because of rather fundamental changes to the underlying model proposed in the
paper, it has been withdrawn from the archive.
|
112,553 | 112,553 | Collective Intelligence for Control of Distributed Dynamical Systems | We consider the El Farol bar problem, also known as the minority game (W. B.
Arthur, ``The American Economic Review'', 84(2): 406--411 (1994), D. Challet
and Y.C. Zhang, ``Physica A'', 256:514 (1998)). We view it as an instance of
the general problem of how to configure the nodal elements of a distributed
dynamical system so that they do not ``work at cross purposes'', in that their
collective dynamics avoids frustration and thereby achieves a provided global
goal. We summarize a mathematical theory for such configuration applicable when
(as in the bar problem) the global goal can be expressed as minimizing a global
energy function and the nodes can be expressed as minimizers of local free
energy functions. We show that a system designed with that theory performs
nearly optimally for the bar problem.
|
112,554 | 112,554 | An Introduction to Collective Intelligence | This paper surveys the emerging science of how to design a ``COllective
INtelligence'' (COIN). A COIN is a large multi-agent system where:
(i) There is little to no centralized communication or control; and
(ii) There is a provided world utility function that rates the possible
histories of the full system.
In particular, we are interested in COINs in which each agent runs a
reinforcement learning (RL) algorithm. Rather than use a conventional modeling
approach (e.g., model the system dynamics, and hand-tune agents to cooperate),
we aim to solve the COIN design problem implicitly, via the ``adaptive''
character of the RL algorithms of each of the agents. This approach introduces
an entirely new, profound design problem: Assuming the RL algorithms are able
to achieve high rewards, what reward functions for the individual agents will,
when pursued by those agents, result in high world utility? In other words,
what reward functions will best ensure that we do not have phenomena like the
tragedy of the commons, Braess's paradox, or the liquidity trap?
Although still very young, research specifically concentrating on the COIN
design problem has already resulted in successes in artificial domains, in
particular in packet-routing, the leader-follower problem, and in variants of
Arthur's El Farol bar problem. It is expected that as it matures and draws upon
other disciplines related to COINs, this research will greatly expand the range
of tasks addressable by human engineers. Moreover, in addition to drawing on
them, such a fully developed scie nce of COIN design may provide much insight
into other already established scientific fields, such as economics, game
theory, and population biology.
|
112,555 | 112,555 | A statistical model for word discovery in child directed speech | A statistical model for segmentation and word discovery in child directed
speech is presented. An incremental unsupervised learning algorithm to infer
word boundaries based on this model is described and results of empirical tests
showing that the algorithm is competitive with other models that have been used
for similar tasks are also presented.
|
112,556 | 112,556 | New Error Bounds for Solomonoff Prediction | Solomonoff sequence prediction is a scheme to predict digits of binary
strings without knowing the underlying probability distribution. We call a
prediction scheme informed when it knows the true probability distribution of
the sequence. Several new relations between universal Solomonoff sequence
prediction and informed prediction and general probabilistic prediction schemes
will be proved. Among others, they show that the number of errors in Solomonoff
prediction is finite for computable distributions, if finite in the informed
case. Deterministic variants will also be studied. The most interesting result
is that the deterministic variant of Solomonoff prediction is optimal compared
to any other probabilistic or deterministic prediction scheme apart from
additive square root corrections only. This makes it well suited even for
difficult prediction problems, where it does not suffice when the number of
errors is minimal to within some factor greater than one. Solomonoff's original
bound and the ones presented here complement each other in a useful way.
|
112,557 | 112,557 | HMM Specialization with Selective Lexicalization | We present a technique which complements Hidden Markov Models by
incorporating some lexicalized states representing syntactically uncommon
words. Our approach examines the distribution of transitions, selects the
uncommon words, and makes lexicalized states for the words. We performed a
part-of-speech tagging experiment on the Brown corpus to evaluate the resultant
language model and discovered that this technique improved the tagging accuracy
by 0.21% at the 95% level of confidence.
|
112,558 | 112,558 | Algorithmic Statistics | While Kolmogorov complexity is the accepted absolute measure of information
content of an individual finite object, a similarly absolute notion is needed
for the relation between an individual data sample and an individual model
summarizing the information in the data, for example, a finite set (or
probability distribution) where the data sample typically came from. The
statistical theory based on such relations between individual objects can be
called algorithmic statistics, in contrast to classical statistical theory that
deals with relations between probabilistic ensembles. We develop the
algorithmic theory of statistic, sufficient statistic, and minimal sufficient
statistic. This theory is based on two-part codes consisting of the code for
the statistic (the model summarizing the regularity, the meaningful
information, in the data) and the model-to-data code. In contrast to the
situation in probabilistic statistical theory, the algorithmic relation of
(minimal) sufficiency is an absolute relation between the individual model and
the individual data sample. We distinguish implicit and explicit descriptions
of the models. We give characterizations of algorithmic (Kolmogorov) minimal
sufficient statistic for all data samples for both description modes--in the
explicit mode under some constraints. We also strengthen and elaborate earlier
results on the ``Kolmogorov structure function'' and ``absolutely
non-stochastic objects''--those rare objects for which the simplest models that
summarize their relevant information (minimal sufficient statistics) are at
least as complex as the objects themselves. We demonstrate a close relation
between the probabilistic notions and the algorithmic ones.
|
112,559 | 112,559 | Learning Complexity Dimensions for a Continuous-Time Control System | This paper takes a computational learning theory approach to a problem of
linear systems identification. It is assumed that input signals have only a
finite number k of frequency components, and systems to be identified have
dimension no greater than n. The main result establishes that the sample
complexity needed for identification scales polynomially with n and
logarithmically with k.
|
112,560 | 112,560 | Mathematics of learning | We study the convergence properties of a pair of learning algorithms
(learning with and without memory). This leads us to study the dominant
eigenvalue of a class of random matrices. This turns out to be related to the
roots of the derivative of random polynomials (generated by picking their roots
uniformly at random in the interval [0, 1], although our results extend to
other distributions). This, in turn, requires the study of the statistical
behavior of the harmonic mean of random variables as above, which leads us to
delicate question of the rate of convergence to stable laws and tail estimates
for stable laws. The reader can find the proofs of most of the results
announced here in the paper entitled "Harmonic mean, random polynomials, and
random matrices", by the same authors.
|
112,561 | 112,561 | Harmonic mean, random polynomials and stochastic matrices | Motivated by a problem in learning theory, we are led to study the dominant
eigenvalue of a class of random matrices. This turns out to be related to the
roots of the derivative of random polynomials (generated by picking their roots
uniformly at random in the interval [0, 1], although our results extend to
other distributions). This, in turn, requires the study of the statistical
behavior of the harmonic mean of random variables as above, and that, in turn,
leads us to delicate question of the rate of convergence to stable laws and
tail estimates for stable laws.
|
112,562 | 112,562 | Robust Estimators under the Imprecise Dirichlet Model | Walley's Imprecise Dirichlet Model (IDM) for categorical data overcomes
several fundamental problems which other approaches to uncertainty suffer from.
Yet, to be useful in practice, one needs efficient ways for computing the
imprecise=robust sets or intervals. The main objective of this work is to
derive exact, conservative, and approximate, robust and credible interval
estimates under the IDM for a large class of statistical estimators, including
the entropy and mutual information.
|
112,563 | 112,563 | A tutorial introduction to the minimum description length principle | This tutorial provides an overview of and introduction to Rissanen's Minimum
Description Length (MDL) Principle. The first chapter provides a conceptual,
entirely non-technical introduction to the subject. It serves as a basis for
the technical introduction given in the second chapter, in which all the ideas
of the first chapter are made mathematically precise. The main ideas are
discussed in great conceptual and technical detail. This tutorial is an
extended version of the first two chapters of the collection "Advances in
Minimum Description Length: Theory and Application" (edited by P.Grunwald, I.J.
Myung and M. Pitt, to be published by the MIT Press, Spring 2005).
|
112,564 | 112,564 | Suboptimal behaviour of Bayes and MDL in classification under
misspecification | We show that forms of Bayesian and MDL inference that are often applied to
classification problems can be *inconsistent*. This means there exists a
learning problem such that for all amounts of data the generalization errors of
the MDL classifier and the Bayes classifier relative to the Bayesian posterior
both remain bounded away from the smallest achievable generalization error.
|
112,565 | 112,565 | Learning a Machine for the Decision in a Partially Observable Markov
Universe | In this paper, we are interested in optimal decisions in a partially
observable Markov universe. Our viewpoint departs from the dynamic programming
viewpoint: we are directly approximating an optimal strategic tree depending on
the observation. This approximation is made by means of a parameterized
probabilistic law. In this paper, a particular family of hidden Markov models,
with input and output, is considered as a learning framework. A method for
optimizing the parameters of these HMMs is proposed and applied. This
optimization method is based on the cross-entropic principle.
|
112,566 | 112,566 | Fast Non-Parametric Bayesian Inference on Infinite Trees | Given i.i.d. data from an unknown distribution, we consider the problem of
predicting future items. An adaptive way to estimate the probability density is
to recursively subdivide the domain to an appropriate data-dependent
granularity. A Bayesian would assign a data-independent prior probability to
"subdivide", which leads to a prior over infinite(ly many) trees. We derive an
exact, fast, and simple inference algorithm for such a prior, for the data
evidence, the predictive distribution, the effective model dimension, and other
quantities.
|
112,567 | 112,567 | Strong Asymptotic Assertions for Discrete MDL in Regression and
Classification | We study the properties of the MDL (or maximum penalized complexity)
estimator for Regression and Classification, where the underlying model class
is countable. We show in particular a finite bound on the Hellinger losses
under the only assumption that there is a "true" model contained in the class.
This implies almost sure convergence of the predictive distribution to the true
one at a fast rate. It corresponds to Solomonoff's central theorem of universal
induction, however with a bound that is exponentially larger.
|
112,568 | 112,568 | Combinations and Mixtures of Optimal Policies in Unichain Markov
Decision Processes are Optimal | We show that combinations of optimal (stationary) policies in unichain Markov
decision processes are optimal. That is, let M be a unichain Markov decision
process with state space S, action space A and policies \pi_j^*: S -> A (1\leq
j\leq n) with optimal average infinite horizon reward. Then any combination \pi
of these policies, where for each state i in S there is a j such that
\pi(i)=\pi_j^*(i), is optimal as well. Furthermore, we prove that any mixture
of optimal policies, where at each visit in a state i an arbitrary action
\pi_j^*(i) of an optimal policy is chosen, yields optimal average reward, too.
|
112,569 | 112,569 | MDL Convergence Speed for Bernoulli Sequences | The Minimum Description Length principle for online sequence
estimation/prediction in a proper learning setup is studied. If the underlying
model class is discrete, then the total expected square loss is a particularly
interesting performance measure: (a) this quantity is finitely bounded,
implying convergence with probability one, and (b) it additionally specifies
the convergence speed. For MDL, in general one can only have loss bounds which
are finite but exponentially larger than those for Bayes mixtures. We show that
this is even the case if the model class contains only Bernoulli distributions.
We derive a new upper bound on the prediction error for countable Bernoulli
classes. This implies a small bound (comparable to the one for Bayes mixtures)
for certain important model classes. We discuss the application to Machine
Learning tasks such as classification and hypothesis testing, and
generalization to countable classes of i.i.d. models.
|
112,570 | 112,570 | Generalization error bounds in semi-supervised classification under the
cluster assumption | We consider semi-supervised classification when part of the available data is
unlabeled. These unlabeled data can be useful for the classification problem
when we make an assumption relating the behavior of the regression function to
that of the marginal distribution. Seeger (2000) proposed the well-known
"cluster assumption" as a reasonable one. We propose a mathematical formulation
of this assumption and a method based on density level sets estimation that
takes advantage of it to achieve fast rates of convergence both in the number
of unlabeled examples and the number of labeled examples.
|
112,571 | 112,571 | Cross-Entropic Learning of a Machine for the Decision in a Partially
Observable Universe | Revision of the paper previously entitled "Learning a Machine for the
Decision in a Partially Observable Markov Universe" In this paper, we are
interested in optimal decisions in a partially observable universe. Our
approach is to directly approximate an optimal strategic tree depending on the
observation. This approximation is made by means of a parameterized
probabilistic law. A particular family of hidden Markov models, with input
\emph{and} output, is considered as a model of policy. A method for optimizing
the parameters of these HMMs is proposed and applied. This optimization is
based on the cross-entropic principle for rare events simulation developed by
Rubinstein.
|
112,572 | 112,572 | Bayesian Regression of Piecewise Constant Functions | We derive an exact and efficient Bayesian regression algorithm for piecewise
constant functions of unknown segment number, boundary location, and levels. It
works for any noise and segment level prior, e.g. Cauchy which can handle
outliers. We derive simple but good estimates for the in-segment variance. We
also propose a Bayesian regression curve as a better way of smoothing data
without blurring boundaries. The Bayesian approach also allows straightforward
determination of the evidence, break probabilities and error estimates, useful
for model selection and significance and robustness studies. We discuss the
performance on synthetic and real-world examples. Many possible extensions will
be discussed.
|
112,573 | 112,573 | Entropy And Vision | In vector quantization the number of vectors used to construct the codebook
is always an undefined problem, there is always a compromise between the number
of vectors and the quantity of information lost during the compression. In this
text we present a minimum of Entropy principle that gives solution to this
compromise and represents an Entropy point of view of signal compression in
general. Also we present a new adaptive Object Quantization technique that is
the same for the compression and the perception.
|
112,574 | 112,574 | Graph Laplacians and their convergence on random neighborhood graphs | Given a sample from a probability measure with support on a submanifold in
Euclidean space one can construct a neighborhood graph which can be seen as an
approximation of the submanifold. The graph Laplacian of such a graph is used
in several machine learning methods like semi-supervised learning,
dimensionality reduction and clustering. In this paper we determine the
pointwise limit of three different graph Laplacians used in the literature as
the sample size increases and the neighborhood size approaches zero. We show
that for a uniform measure on the submanifold all graph Laplacians have the
same limit up to constants. However in the case of a non-uniform measure on the
submanifold only the so called random walk graph Laplacian converges to the
weighted Laplace-Beltrami operator.
|
112,575 | 112,575 | Occam's hammer: a link between randomized learning and multiple testing
FDR control | We establish a generic theoretical tool to construct probabilistic bounds for
algorithms where the output is a subset of objects from an initial pool of
candidates (or more generally, a probability distribution on said pool). This
general device, dubbed "Occam's hammer'', acts as a meta layer when a
probabilistic bound is already known on the objects of the pool taken
individually, and aims at controlling the proportion of the objects in the set
output not satisfying their individual bound. In this regard, it can be seen as
a non-trivial generalization of the "union bound with a prior'' ("Occam's
razor''), a familiar tool in learning theory. We give applications of this
principle to randomized classifiers (providing an interesting alternative
approach to PAC-Bayes bounds) and multiple testing (where it allows to retrieve
exactly and extend the so-called Benjamini-Yekutieli testing procedure).
|
112,576 | 112,576 | Cross-Entropy method: convergence issues for extended implementation | The cross-entropy method (CE) developed by R. Rubinstein is an elegant
practical principle for simulating rare events. The method approximates the
probability of the rare event by means of a family of probabilistic models. The
method has been extended to optimization, by considering an optimal event as a
rare event. CE works rather good when dealing with deterministic function
optimization. Now, it appears that two conditions are needed for a good
convergence of the method. First, it is necessary to have a family of models
sufficiently flexible for discriminating the optimal events. Indirectly, it
appears also that the function to be optimized should be deterministic. The
purpose of this paper is to consider the case of partially discriminating model
family, and of stochastic functions. It will be shown on simple examples that
the CE could fail when relaxing these hypotheses. Alternative improvements of
the CE method are investigated and compared on random examples in order to
handle this issue.
|
112,577 | 112,577 | Strategies for prediction under imperfect monitoring | We propose simple randomized strategies for sequential prediction under
imperfect monitoring, that is, when the forecaster does not have access to the
past outcomes but rather to a feedback signal. The proposed strategies are
consistent in the sense that they achieve, asymptotically, the best possible
average reward. It was Rustichini (1999) who first proved the existence of such
consistent predictors. The forecasters presented here offer the first
constructive proof of consistency. Moreover, the proposed algorithms are
computationally efficient. We also establish upper bounds for the rates of
convergence. In the case of deterministic feedback, these rates are optimal up
to logarithmic terms.
|
112,578 | 112,578 | The Loss Rank Principle for Model Selection | We introduce a new principle for model selection in regression and
classification. Many regression models are controlled by some smoothness or
flexibility or complexity parameter c, e.g. the number of neighbors to be
averaged over in k nearest neighbor (kNN) regression or the polynomial degree
in regression with polynomials. Let f_D^c be the (best) regressor of complexity
c on data D. A more flexible regressor can fit more data D' well than a more
rigid one. If something (here small loss) is easy to achieve it's typically
worth less. We define the loss rank of f_D^c as the number of other
(fictitious) data D' that are fitted better by f_D'^c than D is fitted by
f_D^c. We suggest selecting the model complexity c that has minimal loss rank
(LoRP). Unlike most penalized maximum likelihood variants (AIC,BIC,MDL), LoRP
only depends on the regression function and loss function. It works without a
stochastic noise model, and is directly applicable to any non-parametric
regressor, like kNN. In this paper we formalize, discuss, and motivate LoRP,
study it for specific regression problems, in particular linear ones, and
compare it to other model selection schemes.
|
112,579 | 112,579 | Information Bottlenecks, Causal States, and Statistical Relevance Bases:
How to Represent Relevant Information in Memoryless Transduction | Discovering relevant, but possibly hidden, variables is a key step in
constructing useful and predictive theories about the natural world. This brief
note explains the connections between three approaches to this problem: the
recently introduced information-bottleneck method, the computational mechanics
approach to inferring optimal models, and Salmon's statistical relevance basis.
|
112,580 | 112,580 | A Model for Prejudiced Learning in Noisy Environments | Based on the heuristics that maintaining presumptions can be beneficial in
uncertain environments, we propose a set of basic axioms for learning systems
to incorporate the concept of prejudice. The simplest, memoryless model of a
deterministic learning rule obeying the axioms is constructed, and shown to be
equivalent to the logistic map. The system's performance is analysed in an
environment in which it is subject to external randomness, weighing learning
defectiveness against stability gained. The corresponding random dynamical
system with inhomogeneous, additive noise is studied, and shown to exhibit the
phenomena of noise induced stability and stochastic bifurcations. The overall
results allow for the interpretation that prejudice in uncertain environments
entails a considerable portion of stubbornness as a secondary phenomenon.
|
112,581 | 112,581 | Metrics for more than two points at once | The conventional definition of a topological metric over a space specifies
properties that must be obeyed by any measure of "how separated" two points in
that space are. Here it is shown how to extend that definition, and in
particular the triangle inequality, to concern arbitrary numbers of points.
Such a measure of how separated the points within a collection are can be
bootstrapped, to measure "how separated" from each other are two (or more)
collections. The measure presented here also allows fractional membership of an
element in a collection. This means it directly concerns measures of ``how
spread out" a probability distribution over a space is. When such a measure is
bootstrapped to compare two collections, it allows us to measure how separated
two probability distributions are, or more generally, how separated a
distribution of distributions is.
|
112,582 | 112,582 | Combinatorial Approach to Object Analysis | We present a perceptional mathematical model for image and signal analysis. A
resemblance measure is defined, and submitted to an innovating combinatorial
optimization algorithm. Numerical Simulations are also presented
|
112,583 | 112,583 | The information bottleneck method | We define the relevant information in a signal $x\in X$ as being the
information that this signal provides about another signal $y\in \Y$. Examples
include the information that face images provide about the names of the people
portrayed, or the information that speech sounds provide about the words
spoken. Understanding the signal $x$ requires more than just predicting $y$, it
also requires specifying which features of $\X$ play a role in the prediction.
We formalize this problem as that of finding a short code for $\X$ that
preserves the maximum information about $\Y$. That is, we squeeze the
information that $\X$ provides about $\Y$ through a `bottleneck' formed by a
limited set of codewords $\tX$. This constrained optimization problem can be
seen as a generalization of rate distortion theory in which the distortion
measure $d(x,\x)$ emerges from the joint statistics of $\X$ and $\Y$. This
approach yields an exact set of self consistent equations for the coding rules
$X \to \tX$ and $\tX \to \Y$. Solutions to these equations can be found by a
convergent re-estimation method that generalizes the Blahut-Arimoto algorithm.
Our variational principle provides a surprisingly rich framework for discussing
a variety of problems in signal processing and learning, as will be described
in detail elsewhere.
|
112,584 | 112,584 | Predictability, complexity and learning | We define {\em predictive information} $I_{\rm pred} (T)$ as the mutual
information between the past and the future of a time series. Three
qualitatively different behaviors are found in the limit of large observation
times $T$: $I_{\rm pred} (T)$ can remain finite, grow logarithmically, or grow
as a fractional power law. If the time series allows us to learn a model with a
finite number of parameters, then $I_{\rm pred} (T)$ grows logarithmically with
a coefficient that counts the dimensionality of the model space. In contrast,
power--law growth is associated, for example, with the learning of infinite
parameter (or nonparametric) models such as continuous functions with
smoothness constraints. There are connections between the predictive
information and measures of complexity that have been defined both in learning
theory and in the analysis of physical systems through statistical mechanics
and dynamical systems theory. Further, in the same way that entropy provides
the unique measure of available information consistent with some simple and
plausible conditions, we argue that the divergent part of $I_{\rm pred} (T)$
provides the unique measure for the complexity of dynamics underlying a time
series. Finally, we discuss how these ideas may be useful in different problems
in physics, statistics, and biology.
|
112,585 | 112,585 | Information theory and learning: a physical approach | We try to establish a unified information theoretic approach to learning and
to explore some of its applications. First, we define {\em predictive
information} as the mutual information between the past and the future of a
time series, discuss its behavior as a function of the length of the series,
and explain how other quantities of interest studied previously in learning
theory - as well as in dynamical systems and statistical mechanics - emerge
from this universally definable concept. We then prove that predictive
information provides the {\em unique measure for the complexity} of dynamics
underlying the time series and show that there are classes of models
characterized by {\em power-law growth of the predictive information} that are
qualitatively more complex than any of the systems that have been investigated
before. Further, we investigate numerically the learning of a nonparametric
probability density, which is an example of a problem with power-law
complexity, and show that the proper Bayesian formulation of this problem
provides for the `Occam' factors that punish overly complex models and thus
allow one {\em to learn not only a solution within a specific model class, but
also the class itself} using the data only and with very few a priori
assumptions. We study a possible {\em information theoretic method} that
regularizes the learning of an undersampled discrete variable, and show that
learning in such a setup goes through stages of very different complexities.
Finally, we discuss how all of these ideas may be useful in various problems in
physics, statistics, and, most importantly, biology.
|
112,586 | 112,586 | Structural Inference of Hierarchies in Networks | One property of networks that has received comparatively little attention is
hierarchy, i.e., the property of having vertices that cluster together in
groups, which then join to form groups of groups, and so forth, up through all
levels of organization in the network. Here, we give a precise definition of
hierarchical structure, give a generic model for generating arbitrary
hierarchical structure in a random graph, and describe a statistically
principled way to learn the set of hierarchical features that most plausibly
explain a particular real-world network. By applying this approach to two
example networks, we demonstrate its advantages for the interpretation of
network data, the annotation of graphs with edge, vertex and community
properties, and the generation of generic null models for further hypothesis
testing.
|
112,587 | 112,587 | Parametric Inference for Biological Sequence Analysis | One of the major successes in computational biology has been the unification,
using the graphical model formalism, of a multitude of algorithms for
annotating and comparing biological sequences. Graphical models that have been
applied towards these problems include hidden Markov models for annotation,
tree models for phylogenetics, and pair hidden Markov models for alignment. A
single algorithm, the sum-product algorithm, solves many of the inference
problems associated with different statistical models. This paper introduces
the \emph{polytope propagation algorithm} for computing the Newton polytope of
an observation from a graphical model. This algorithm is a geometric version of
the sum-product algorithm and is used to analyze the parametric behavior of
maximum a posteriori inference calculations for graphical models.
|
112,588 | 112,588 | Fluctuation-dissipation theorem and models of learning | Advances in statistical learning theory have resulted in a multitude of
different designs of learning machines. But which ones are implemented by
brains and other biological information processors? We analyze how various
abstract Bayesian learners perform on different data and argue that it is
difficult to determine which learning-theoretic computation is performed by a
particular organism using just its performance in learning a stationary target
(learning curve). Basing on the fluctuation-dissipation relation in statistical
physics, we then discuss a different experimental setup that might be able to
solve the problem.
|
112,589 | 112,589 | Metric learning pairwise kernel for graph inference | Much recent work in bioinformatics has focused on the inference of various
types of biological networks, representing gene regulation, metabolic
processes, protein-protein interactions, etc. A common setting involves
inferring network edges in a supervised fashion from a set of high-confidence
edges, possibly characterized by multiple, heterogeneous data sets (protein
sequence, gene expression, etc.). Here, we distinguish between two modes of
inference in this setting: direct inference based upon similarities between
nodes joined by an edge, and indirect inference based upon similarities between
one pair of nodes and another pair of nodes. We propose a supervised approach
for the direct case by translating it into a distance metric learning problem.
A relaxation of the resulting convex optimization problem leads to the support
vector machine (SVM) algorithm with a particular kernel for pairs, which we
call the metric learning pairwise kernel (MLPK). We demonstrate, using several
real biological networks, that this direct approach often improves upon the
state-of-the-art SVM for indirect inference with the tensor product pairwise
kernel.
|
112,590 | 112,590 | Algorithmic Theories of Everything | The probability distribution P from which the history of our universe is
sampled represents a theory of everything or TOE. We assume P is formally
describable. Since most (uncountably many) distributions are not, this imposes
a strong inductive bias. We show that P(x) is small for any universe x lacking
a short description, and study the spectrum of TOEs spanned by two Ps, one
reflecting the most compact constructive descriptions, the other the fastest
way of computing everything. The former derives from generalizations of
traditional computability, Solomonoff's algorithmic probability, Kolmogorov
complexity, and objects more random than Chaitin's Omega, the latter from
Levin's universal search and a natural resource-oriented postulate: the
cumulative prior probability of all x incomputable within time t by this
optimal algorithm should be 1/t. Between both Ps we find a universal
cumulatively enumerable measure that dominates traditional enumerable measures;
any such CEM must assign low probability to any universe lacking a short
enumerating program. We derive P-specific consequences for evolving observers,
inductive reasoning, quantum physics, philosophy, and the expected duration of
our universe.
|
112,591 | 112,591 | Improved Bounds on Quantum Learning Algorithms | In this article we give several new results on the complexity of algorithms
that learn Boolean functions from quantum queries and quantum examples.
Hunziker et al. conjectured that for any class C of Boolean functions, the
number of quantum black-box queries which are required to exactly identify an
unknown function from C is $O(\frac{\log |C|}{\sqrt{{\hat{\gamma}}^{C}}})$,
where $\hat{\gamma}^{C}$ is a combinatorial parameter of the class C. We
essentially resolve this conjecture in the affirmative by giving a quantum
algorithm that, for any class C, identifies any unknown function from C using
$O(\frac{\log |C| \log \log |C|}{\sqrt{{\hat{\gamma}}^{C}}})$ quantum black-box
queries.
We consider a range of natural problems intermediate between the exact
learning problem (in which the learner must obtain all bits of information
about the black-box function) and the usual problem of computing a predicate
(in which the learner must obtain only one bit of information about the
black-box function). We give positive and negative results on when the quantum
and classical query complexities of these intermediate problems are
polynomially related to each other.
Finally, we improve the known lower bounds on the number of quantum examples
(as opposed to quantum black-box queries) required for $(\epsilon,\delta)$-PAC
learning any concept class of Vapnik-Chervonenkis dimension d over the domain
$\{0,1\}^n$ from $\Omega(\frac{d}{n})$ to $\Omega(\frac{1}{\epsilon}\log
\frac{1}{\delta}+d+\frac{\sqrt{d}}{\epsilon})$. This new lower bound comes
closer to matching known upper bounds for classical PAC learning.
|
0 | null | ZSON: Zero-Shot Object-Goal Navigation using Multimodal Goal Embeddings | We present a scalable approach for learning open-world object-goal navigation
(ObjectNav) -- the task of asking a virtual robot (agent) to find any instance
of an object in an unexplored environment (e.g., "find a sink"). Our approach
is entirely zero-shot -- i.e., it does not require ObjectNav rewards or
demonstrations of any kind. Instead, we train on the image-goal navigation
(ImageNav) task, in which agents find the location where a picture (i.e., goal
image) was captured. Specifically, we encode goal images into a multimodal,
semantic embedding space to enable training semantic-goal navigation
(SemanticNav) agents at scale in unannotated 3D environments (e.g., HM3D).
After training, SemanticNav agents can be instructed to find objects described
in free-form natural language (e.g., "sink", "bathroom sink", etc.) by
projecting language goals into the same multimodal, semantic embedding space.
As a result, our approach enables open-world ObjectNav. We extensively evaluate
our agents on three ObjectNav datasets (Gibson, HM3D, and MP3D) and observe
absolute improvements in success of 4.2% - 20.0% over existing zero-shot
methods. For reference, these gains are similar or better than the 5%
improvement in success between the Habitat 2020 and 2021 ObjectNav challenge
winners. In an open-world setting, we discover that our agents can generalize
to compound instructions with a room explicitly mentioned (e.g., "Find a
kitchen sink") and when the target room can be inferred (e.g., "Find a sink and
a stove"). |
1 | null | Predicting the Stability of Hierarchical Triple Systems with Convolutional Neural Networks | Understanding the long-term evolution of hierarchical triple systems is
challenging due to its inherent chaotic nature, and it requires computationally
expensive simulations. Here we propose a convolutional neural network model to
predict the stability of hierarchical triples by looking at their evolution
during the first $5 \times 10^5$ inner binary orbits. We employ the regularized
few-body code \textsc{tsunami} to simulate $5\times 10^6$ hierarchical triples,
from which we generate a large training and test dataset. We develop twelve
different network configurations that use different combinations of the
triples' orbital elements and compare their performances. Our best model uses 6
time-series, namely, the semimajor axes ratio, the inner and outer
eccentricities, the mutual inclination and the arguments of pericenter. This
model achieves an area under the curve of over $95\%$ and informs of the
relevant parameters to study triple systems stability. All trained models are
made publicly available, allowing to predict the stability of hierarchical
triple systems $200$ times faster than pure $N$-body methods. |
2 | null | Debiasing Learning for Membership Inference Attacks Against Recommender Systems | Learned recommender systems may inadvertently leak information about their
training data, leading to privacy violations. We investigate privacy threats
faced by recommender systems through the lens of membership inference. In such
attacks, an adversary aims to infer whether a user's data is used to train the
target recommender. To achieve this, previous work has used a shadow
recommender to derive training data for the attack model, and then predicts the
membership by calculating difference vectors between users' historical
interactions and recommended items. State-of-the-art methods face two
challenging problems: (1) training data for the attack model is biased due to
the gap between shadow and target recommenders, and (2) hidden states in
recommenders are not observational, resulting in inaccurate estimations of
difference vectors. To address the above limitations, we propose a Debiasing
Learning for Membership Inference Attacks against recommender systems (DL-MIA)
framework that has four main components: (1) a difference vector generator, (2)
a disentangled encoder, (3) a weight estimator, and (4) an attack model. To
mitigate the gap between recommenders, a variational auto-encoder (VAE) based
disentangled encoder is devised to identify recommender invariant and specific
features. To reduce the estimation bias, we design a weight estimator,
assigning a truth-level score for each difference vector to indicate estimation
accuracy. We evaluate DL-MIA against both general recommenders and sequential
recommenders on three real-world datasets. Experimental results show that
DL-MIA effectively alleviates training and estimation biases simultaneously,
and achieves state-of-the-art attack performance. |
3 | null | Data Leakage in Federated Averaging | Recent attacks have shown that user data can be reconstructed from FedSGD
updates, thus breaking privacy. However, these attacks are of limited practical
relevance as federated learning typically uses the FedAvg algorithm. It is
generally accepted that reconstructing data from FedAvg updates is much harder
than FedSGD as: (i) there are unobserved intermediate weight updates, (ii) the
order of inputs matters, and (iii) the order of labels changes every epoch. In
this work, we propose a new optimization-based attack which successfully
attacks FedAvg by addressing the above challenges. First, we solve the
optimization problem using automatic differentiation that forces a simulation
of the client's update for the reconstructed labels and inputs so as to match
the received client update. Second, we address the unknown input order by
treating images at different epochs as independent during optimization, while
relating them with a permutation invariant prior. Third, we reconstruct the
labels by estimating the parameters of existing FedSGD attacks at every FedAvg
step. On the popular FEMNIST dataset, we demonstrate that on average we
successfully reconstruct >45% of the client's images from realistic FedAvg
updates computed on 10 local epochs of 10 batches each with 5 images, compared
to only <10% using the baseline. These findings indicate that many real-world
federated learning implementations based on FedAvg are vulnerable. |
4 | null | On Certifying and Improving Generalization to Unseen Domains | Domain Generalization (DG) aims to learn models whose performance remains
high on unseen domains encountered at test-time by using data from multiple
related source domains. Many existing DG algorithms reduce the divergence
between source distributions in a representation space to potentially align the
unseen domain close to the sources. This is motivated by the analysis that
explains generalization to unseen domains using distributional distance (such
as the Wasserstein distance) to the sources. However, due to the openness of
the DG objective, it is challenging to evaluate DG algorithms comprehensively
using a few benchmark datasets. In particular, we demonstrate that the accuracy
of the models trained with DG methods varies significantly across unseen
domains, generated from popular benchmark datasets. This highlights that the
performance of DG methods on a few benchmark datasets may not be representative
of their performance on unseen domains in the wild. To overcome this roadblock,
we propose a universal certification framework based on distributionally robust
optimization (DRO) that can efficiently certify the worst-case performance of
any DG method. This enables a data-independent evaluation of a DG method
complementary to the empirical evaluations on benchmark datasets. Furthermore,
we propose a training algorithm that can be used with any DG method to provably
improve their certified performance. Our empirical evaluation demonstrates the
effectiveness of our method at significantly improving the worst-case loss
(i.e., reducing the risk of failure of these models in the wild) without
incurring a significant performance drop on benchmark datasets. |
5 | null | From Tensor Network Quantum States to Tensorial Recurrent Neural Networks | We show that any matrix product state (MPS) can be exactly represented by a
recurrent neural network (RNN) with a linear memory update. We generalize this
RNN architecture to 2D lattices using a multilinear memory update. It supports
perfect sampling and wave function evaluation in polynomial time, and can
represent an area law of entanglement entropy. Numerical evidence shows that it
can encode the wave function using a bond dimension lower by orders of
magnitude when compared to MPS, with an accuracy that can be systematically
improved by increasing the bond dimension. |
6 | null | Out of distribution robustness with pre-trained Bayesian neural networks | We develop ShiftMatch, a new training-data-dependent likelihood for out of
distribution (OOD) robustness in Bayesian neural networks (BNNs). ShiftMatch is
inspired by the training-data-dependent "EmpCov" priors from Izmailov et al.
(2021a) and efficiently matches test-time spatial correlations to those at
training time. Critically, ShiftMatch is designed to leave neural network
training unchanged, allowing it to use publically available samples from
pretrained BNNs. Using pre-trained HMC samples, ShiftMatch gives strong
performance improvements on CIFAR-10-C, outperforms EmpCov priors, and is
perhaps the first Bayesian method capable of convincingly outperforming plain
deep ensembles. ShiftMatch can be integrated with non-Bayesian methods like
deep ensembles, where it offers smaller, but still considerable, performance
improvements. Overall, Bayesian ShiftMatch gave slightly better accuracy than
ensembles with ShiftMatch, though they both had very similar log-likelihoods. |
7 | null | Quantifying Inherent Randomness in Machine Learning Algorithms | Most machine learning (ML) algorithms have several stochastic elements, and
their performances are affected by these sources of randomness. This paper uses
an empirical study to systematically examine the effects of two sources:
randomness in model training and randomness in the partitioning of a dataset
into training and test subsets. We quantify and compare the magnitude of the
variation in predictive performance for the following ML algorithms: Random
Forests (RFs), Gradient Boosting Machines (GBMs), and Feedforward Neural
Networks (FFNNs). Among the different algorithms, randomness in model training
causes larger variation for FFNNs compared to tree-based methods. This is to be
expected as FFNNs have more stochastic elements that are part of their model
initialization and training. We also found that random splitting of datasets
leads to higher variation compared to the inherent randomness from model
training. The variation from data splitting can be a major issue if the
original dataset has considerable heterogeneity.
Keywords: Model Training, Reproducibility, Variation |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.