title
stringlengths 7
246
| abstract
stringlengths 6
3.31k
|
|---|---|
Learning to encode motion using spatio-temporal synchrony | We consider the task of learning to extract motion from videos. To this end,
we show that the detection of spatial transformations can be viewed as the
detection of synchrony between the image sequence and a sequence of features
undergoing the motion we wish to detect. We show that learning about synchrony
is possible using very fast, local learning rules, by introducing
multiplicative "gating" interactions between hidden units across frames. This
makes it possible to achieve competitive performance in a wide variety of
motion estimation tasks, using a small fraction of the time required to learn
features, and to outperform hand-crafted spatio-temporal features by a large
margin. We also show how learning about synchrony can be viewed as performing
greedy parameter estimation in the well-known motion energy model.
|
Confidence Intervals and Hypothesis Testing for High-Dimensional
Regression | Fitting high-dimensional statistical models often requires the use of
non-linear parameter estimation procedures. As a consequence, it is generally
impossible to obtain an exact characterization of the probability distribution
of the parameter estimates. This in turn implies that it is extremely
challenging to quantify the \emph{uncertainty} associated with a certain
parameter estimate. Concretely, no commonly accepted procedure exists for
computing classical measures of uncertainty and statistical significance as
confidence intervals or $p$-values for these models.
We consider here high-dimensional linear regression problem, and propose an
efficient algorithm for constructing confidence intervals and $p$-values. The
resulting confidence intervals have nearly optimal size. When testing for the
null hypothesis that a certain parameter is vanishing, our method has nearly
optimal power.
Our approach is based on constructing a `de-biased' version of regularized
M-estimators. The new construction improves over recent work in the field in
that it does not assume a special structure on the design matrix. We test our
method on synthetic data and a high-throughput genomic data set about
riboflavin production rate.
|
Bregman Alternating Direction Method of Multipliers | The mirror descent algorithm (MDA) generalizes gradient descent by using a
Bregman divergence to replace squared Euclidean distance. In this paper, we
similarly generalize the alternating direction method of multipliers (ADMM) to
Bregman ADMM (BADMM), which allows the choice of different Bregman divergences
to exploit the structure of problems. BADMM provides a unified framework for
ADMM and its variants, including generalized ADMM, inexact ADMM and Bethe ADMM.
We establish the global convergence and the $O(1/T)$ iteration complexity for
BADMM. In some cases, BADMM can be faster than ADMM by a factor of
$O(n/\log(n))$. In solving the linear program of mass transportation problem,
BADMM leads to massive parallelism and can easily run on GPU. BADMM is several
times faster than highly optimized commercial software Gurobi.
|
Sparse Inverse Covariance Matrix Estimation Using Quadratic
Approximation | The L1-regularized Gaussian maximum likelihood estimator (MLE) has been shown
to have strong statistical guarantees in recovering a sparse inverse covariance
matrix, or alternatively the underlying graph structure of a Gaussian Markov
Random Field, from very limited samples. We propose a novel algorithm for
solving the resulting optimization problem which is a regularized
log-determinant program. In contrast to recent state-of-the-art methods that
largely use first order gradient information, our algorithm is based on
Newton's method and employs a quadratic approximation, but with some
modifications that leverage the structure of the sparse Gaussian MLE problem.
We show that our method is superlinearly convergent, and present experimental
results using synthetic and real-world application data that demonstrate the
considerable improvements in performance of our method when compared to other
state-of-the-art methods.
|
Relaxed Sparse Eigenvalue Conditions for Sparse Estimation via
Non-convex Regularized Regression | Non-convex regularizers usually improve the performance of sparse estimation
in practice. To prove this fact, we study the conditions of sparse estimations
for the sharp concave regularizers which are a general family of non-convex
regularizers including many existing regularizers. For the global solutions of
the regularized regression, our sparse eigenvalue based conditions are weaker
than that of L1-regularization for parameter estimation and sparseness
estimation. For the approximate global and approximate stationary (AGAS)
solutions, almost the same conditions are also enough. We show that the desired
AGAS solutions can be obtained by coordinate descent (CD) based methods.
Finally, we perform some experiments to show the performance of CD methods on
giving AGAS solutions and the degree of weakness of the estimation conditions
required by the sharp concave regularizers.
|
Constrained fractional set programs and their application in local
clustering and community detection | The (constrained) minimization of a ratio of set functions is a problem
frequently occurring in clustering and community detection. As these
optimization problems are typically NP-hard, one uses convex or spectral
relaxations in practice. While these relaxations can be solved globally
optimally, they are often too loose and thus lead to results far away from the
optimum. In this paper we show that every constrained minimization problem of a
ratio of non-negative set functions allows a tight relaxation into an
unconstrained continuous optimization problem. This result leads to a flexible
framework for solving constrained problems in network analysis. While a
globally optimal solution for the resulting non-convex problem cannot be
guaranteed, we outperform the loose convex or spectral relaxations by a large
margin on constrained local clustering problems.
|
Classifying Single-Trial EEG during Motor Imagery with a Small Training
Set | Before the operation of a motor imagery based brain-computer interface (BCI)
adopting machine learning techniques, a cumbersome training procedure is
unavoidable. The development of a practical BCI posed the challenge of
classifying single-trial EEG with a small training set. In this letter, we
addressed this problem by employing a series of signal processing and machine
learning approaches to alleviate overfitting and obtained test accuracy similar
to training accuracy on the datasets from BCI Competition III and our own
experiments.
|
Hyperparameter Optimization and Boosting for Classifying Facial
Expressions: How good can a "Null" Model be? | One of the goals of the ICML workshop on representation and learning is to
establish benchmark scores for a new data set of labeled facial expressions.
This paper presents the performance of a "Null" model consisting of
convolutions with random weights, PCA, pooling, normalization, and a linear
readout. Our approach focused on hyperparameter optimization rather than novel
model components. On the Facial Expression Recognition Challenge held by the
Kaggle website, our hyperparameter optimization approach achieved a score of
60% accuracy on the test data. This paper also introduces a new ensemble
construction variant that combines hyperparameter optimization with the
construction of ensembles. This algorithm constructed an ensemble of four
models that scored 65.5% accuracy. These scores rank 12th and 5th respectively
among the 56 challenge participants. It is worth noting that our approach was
developed prior to the release of the data set, and applied without
modification; our strong competition performance suggests that the TPE
hyperparameter optimization algorithm and domain expertise encoded in our Null
model can generalize to new image classification data sets.
|
Approximation Algorithms for Bayesian Multi-Armed Bandit Problems | In this paper, we consider several finite-horizon Bayesian multi-armed bandit
problems with side constraints which are computationally intractable (NP-Hard)
and for which no optimal (or near optimal) algorithms are known to exist with
sub-exponential running time. All of these problems violate the standard
exchange property, which assumes that the reward from the play of an arm is not
contingent upon when the arm is played. Not only are index policies suboptimal
in these contexts, there has been little analysis of such policies in these
problem settings. We show that if we consider near-optimal policies, in the
sense of approximation algorithms, then there exists (near) index policies.
Conceptually, if we can find policies that satisfy an approximate version of
the exchange property, namely, that the reward from the play of an arm depends
on when the arm is played to within a constant factor, then we have an avenue
towards solving these problems. However such an approximate version of the
idling bandit property does not hold on a per-play basis and are shown to hold
in a global sense. Clearly, such a property is not necessarily true of
arbitrary single arm policies and finding such single arm policies is
nontrivial. We show that by restricting the state spaces of arms we can find
single arm policies and that these single arm policies can be combined into
global (near) index policies where the approximate version of the exchange
property is true in expectation. The number of different bandit problems that
can be addressed by this technique already demonstrate its wide applicability.
|
Outlying Property Detection with Numerical Attributes | The outlying property detection problem is the problem of discovering the
properties distinguishing a given object, known in advance to be an outlier in
a database, from the other database objects. In this paper, we analyze the
problem within a context where numerical attributes are taken into account,
which represents a relevant case left open in the literature. We introduce a
measure to quantify the degree the outlierness of an object, which is
associated with the relative likelihood of the value, compared to the to the
relative likelihood of other objects in the database. As a major contribution,
we present an efficient algorithm to compute the outlierness relative to
significant subsets of the data. The latter subsets are characterized in a
"rule-based" fashion, and hence the basis for the underlying explanation of the
outlierness.
|
Online Alternating Direction Method (longer version) | Online optimization has emerged as powerful tool in large scale optimization.
In this pa- per, we introduce efficient online optimization algorithms based on
the alternating direction method (ADM), which can solve online convex
optimization under linear constraints where the objective could be non-smooth.
We introduce new proof techniques for ADM in the batch setting, which yields a
O(1/T) convergence rate for ADM and forms the basis for regret anal- ysis in
the online setting. We consider two scenarios in the online setting, based on
whether an additional Bregman divergence is needed or not. In both settings, we
establish regret bounds for both the objective function as well as constraints
violation for general and strongly convex functions. We also consider inexact
ADM updates where certain terms are linearized to yield efficient updates and
show the stochastic convergence rates. In addition, we briefly discuss that
online ADM can be used as projection- free online learning algorithm in some
scenarios. Preliminary results are presented to illustrate the performance of
the proposed algorithms.
|
Spectral Experts for Estimating Mixtures of Linear Regressions | Discriminative latent-variable models are typically learned using EM or
gradient-based optimization, which suffer from local optima. In this paper, we
develop a new computationally efficient and provably consistent estimator for a
mixture of linear regressions, a simple instance of a discriminative
latent-variable model. Our approach relies on a low-rank linear regression to
recover a symmetric tensor, which can be factorized into the parameters using a
tensor power method. We prove rates of convergence for our estimator and
provide an empirical evaluation illustrating its strengths relative to local
optimization (EM).
|
Cluster coloring of the Self-Organizing Map: An information
visualization perspective | This paper takes an information visualization perspective to visual
representations in the general SOM paradigm. This involves viewing SOM-based
visualizations through the eyes of Bertin's and Tufte's theories on data
graphics. The regular grid shape of the Self-Organizing Map (SOM), while being
a virtue for linking visualizations to it, restricts representation of cluster
structures. From the viewpoint of information visualization, this paper
provides a general, yet simple, solution to projection-based coloring of the
SOM that reveals structures. First, the proposed color space is easy to
construct and customize to the purpose of use, while aiming at being
perceptually correct and informative through two separable dimensions. Second,
the coloring method is not dependent on any specific method of projection, but
is rather modular to fit any objective function suitable for the task at hand.
The cluster coloring is illustrated on two datasets: the iris data, and welfare
and poverty indicators.
|
On-line PCA with Optimal Regrets | We carefully investigate the on-line version of PCA, where in each trial a
learning algorithm plays a k-dimensional subspace, and suffers the compression
loss on the next instance when projected into the chosen subspace. In this
setting, we analyze two popular on-line algorithms, Gradient Descent (GD) and
Exponentiated Gradient (EG). We show that both algorithms are essentially
optimal in the worst-case. This comes as a surprise, since EG is known to
perform sub-optimally when the instances are sparse. This different behavior of
EG for PCA is mainly related to the non-negativity of the loss in this case,
which makes the PCA setting qualitatively different from other settings studied
in the literature. Furthermore, we show that when considering regret bounds as
function of a loss budget, EG remains optimal and strictly outperforms GD.
Next, we study the extension of the PCA setting, in which the Nature is allowed
to play with dense instances, which are positive matrices with bounded largest
eigenvalue. Again we can show that EG is optimal and strictly better than GD in
this setting.
|
Stability of Multi-Task Kernel Regression Algorithms | We study the stability properties of nonlinear multi-task regression in
reproducing Hilbert spaces with operator-valued kernels. Such kernels, a.k.a.
multi-task kernels, are appropriate for learning prob- lems with nonscalar
outputs like multi-task learning and structured out- put prediction. We show
that multi-task kernel regression algorithms are uniformly stable in the
general case of infinite-dimensional output spaces. We then derive under mild
assumption on the kernel generaliza- tion bounds of such algorithms, and we
show their consistency even with non Hilbert-Schmidt operator-valued kernels .
We demonstrate how to apply the results to various multi-task kernel regression
methods such as vector-valued SVR and functional ridge regression.
|
On Finding the Largest Mean Among Many | Sampling from distributions to find the one with the largest mean arises in a
broad range of applications, and it can be mathematically modeled as a
multi-armed bandit problem in which each distribution is associated with an
arm. This paper studies the sample complexity of identifying the best arm
(largest mean) in a multi-armed bandit problem. Motivated by large-scale
applications, we are especially interested in identifying situations where the
total number of samples that are necessary and sufficient to find the best arm
scale linearly with the number of arms. We present a single-parameter
multi-armed bandit model that spans the range from linear to superlinear sample
complexity. We also give a new algorithm for best arm identification, called
PRISM, with linear sample complexity for a wide range of mean distributions.
The algorithm, like most exploration procedures for multi-armed bandits, is
adaptive in the sense that the next arms to sample are selected based on
previous samples. We compare the sample complexity of adaptive procedures with
simpler non-adaptive procedures using new lower bounds. For many problem
instances, the increased sample complexity required by non-adaptive procedures
is a polynomial factor of the number of arms.
|
Parallel Coordinate Descent Newton Method for Efficient
$\ell_1$-Regularized Minimization | The recent years have witnessed advances in parallel algorithms for large
scale optimization problems. Notwithstanding demonstrated success, existing
algorithms that parallelize over features are usually limited by divergence
issues under high parallelism or require data preprocessing to alleviate these
problems. In this work, we propose a Parallel Coordinate Descent Newton
algorithm using multidimensional approximate Newton steps (PCDN), where the
off-diagonal elements of the Hessian are set to zero to enable parallelization.
It randomly partitions the feature set into $b$ bundles/subsets with size of
$P$, and sequentially processes each bundle by first computing the descent
directions for each feature in parallel and then conducting $P$-dimensional
line search to obtain the step size. We show that: (1) PCDN is guaranteed to
converge globally despite increasing parallelism; (2) PCDN converges to the
specified accuracy $\epsilon$ within the limited iteration number of
$T_\epsilon$, and $T_\epsilon$ decreases with increasing parallelism (bundle
size $P$). Using the implementation technique of maintaining intermediate
quantities, we minimize the data transfer and synchronization cost of the
$P$-dimensional line search. For concreteness, the proposed PCDN algorithm is
applied to $\ell_1$-regularized logistic regression and $\ell_2$-loss SVM.
Experimental evaluations on six benchmark datasets show that the proposed PCDN
algorithm exploits parallelism well and outperforms the state-of-the-art
methods in speed without losing accuracy.
|
Bioclimating Modelling: A Machine Learning Perspective | Many machine learning (ML) approaches are widely used to generate bioclimatic
models for prediction of geographic range of organism as a function of climate.
Applications such as prediction of range shift in organism, range of invasive
species influenced by climate change are important parameters in understanding
the impact of climate change. However, success of machine learning-based
approaches depends on a number of factors. While it can be safely said that no
particular ML technique can be effective in all applications and success of a
technique is predominantly dependent on the application or the type of the
problem, it is useful to understand their behaviour to ensure informed choice
of techniques. This paper presents a comprehensive review of machine
learning-based bioclimatic model generation and analyses the factors
influencing success of such models. Considering the wide use of statistical
techniques, in our discussion we also include conventional statistical
techniques used in bioclimatic modelling.
|
Joint estimation of sparse multivariate regression and conditional
graphical models | Multivariate regression model is a natural generalization of the classical
univari- ate regression model for fitting multiple responses. In this paper, we
propose a high- dimensional multivariate conditional regression model for
constructing sparse estimates of the multivariate regression coefficient matrix
that accounts for the dependency struc- ture among the multiple responses. The
proposed method decomposes the multivariate regression problem into a series of
penalized conditional log-likelihood of each response conditioned on the
covariates and other responses. It allows simultaneous estimation of the sparse
regression coefficient matrix and the sparse inverse covariance matrix. The
asymptotic selection consistency and normality are established for the
diverging dimension of the covariates and number of responses. The
effectiveness of the pro- posed method is also demonstrated in a variety of
simulated examples as well as an application to the Glioblastoma multiforme
cancer data.
|
Hacking Smart Machines with Smarter Ones: How to Extract Meaningful Data
from Machine Learning Classifiers | Machine Learning (ML) algorithms are used to train computers to perform a
variety of complex tasks and improve with experience. Computers learn how to
recognize patterns, make unintended decisions, or react to a dynamic
environment. Certain trained machines may be more effective than others because
they are based on more suitable ML algorithms or because they were trained
through superior training sets. Although ML algorithms are known and publicly
released, training sets may not be reasonably ascertainable and, indeed, may be
guarded as trade secrets. While much research has been performed about the
privacy of the elements of training sets, in this paper we focus our attention
on ML classifiers and on the statistical information that can be unconsciously
or maliciously revealed from them. We show that it is possible to infer
unexpected but useful information from ML classifiers. In particular, we build
a novel meta-classifier and train it to hack other classifiers, obtaining
meaningful information about their training sets. This kind of information
leakage can be exploited, for example, by a vendor to build more effective
classifiers or to simply acquire trade secrets from a competitor's apparatus,
potentially violating its intellectual property rights.
|
Table of Content detection using Machine Learning | Table of content (TOC) detection has drawn attention now a day because it
plays an important role in digitization of multipage document. Generally book
document is multipage document. So it becomes necessary to detect Table of
Content page for easy navigation of multipage document and also to make
information retrieval faster for desirable data from the multipage document.
All the Table of content pages follow the different layout, different way of
presenting the contents of the document like chapter, section, subsection etc.
This paper introduces a new method to detect Table of content using machine
learning technique with different features. With the main aim to detect Table
of Content pages is to structure the document according to their contents.
|
A Fuzzy Based Approach to Text Mining and Document Clustering | Fuzzy logic deals with degrees of truth. In this paper, we have shown how to
apply fuzzy logic in text mining in order to perform document clustering. We
took an example of document clustering where the documents had to be clustered
into two categories. The method involved cleaning up the text and stemming of
words. Then, we chose m number of features which differ significantly in their
word frequencies (WF), normalized by document length, between documents
belonging to these two clusters. The documents to be clustered were represented
as a collection of m normalized WF values. Fuzzy c-means (FCM) algorithm was
used to cluster these documents into two clusters. After the FCM execution
finished, the documents in the two clusters were analysed for the values of
their respective m features. It was known that documents belonging to a
document type, say X, tend to have higher WF values for some particular
features. If the documents belonging to a cluster had higher WF values for
those same features, then that cluster was said to represent X. By fuzzy logic,
we not only get the cluster name, but also the degree to which a document
belongs to a cluster.
|
Stochastic Majorization-Minimization Algorithms for Large-Scale
Optimization | Majorization-minimization algorithms consist of iteratively minimizing a
majorizing surrogate of an objective function. Because of its simplicity and
its wide applicability, this principle has been very popular in statistics and
in signal processing. In this paper, we intend to make this principle scalable.
We introduce a stochastic majorization-minimization scheme which is able to
deal with large-scale or possibly infinite data sets. When applied to convex
optimization problems under suitable assumptions, we show that it achieves an
expected convergence rate of $O(1/\sqrt{n})$ after $n$ iterations, and of
$O(1/n)$ for strongly convex functions. Equally important, our scheme almost
surely converges to stationary points for a large class of non-convex problems.
We develop several efficient algorithms based on our framework. First, we
propose a new stochastic proximal gradient method, which experimentally matches
state-of-the-art solvers for large-scale $\ell_1$-logistic regression. Second,
we develop an online DC programming algorithm for non-convex sparse estimation.
Finally, we demonstrate the effectiveness of our approach for solving
large-scale structured matrix factorization problems.
|
Multiarmed Bandits With Limited Expert Advice | We solve the COLT 2013 open problem of \citet{SCB} on minimizing regret in
the setting of advice-efficient multiarmed bandits with expert advice. We give
an algorithm for the setting of K arms and N experts out of which we are
allowed to query and use only M experts' advices in each round, which has a
regret bound of \tilde{O}\bigP{\sqrt{\frac{\min\{K, M\} N}{M} T}} after T
rounds. We also prove that any algorithm for this problem must have expected
regret at least \tilde{\Omega}\bigP{\sqrt{\frac{\min\{K, M\} N}{M}T}}, thus
showing that our upper bound is nearly tight.
|
Galerkin Methods for Complementarity Problems and Variational
Inequalities | Complementarity problems and variational inequalities arise in a wide variety
of areas, including machine learning, planning, game theory, and physical
simulation. In all of these areas, to handle large-scale problem instances, we
need fast approximate solution methods. One promising idea is Galerkin
approximation, in which we search for the best answer within the span of a
given set of basis functions. Bertsekas proposed one possible Galerkin method
for variational inequalities. However, this method can exhibit two problems in
practice: its approximation error is worse than might be expected based on the
ability of the basis to represent the desired solution, and each iteration
requires a projection step that is not always easy to implement efficiently.
So, in this paper, we present a new Galerkin method with improved behavior: our
new error bounds depend directly on the distance from the true solution to the
subspace spanned by our basis, and the only projections we require are onto the
feasible region or onto the span of our basis.
|
From-Below Approximations in Boolean Matrix Factorization: Geometry and
New Algorithm | We present new results on Boolean matrix factorization and a new algorithm
based on these results. The results emphasize the significance of
factorizations that provide from-below approximations of the input matrix.
While the previously proposed algorithms do not consider the possibly different
significance of different matrix entries, our results help measure such
significance and suggest where to focus when computing factors. An experimental
evaluation of the new algorithm on both synthetic and real data demonstrates
its good performance in terms of good coverage by the first k factors as well
as a small number of factors needed for exact decomposition and indicates that
the algorithm outperforms the available ones in these terms. We also propose
future research topics.
|
Machine Teaching for Bayesian Learners in the Exponential Family | What if there is a teacher who knows the learning goal and wants to design
good training data for a machine learner? We propose an optimal teaching
framework aimed at learners who employ Bayesian models. Our framework is
expressed as an optimization problem over teaching examples that balance the
future loss of the learner and the effort of the teacher. This optimization
problem is in general hard. In the case where the learner employs conjugate
exponential family models, we present an approximate algorithm for finding the
optimal teaching set. Our algorithm optimizes the aggregate sufficient
statistics, then unpacks them into actual teaching examples. We give several
examples to illustrate our framework.
|
Class Proportion Estimation with Application to Multiclass Anomaly
Rejection | This work addresses two classification problems that fall under the heading
of domain adaptation, wherein the distributions of training and testing
examples differ. The first problem studied is that of class proportion
estimation, which is the problem of estimating the class proportions in an
unlabeled testing data set given labeled examples of each class. Compared to
previous work on this problem, our approach has the novel feature that it does
not require labeled training data from one of the classes. This property allows
us to address the second domain adaptation problem, namely, multiclass anomaly
rejection. Here, the goal is to design a classifier that has the option of
assigning a "reject" label, indicating that the instance did not arise from a
class present in the training data. We establish consistent learning strategies
for both of these domain adaptation problems, which to our knowledge are the
first of their kind. We also implement the class proportion estimation
technique and demonstrate its performance on several benchmark data sets.
|
Song-based Classification techniques for Endangered Bird Conservation | The work presented in this paper is part of a global framework which long
term goal is to design a wireless sensor network able to support the
observation of a population of endangered birds. We present the first stage for
which we have conducted a knowledge discovery approach on a sample of
acoustical data. We use MFCC features extracted from bird songs and we exploit
two knowledge discovery techniques. One that relies on clustering-based
approaches, that highlights the homogeneity in the songs of the species. The
other, based on predictive modeling, that demonstrates the good performances of
various machine learning techniques for the identification process. The
knowledge elicited provides promising results to consider a widespread study
and to elicit guidelines for designing a first version of the automatic
approach for data collection based on acoustic sensors.
|
A Statistical Perspective on Algorithmic Leveraging | One popular method for dealing with large-scale data sets is sampling. For
example, by using the empirical statistical leverage scores as an importance
sampling distribution, the method of algorithmic leveraging samples and
rescales rows/columns of data matrices to reduce the data size before
performing computations on the subproblem. This method has been successful in
improving computational efficiency of algorithms for matrix problems such as
least-squares approximation, least absolute deviations approximation, and
low-rank matrix approximation. Existing work has focused on algorithmic issues
such as worst-case running times and numerical issues associated with providing
high-quality implementations, but none of it addresses statistical aspects of
this method.
In this paper, we provide a simple yet effective framework to evaluate the
statistical properties of algorithmic leveraging in the context of estimating
parameters in a linear regression model with a fixed number of predictors. We
show that from the statistical perspective of bias and variance, neither
leverage-based sampling nor uniform sampling dominates the other. This result
is particularly striking, given the well-known result that, from the
algorithmic perspective of worst-case analysis, leverage-based sampling
provides uniformly superior worst-case algorithmic results, when compared with
uniform sampling. Based on these theoretical results, we propose and analyze
two new leveraging algorithms. A detailed empirical evaluation of existing
leverage-based methods as well as these two new methods is carried out on both
synthetic and real data sets. The empirical results indicate that our theory is
a good predictor of practical performance of existing and new leverage-based
algorithms and that the new algorithms achieve improved performance.
|
Model Reframing by Feature Context Change | The feature space (including both input and output variables) characterises a
data mining problem. In predictive (supervised) problems, the quality and
availability of features determines the predictability of the dependent
variable, and the performance of data mining models in terms of
misclassification or regression error. Good features, however, are usually
difficult to obtain. It is usual that many instances come with missing values,
either because the actual value for a given attribute was not available or
because it was too expensive. This is usually interpreted as a utility or
cost-sensitive learning dilemma, in this case between misclassification (or
regression error) costs and attribute tests costs. Both misclassification cost
(MC) and test cost (TC) can be integrated into a single measure, known as joint
cost (JC). We introduce methods and plots (such as the so-called JROC plots)
that can work with any of-the-shelf predictive technique, including ensembles,
such that we re-frame the model to use the appropriate subset of attributes
(the feature configuration) during deployment time. In other words, models are
trained with the available attributes (once and for all) and then deployed by
setting missing values on the attributes that are deemed ineffective for
reducing the joint cost. As the number of feature configuration combinations
grows exponentially with the number of features we introduce quadratic methods
that are able to approximate the optimal configuration and model choices, as
shown by the experimental results.
|
Deep Learning by Scattering | We introduce general scattering transforms as mathematical models of deep
neural networks with l2 pooling. Scattering networks iteratively apply complex
valued unitary operators, and the pooling is performed by a complex modulus. An
expected scattering defines a contractive representation of a high-dimensional
probability distribution, which preserves its mean-square norm. We show that
unsupervised learning can be casted as an optimization of the space contraction
to preserve the volume occupied by unlabeled examples, at each layer of the
network. Supervised learning and classification are performed with an averaged
scattering, which provides scattering estimations for multiple classes.
|
Correlated random features for fast semi-supervised learning | This paper presents Correlated Nystrom Views (XNV), a fast semi-supervised
algorithm for regression and classification. The algorithm draws on two main
ideas. First, it generates two views consisting of computationally inexpensive
random features. Second, XNV applies multiview regression using Canonical
Correlation Analysis (CCA) on unlabeled data to bias the regression towards
useful features. It has been shown that, if the views contains accurate
estimators, CCA regression can substantially reduce variance with a minimal
increase in bias. Random views are justified by recent theoretical and
empirical work showing that regression with random features closely
approximates kernel regression, implying that random views can be expected to
contain accurate estimators. We show that XNV consistently outperforms a
state-of-the-art algorithm for semi-supervised learning: substantially
improving predictive performance and reducing the variability of performance on
a wide variety of real-world datasets, whilst also reducing runtime by orders
of magnitude.
|
Synthesizing Manipulation Sequences for Under-Specified Tasks using
Unrolled Markov Random Fields | Many tasks in human environments require performing a sequence of navigation
and manipulation steps involving objects. In unstructured human environments,
the location and configuration of the objects involved often change in
unpredictable ways. This requires a high-level planning strategy that is robust
and flexible in an uncertain environment. We propose a novel dynamic planning
strategy, which can be trained from a set of example sequences. High level
tasks are expressed as a sequence of primitive actions or controllers (with
appropriate parameters). Our score function, based on Markov Random Field
(MRF), captures the relations between environment, controllers, and their
arguments. By expressing the environment using sets of attributes, the approach
generalizes well to unseen scenarios. We train the parameters of our MRF using
a maximum margin learning method. We provide a detailed empirical validation of
our overall framework demonstrating successful plan strategies for a variety of
tasks.
|
Fourier PCA and Robust Tensor Decomposition | Fourier PCA is Principal Component Analysis of a matrix obtained from higher
order derivatives of the logarithm of the Fourier transform of a
distribution.We make this method algorithmic by developing a tensor
decomposition method for a pair of tensors sharing the same vectors in rank-$1$
decompositions. Our main application is the first provably polynomial-time
algorithm for underdetermined ICA, i.e., learning an $n \times m$ matrix $A$
from observations $y=Ax$ where $x$ is drawn from an unknown product
distribution with arbitrary non-Gaussian components. The number of component
distributions $m$ can be arbitrarily higher than the dimension $n$ and the
columns of $A$ only need to satisfy a natural and efficiently verifiable
nondegeneracy condition. As a second application, we give an alternative
algorithm for learning mixtures of spherical Gaussians with linearly
independent means. These results also hold in the presence of Gaussian noise.
|
Design of an Agent for Answering Back in Smart Phones | The objective of the paper is to design an agent which provides efficient
response to the caller when a call goes unanswered in smartphones. The agent
provides responses through text messages, email etc stating the most likely
reason as to why the callee is unable to answer a call. Responses are composed
taking into consideration the importance of the present call and the situation
the callee is in at the moment like driving, sleeping, at work etc. The agent
makes decisons in the compostion of response messages based on the patterns it
has come across in the learning environment. Initially the user helps the agent
to compose response messages. The agent associates this message to the percept
it recieves with respect to the environment the callee is in. The user may
thereafter either choose to make to response system automatic or choose to
recieve suggestions from the agent for responses messages and confirm what is
to be sent to the caller.
|
A Randomized Nonmonotone Block Proximal Gradient Method for a Class of
Structured Nonlinear Programming | We propose a randomized nonmonotone block proximal gradient (RNBPG) method
for minimizing the sum of a smooth (possibly nonconvex) function and a
block-separable (possibly nonconvex nonsmooth) function. At each iteration,
this method randomly picks a block according to any prescribed probability
distribution and solves typically several associated proximal subproblems that
usually have a closed-form solution, until a certain progress on objective
value is achieved. In contrast to the usual randomized block coordinate descent
method [23,20], our method has a nonmonotone flavor and uses variable stepsizes
that can partially utilize the local curvature information of the smooth
component of objective function. We show that any accumulation point of the
solution sequence of the method is a stationary point of the problem {\it
almost surely} and the method is capable of finding an approximate stationary
point with high probability. We also establish a sublinear rate of convergence
for the method in terms of the minimal expected squared norm of certain
proximal gradients over the iterations. When the problem under consideration is
convex, we show that the expected objective values generated by RNBPG converge
to the optimal value of the problem. Under some assumptions, we further
establish a sublinear and linear rate of convergence on the expected objective
values generated by a monotone version of RNBPG. Finally, we conduct some
preliminary experiments to test the performance of RNBPG on the
$\ell_1$-regularized least-squares problem and a dual SVM problem in machine
learning. The computational results demonstrate that our method substantially
outperforms the randomized block coordinate {\it descent} method with fixed or
variable stepsizes.
|
Understanding the Predictive Power of Computational Mechanics and Echo
State Networks in Social Media | There is a large amount of interest in understanding users of social media in
order to predict their behavior in this space. Despite this interest, user
predictability in social media is not well-understood. To examine this
question, we consider a network of fifteen thousand users on Twitter over a
seven week period. We apply two contrasting modeling paradigms: computational
mechanics and echo state networks. Both methods attempt to model the behavior
of users on the basis of their past behavior. We demonstrate that the behavior
of users on Twitter can be well-modeled as processes with self-feedback. We
find that the two modeling approaches perform very similarly for most users,
but that they differ in performance on a small subset of the users. By
exploring the properties of these performance-differentiated users, we
highlight the challenges faced in applying predictive models to dynamic social
data.
|
Scaling Up Robust MDPs by Reinforcement Learning | We consider large-scale Markov decision processes (MDPs) with parameter
uncertainty, under the robust MDP paradigm. Previous studies showed that robust
MDPs, based on a minimax approach to handle uncertainty, can be solved using
dynamic programming for small to medium sized problems. However, due to the
"curse of dimensionality", MDPs that model real-life problems are typically
prohibitively large for such approaches. In this work we employ a reinforcement
learning approach to tackle this planning problem: we develop a robust
approximate dynamic programming method based on a projected fixed point
equation to approximately solve large scale robust MDPs. We show that the
proposed method provably succeeds under certain technical conditions, and
demonstrate its effectiveness through simulation of an option pricing problem.
To the best of our knowledge, this is the first attempt to scale up the robust
MDPs paradigm.
|
Solving Relational MDPs with Exogenous Events and Additive Rewards | We formalize a simple but natural subclass of service domains for relational
planning problems with object-centered, independent exogenous events and
additive rewards capturing, for example, problems in inventory control.
Focusing on this subclass, we present a new symbolic planning algorithm which
is the first algorithm that has explicit performance guarantees for relational
MDPs with exogenous events. In particular, under some technical conditions, our
planning algorithm provides a monotonic lower bound on the optimal value
function. To support this algorithm we present novel evaluation and reduction
techniques for generalized first order decision diagrams, a knowledge
representation for real-valued functions over relational world states. Our
planning algorithm uses a set of focus states, which serves as a training set,
to simplify and approximate the symbolic solution, and can thus be seen to
perform learning for planning. A preliminary experimental evaluation
demonstrates the validity of our approach.
|
Traffic data reconstruction based on Markov random field modeling | We consider the traffic data reconstruction problem. Suppose we have the
traffic data of an entire city that are incomplete because some road data are
unobserved. The problem is to reconstruct the unobserved parts of the data. In
this paper, we propose a new method to reconstruct incomplete traffic data
collected from various traffic sensors. Our approach is based on Markov random
field modeling of road traffic. The reconstruction is achieved by using
mean-field method and a machine learning method. We numerically verify the
performance of our method using realistic simulated traffic data for the real
road network of Sendai, Japan.
|
A Survey on Metric Learning for Feature Vectors and Structured Data | The need for appropriate ways to measure the distance or similarity between
data is ubiquitous in machine learning, pattern recognition and data mining,
but handcrafting such good metrics for specific problems is generally
difficult. This has led to the emergence of metric learning, which aims at
automatically learning a metric from data and has attracted a lot of interest
in machine learning and related fields for the past ten years. This survey
paper proposes a systematic review of the metric learning literature,
highlighting the pros and cons of each approach. We pay particular attention to
Mahalanobis distance metric learning, a well-studied and successful framework,
but additionally present a wide range of methods that have recently emerged as
powerful alternatives, including nonlinear metric learning, similarity learning
and local metric learning. Recent trends and extensions, such as
semi-supervised metric learning, metric learning for histogram data and the
derivation of generalization guarantees, are also covered. Finally, this survey
addresses metric learning for structured data, in particular edit distance
learning, and attempts to give an overview of the remaining challenges in
metric learning for the years to come.
|
Evaluation Measures for Hierarchical Classification: a unified view and
novel approaches | Hierarchical classification addresses the problem of classifying items into a
hierarchy of classes. An important issue in hierarchical classification is the
evaluation of different classification algorithms, which is complicated by the
hierarchical relations among the classes. Several evaluation measures have been
proposed for hierarchical classification using the hierarchy in different ways.
This paper studies the problem of evaluation in hierarchical classification by
analyzing and abstracting the key components of the existing performance
measures. It also proposes two alternative generic views of hierarchical
evaluation and introduces two corresponding novel measures. The proposed
measures, along with the state-of-the art ones, are empirically tested on three
large datasets from the domain of text classification. The empirical results
illustrate the undesirable behavior of existing approaches and how the proposed
methods overcome most of these methods across a range of cases.
|
Memory Limited, Streaming PCA | We consider streaming, one-pass principal component analysis (PCA), in the
high-dimensional regime, with limited memory. Here, $p$-dimensional samples are
presented sequentially, and the goal is to produce the $k$-dimensional subspace
that best approximates these points. Standard algorithms require $O(p^2)$
memory; meanwhile no algorithm can do better than $O(kp)$ memory, since this is
what the output itself requires. Memory (or storage) complexity is most
meaningful when understood in the context of computational and sample
complexity. Sample complexity for high-dimensional PCA is typically studied in
the setting of the {\em spiked covariance model}, where $p$-dimensional points
are generated from a population covariance equal to the identity (white noise)
plus a low-dimensional perturbation (the spike) which is the signal to be
recovered. It is now well-understood that the spike can be recovered when the
number of samples, $n$, scales proportionally with the dimension, $p$. Yet, all
algorithms that provably achieve this, have memory complexity $O(p^2)$.
Meanwhile, algorithms with memory-complexity $O(kp)$ do not have provable
bounds on sample complexity comparable to $p$. We present an algorithm that
achieves both: it uses $O(kp)$ memory (meaning storage of any kind) and is able
to compute the $k$-dimensional spike with $O(p \log p)$ sample-complexity --
the first algorithm of its kind. While our theoretical analysis focuses on the
spiked covariance model, our simulations show that our algorithm is successful
on much more general models for the data.
|
Simple one-pass algorithm for penalized linear regression with
cross-validation on MapReduce | In this paper, we propose a one-pass algorithm on MapReduce for penalized
linear regression
\[f_\lambda(\alpha, \beta) = \|Y - \alpha\mathbf{1} - X\beta\|_2^2 +
p_{\lambda}(\beta)\] where $\alpha$ is the intercept which can be omitted
depending on application; $\beta$ is the coefficients and $p_{\lambda}$ is the
penalized function with penalizing parameter $\lambda$. $f_\lambda(\alpha,
\beta)$ includes interesting classes such as Lasso, Ridge regression and
Elastic-net. Compared to latest iterative distributed algorithms requiring
multiple MapReduce jobs, our algorithm achieves huge performance improvement;
moreover, our algorithm is exact compared to the approximate algorithms such as
parallel stochastic gradient decent. Moreover, what our algorithm distinguishes
with others is that it trains the model with cross validation to choose optimal
$\lambda$ instead of user specified one.
Key words: penalized linear regression, lasso, elastic-net, ridge, MapReduce
|
Concentration and Confidence for Discrete Bayesian Sequence Predictors | Bayesian sequence prediction is a simple technique for predicting future
symbols sampled from an unknown measure on infinite sequences over a countable
alphabet. While strong bounds on the expected cumulative error are known, there
are only limited results on the distribution of this error. We prove tight
high-probability bounds on the cumulative error, which is measured in terms of
the Kullback-Leibler (KL) divergence. We also consider the problem of
constructing upper confidence bounds on the KL and Hellinger errors similar to
those constructed from Hoeffding-like bounds in the i.i.d. case. The new
results are applied to show that Bayesian sequence prediction can be used in
the Knows What It Knows (KWIK) framework with bounds that match the
state-of-the-art.
|
Semi-supervised clustering methods | Cluster analysis methods seek to partition a data set into homogeneous
subgroups. It is useful in a wide variety of applications, including document
processing and modern genetics. Conventional clustering methods are
unsupervised, meaning that there is no outcome variable nor is anything known
about the relationship between the observations in the data set. In many
situations, however, information about the clusters is available in addition to
the values of the features. For example, the cluster labels of some
observations may be known, or certain observations may be known to belong to
the same cluster. In other cases, one may wish to identify clusters that are
associated with a particular outcome variable. This review describes several
clustering algorithms (known as "semi-supervised clustering" methods) that can
be applied in these situations. The majority of these methods are modifications
of the popular k-means clustering method, and several of them will be described
in detail. A brief description of some other semi-supervised clustering
algorithms is also provided.
|
Exploratory Learning | In multiclass semi-supervised learning (SSL), it is sometimes the case that
the number of classes present in the data is not known, and hence no labeled
examples are provided for some classes. In this paper we present variants of
well-known semi-supervised multiclass learning methods that are robust when the
data contains an unknown number of classes. In particular, we present an
"exploratory" extension of expectation-maximization (EM) that explores
different numbers of classes while learning. "Exploratory" SSL greatly improves
performance on three datasets in terms of F1 on the classes with seed examples
i.e., the classes which are expected to be in the data. Our Exploratory EM
algorithm also outperforms a SSL method based non-parametric Bayesian
clustering.
|
WebSets: Extracting Sets of Entities from the Web Using Unsupervised
Information Extraction | We describe a open-domain information extraction method for extracting
concept-instance pairs from an HTML corpus. Most earlier approaches to this
problem rely on combining clusters of distributionally similar terms and
concept-instance pairs obtained with Hearst patterns. In contrast, our method
relies on a novel approach for clustering terms found in HTML tables, and then
assigning concept names to these clusters using Hearst patterns. The method can
be efficiently applied to a large corpus, and experimental results on several
datasets show that our method can accurately extract large numbers of
concept-instance pairs.
|
Algorithms of the LDA model [REPORT] | We review three algorithms for Latent Dirichlet Allocation (LDA). Two of them
are variational inference algorithms: Variational Bayesian inference and Online
Variational Bayesian inference and one is Markov Chain Monte Carlo (MCMC)
algorithm -- Collapsed Gibbs sampling. We compare their time complexity and
performance. We find that online variational Bayesian inference is the fastest
algorithm and still returns reasonably good results.
|
Learning directed acyclic graphs based on sparsest permutations | We consider the problem of learning a Bayesian network or directed acyclic
graph (DAG) model from observational data. A number of constraint-based,
score-based and hybrid algorithms have been developed for this purpose. For
constraint-based methods, statistical consistency guarantees typically rely on
the faithfulness assumption, which has been show to be restrictive especially
for graphs with cycles in the skeleton. However, there is only limited work on
consistency guarantees for score-based and hybrid algorithms and it has been
unclear whether consistency guarantees can be proven under weaker conditions
than the faithfulness assumption. In this paper, we propose the sparsest
permutation (SP) algorithm. This algorithm is based on finding the causal
ordering of the variables that yields the sparsest DAG. We prove that this new
score-based method is consistent under strictly weaker conditions than the
faithfulness assumption. We also demonstrate through simulations on small DAGs
that the SP algorithm compares favorably to the constraint-based PC and SGS
algorithms as well as the score-based Greedy Equivalence Search and hybrid
Max-Min Hill-Climbing method. In the Gaussian setting, we prove that our
algorithm boils down to finding the permutation of the variables with sparsest
Cholesky decomposition for the inverse covariance matrix. Using this
connection, we show that in the oracle setting, where the true covariance
matrix is known, the SP algorithm is in fact equivalent to $\ell_0$-penalized
maximum likelihood estimation.
|
Challenges in Representation Learning: A report on three machine
learning contests | The ICML 2013 Workshop on Challenges in Representation Learning focused on
three challenges: the black box learning challenge, the facial expression
recognition challenge, and the multimodal learning challenge. We describe the
datasets created for these challenges and summarize the results of the
competitions. We provide suggestions for organizers of future challenges and
some comments on what kind of knowledge can be gained from machine learning
competitions.
|
An Empirical Study into Annotator Agreement, Ground Truth Estimation,
and Algorithm Evaluation | Although agreement between annotators has been studied in the past from a
statistical viewpoint, little work has attempted to quantify the extent to
which this phenomenon affects the evaluation of computer vision (CV) object
detection algorithms. Many researchers utilise ground truth (GT) in experiments
and more often than not this GT is derived from one annotator's opinion. How
does the difference in opinion affect an algorithm's evaluation? Four examples
of typical CV problems are chosen, and a methodology is applied to each to
quantify the inter-annotator variance and to offer insight into the mechanisms
behind agreement and the use of GT. It is found that when detecting linear
objects annotator agreement is very low. The agreement in object position,
linear or otherwise, can be partially explained through basic image properties.
Automatic object detectors are compared to annotator agreement and it is found
that a clear relationship exists. Several methods for calculating GTs from a
number of annotations are applied and the resulting differences in the
performance of the object detectors are quantified. It is found that the rank
of a detector is highly dependent upon the method used to form the GT. It is
also found that although the STAPLE and LSML GT estimation methods appear to
represent the mean of the performance measured using the individual
annotations, when there are few annotations, or there is a large variance in
them, these estimates tend to degrade. Furthermore, one of the most commonly
adopted annotation combination methods--consensus voting--accentuates more
obvious features, which results in an overestimation of the algorithm's
performance. Finally, it is concluded that in some datasets it may not be
possible to state with any confidence that one algorithm outperforms another
when evaluating upon one GT and a method for calculating confidence bounds is
discussed.
|
Quantum support vector machine for big data classification | Supervised machine learning is the classification of new data based on
already classified training examples. In this work, we show that the support
vector machine, an optimized binary classifier, can be implemented on a quantum
computer, with complexity logarithmic in the size of the vectors and the number
of training examples. In cases when classical sampling algorithms require
polynomial time, an exponential speed-up is obtained. At the core of this
quantum big data algorithm is a non-sparse matrix exponentiation technique for
efficiently performing a matrix inversion of the training data inner-product
(kernel) matrix.
|
Online discrete optimization in social networks in the presence of
Knightian uncertainty | We study a model of collective real-time decision-making (or learning) in a
social network operating in an uncertain environment, for which no a priori
probabilistic model is available. Instead, the environment's impact on the
agents in the network is seen through a sequence of cost functions, revealed to
the agents in a causal manner only after all the relevant actions are taken.
There are two kinds of costs: individual costs incurred by each agent and
local-interaction costs incurred by each agent and its neighbors in the social
network. Moreover, agents have inertia: each agent has a default mixed strategy
that stays fixed regardless of the state of the environment, and must expend
effort to deviate from this strategy in order to respond to cost signals coming
from the environment. We construct a decentralized strategy, wherein each agent
selects its action based only on the costs directly affecting it and on the
decisions made by its neighbors in the network. In this setting, we quantify
social learning in terms of regret, which is given by the difference between
the realized network performance over a given time horizon and the best
performance that could have been achieved in hindsight by a fictitious
centralized entity with full knowledge of the environment's evolution. We show
that our strategy achieves the regret that scales polylogarithmically with the
time horizon and polynomially with the number of agents and the maximum number
of neighbors of any agent in the social network.
|
A non-parametric conditional factor regression model for
high-dimensional input and response | In this paper, we propose a non-parametric conditional factor regression
(NCFR)model for domains with high-dimensional input and response. NCFR enhances
linear regression in two ways: a) introducing low-dimensional latent factors
leading to dimensionality reduction and b) integrating an Indian Buffet Process
as a prior for the latent factors to derive unlimited sparse dimensions.
Experimental results comparing NCRF to several alternatives give evidence to
remarkable prediction performance.
|
The Orchive : Data mining a massive bioacoustic archive | The Orchive is a large collection of over 20,000 hours of audio recordings
from the OrcaLab research facility located off the northern tip of Vancouver
Island. It contains recorded orca vocalizations from the 1980 to the present
time and is one of the largest resources of bioacoustic data in the world. We
have developed a web-based interface that allows researchers to listen to these
recordings, view waveform and spectral representations of the audio, label
clips with annotations, and view the results of machine learning classifiers
based on automatic audio features extraction. In this paper we describe such
classifiers that discriminate between background noise, orca calls, and the
voice notes that are present in most of the tapes. Furthermore we show
classification results for individual calls based on a previously existing orca
call catalog. We have also experimentally investigated the scalability of
classifiers over the entire Orchive.
|
Discovering the Markov network structure | In this paper a new proof is given for the supermodularity of information
content. Using the decomposability of the information content an algorithm is
given for discovering the Markov network graph structure endowed by the
pairwise Markov property of a given probability distribution. A discrete
probability distribution is given for which the equivalence of
Hammersley-Clifford theorem is fulfilled although some of the possible vector
realizations are taken on with zero probability. Our algorithm for discovering
the pairwise Markov network is illustrated on this example, too.
|
Distributed Online Big Data Classification Using Context Information | Distributed, online data mining systems have emerged as a result of
applications requiring analysis of large amounts of correlated and
high-dimensional data produced by multiple distributed data sources. We propose
a distributed online data classification framework where data is gathered by
distributed data sources and processed by a heterogeneous set of distributed
learners which learn online, at run-time, how to classify the different data
streams either by using their locally available classification functions or by
helping each other by classifying each other's data. Importantly, since the
data is gathered at different locations, sending the data to another learner to
process incurs additional costs such as delays, and hence this will be only
beneficial if the benefits obtained from a better classification will exceed
the costs. We model the problem of joint classification by the distributed and
heterogeneous learners from multiple data sources as a distributed contextual
bandit problem where each data is characterized by a specific context. We
develop a distributed online learning algorithm for which we can prove
sublinear regret. Compared to prior work in distributed online data mining, our
work is the first to provide analytic regret results characterizing the
performance of the proposed algorithm.
|
Data Fusion by Matrix Factorization | For most problems in science and engineering we can obtain data sets that
describe the observed system from various perspectives and record the behavior
of its individual components. Heterogeneous data sets can be collectively mined
by data fusion. Fusion can focus on a specific target relation and exploit
directly associated data together with contextual data and data about system's
constraints. In the paper we describe a data fusion approach with penalized
matrix tri-factorization (DFMF) that simultaneously factorizes data matrices to
reveal hidden associations. The approach can directly consider any data that
can be expressed in a matrix, including those from feature-based
representations, ontologies, associations and networks. We demonstrate the
utility of DFMF for gene function prediction task with eleven different data
sources and for prediction of pharmacologic actions by fusing six data sources.
Our data fusion algorithm compares favorably to alternative data integration
approaches and achieves higher accuracy than can be obtained from any single
data source alone.
|
Multi-Task Policy Search | Learning policies that generalize across multiple tasks is an important and
challenging research topic in reinforcement learning and robotics. Training
individual policies for every single potential task is often impractical,
especially for continuous task variations, requiring more principled approaches
to share and transfer knowledge among similar tasks. We present a novel
approach for learning a nonlinear feedback policy that generalizes across
multiple tasks. The key idea is to define a parametrized policy as a function
of both the state and the task, which allows learning a single policy that
generalizes across multiple known and unknown tasks. Applications of our novel
approach to reinforcement and imitation learning in real-robot experiments are
shown.
|
Semi-supervised Ranking Pursuit | We propose a novel sparse preference learning/ranking algorithm. Our
algorithm approximates the true utility function by a weighted sum of basis
functions using the squared loss on pairs of data points, and is a
generalization of the kernel matching pursuit method. It can operate both in a
supervised and a semi-supervised setting and allows efficient search for
multiple, near-optimal solutions. Furthermore, we describe the extension of the
algorithm suitable for combined ranking and regression tasks. In our
experiments we demonstrate that the proposed algorithm outperforms several
state-of-the-art learning methods when taking into account unlabeled data and
performs comparably in a supervised learning scenario, while providing sparser
solutions.
|
An Efficient Model Selection for Gaussian Mixture Model in a Bayesian
Framework | In order to cluster or partition data, we often use
Expectation-and-Maximization (EM) or Variational approximation with a Gaussian
Mixture Model (GMM), which is a parametric probability density function
represented as a weighted sum of $\hat{K}$ Gaussian component densities.
However, model selection to find underlying $\hat{K}$ is one of the key
concerns in GMM clustering, since we can obtain the desired clusters only when
$\hat{K}$ is known. In this paper, we propose a new model selection algorithm
to explore $\hat{K}$ in a Bayesian framework. The proposed algorithm builds the
density of the model order which any information criterions such as AIC and BIC
basically fail to reconstruct. In addition, this algorithm reconstructs the
density quickly as compared to the time-consuming Monte Carlo simulation.
|
On the minimal teaching sets of two-dimensional threshold functions | It is known that a minimal teaching set of any threshold function on the
twodimensional rectangular grid consists of 3 or 4 points. We derive exact
formulae for the numbers of functions corresponding to these values and further
refine them in the case of a minimal teaching set of size 3. We also prove that
the average cardinality of the minimal teaching sets of threshold functions is
asymptotically 7/2.
We further present corollaries of these results concerning some special
arrangements of lines in the plane.
|
Investigating the Detection of Adverse Drug Events in a UK General
Practice Electronic Health-Care Database | Data-mining techniques have frequently been developed for Spontaneous
reporting databases. These techniques aim to find adverse drug events
accurately and efficiently. Spontaneous reporting databases are prone to
missing information, under reporting and incorrect entries. This often results
in a detection lag or prevents the detection of some adverse drug events. These
limitations do not occur in electronic health-care databases. In this paper,
existing methods developed for spontaneous reporting databases are implemented
on both a spontaneous reporting database and a general practice electronic
health-care database and compared. The results suggests that the application of
existing methods to the general practice database may help find signals that
have gone undetected when using the spontaneous reporting system database. In
addition the general practice database provides far more supplementary
information, that if incorporated in analysis could provide a wealth of
information for identifying adverse events more accurately.
|
Application of a clustering framework to UK domestic electricity data | This paper takes an approach to clustering domestic electricity load profiles
that has been successfully used with data from Portugal and applies it to UK
data. Clustering techniques are applied and it is found that the preferred
technique in the Portuguese work (a two stage process combining Self Organised
Maps and Kmeans) is not appropriate for the UK data. The work shows that up to
nine clusters of households can be identified with the differences in usage
profiles being visually striking. This demonstrates the appropriateness of
breaking the electricity usage patterns down to more detail than the two load
profiles currently published by the electricity industry. The paper details
initial results using data collected in Milton Keynes around 1990. Further work
is described and will concentrate on building accurate and meaningful clusters
of similar electricity users in order to better direct demand side management
initiatives to the most relevant target customers.
|
AdaBoost and Forward Stagewise Regression are First-Order Convex
Optimization Methods | Boosting methods are highly popular and effective supervised learning methods
which combine weak learners into a single accurate model with good statistical
performance. In this paper, we analyze two well-known boosting methods,
AdaBoost and Incremental Forward Stagewise Regression (FS$_\varepsilon$), by
establishing their precise connections to the Mirror Descent algorithm, which
is a first-order method in convex optimization. As a consequence of these
connections we obtain novel computational guarantees for these boosting
methods. In particular, we characterize convergence bounds of AdaBoost, related
to both the margin and log-exponential loss function, for any step-size
sequence. Furthermore, this paper presents, for the first time, precise
computational complexity results for FS$_\varepsilon$.
|
Constructing Hierarchical Image-tags Bimodal Representations for Word
Tags Alternative Choice | This paper describes our solution to the multi-modal learning challenge of
ICML. This solution comprises constructing three-level representations in three
consecutive stages and choosing correct tag words with a data-specific
strategy. Firstly, we use typical methods to obtain level-1 representations.
Each image is represented using MPEG-7 and gist descriptors with additional
features released by the contest organizers. And the corresponding word tags
are represented by bag-of-words model with a dictionary of 4000 words.
Secondly, we learn the level-2 representations using two stacked RBMs for each
modality. Thirdly, we propose a bimodal auto-encoder to learn the
similarities/dissimilarities between the pairwise image-tags as level-3
representations. Finally, during the test phase, based on one observation of
the dataset, we come up with a data-specific strategy to choose the correct tag
words leading to a leap of an improved overall performance. Our final average
accuracy on the private test set is 100%, which ranks the first place in this
challenge.
|
The Application of a Data Mining Framework to Energy Usage Profiling in
Domestic Residences using UK data | This paper describes a method for defining representative load profiles for
domestic electricity users in the UK. It considers bottom up and clustering
methods and then details the research plans for implementing and improving
existing framework approaches based on the overall usage profile. The work
focuses on adapting and applying analysis framework approaches to UK energy
data in order to determine the effectiveness of creating a few (single figures)
archetypical users with the intention of improving on the current methods of
determining usage profiles. The work is currently in progress and the paper
details initial results using data collected in Milton Keynes around 1990.
Various possible enhancements to the work are considered including a split
based on temperature to reflect the varying UK weather conditions.
|
Creating Personalised Energy Plans. From Groups to Individuals using
Fuzzy C Means Clustering | Changes in the UK electricity market mean that domestic users will be
required to modify their usage behaviour in order that supplies can be
maintained. Clustering allows usage profiles collected at the household level
to be clustered into groups and assigned a stereotypical profile which can be
used to target marketing campaigns. Fuzzy C Means clustering extends this by
allowing each household to be a member of many groups and hence provides the
opportunity to make personalised offers to the household dependent on their
degree of membership of each group. In addition, feedback can be provided on
how user's changing behaviour is moving them towards more "green" or cost
effective stereotypical usage.
|
Examining the Classification Accuracy of TSVMs with ?Feature Selection
in Comparison with the GLAD Algorithm | Gene expression data sets are used to classify and predict patient diagnostic
categories. As we know, it is extremely difficult and expensive to obtain gene
expression labelled examples. Moreover, conventional supervised approaches
cannot function properly when labelled data (training examples) are
insufficient using Support Vector Machines (SVM) algorithms. Therefore, in this
paper, we suggest Transductive Support Vector Machines (TSVMs) as
semi-supervised learning algorithms, learning with both labelled samples data
and unlabelled samples to perform the classification of microarray data. To
prune the superfluous genes and samples we used a feature selection method
called Recursive Feature Elimination (RFE), which is supposed to enhance the
output of classification and avoid the local optimization problem. We examined
the classification prediction accuracy of the TSVM-RFE algorithm in comparison
with the Genetic Learning Across Datasets (GLAD) algorithm, as both are
semi-supervised learning methods. Comparing these two methods, we found that
the TSVM-RFE surpassed both a SVM using RFE and GLAD.
|
Quiet in Class: Classification, Noise and the Dendritic Cell Algorithm | Theoretical analyses of the Dendritic Cell Algorithm (DCA) have yielded
several criticisms about its underlying structure and operation. As a result,
several alterations and fixes have been suggested in the literature to correct
for these findings. A contribution of this work is to investigate the effects
of replacing the classification stage of the DCA (which is known to be flawed)
with a traditional machine learning technique. This work goes on to question
the merits of those unique properties of the DCA that are yet to be thoroughly
analysed. If none of these properties can be found to have a benefit over
traditional approaches, then "fixing" the DCA is arguably less efficient than
simply creating a new algorithm. This work examines the dynamic filtering
property of the DCA and questions the utility of this unique feature for the
anomaly detection problem. It is found that this feature, while advantageous
for noisy, time-ordered classification, is not as useful as a traditional
static filter for processing a synthetic dataset. It is concluded that there
are still unique features of the DCA left to investigate. Areas that may be of
benefit to the Artificial Immune Systems community are suggested.
|
Detect adverse drug reactions for drug Alendronate | Adverse drug reaction (ADR) is widely concerned for public health issue. In
this study we propose an original approach to detect the ADRs using feature
matrix and feature selection. The experiments are carried out on the drug
Simvastatin. Major side effects for the drug are detected and better
performance is achieved compared to other computerized methods. The detected
ADRs are based on the computerized method, further investigation is needed.
|
Discovering Sequential Patterns in a UK General Practice Database | The wealth of computerised medical information becoming readily available
presents the opportunity to examine patterns of illnesses, therapies and
responses. These patterns may be able to predict illnesses that a patient is
likely to develop, allowing the implementation of preventative actions. In this
paper sequential rule mining is applied to a General Practice database to find
rules involving a patients age, gender and medical history. By incorporating
these rules into current health-care a patient can be highlighted as
susceptible to a future illness based on past or current illnesses, gender and
year of birth. This knowledge has the ability to greatly improve health-care
and reduce health-care costs.
|
Dropout Training as Adaptive Regularization | Dropout and other feature noising schemes control overfitting by artificially
corrupting the training data. For generalized linear models, dropout performs a
form of adaptive regularization. Using this viewpoint, we show that the dropout
regularizer is first-order equivalent to an L2 regularizer applied after
scaling the features by an estimate of the inverse diagonal Fisher information
matrix. We also establish a connection to AdaGrad, an online learning
algorithm, and find that a close relative of AdaGrad operates by repeatedly
solving linear dropout-regularized problems. By casting dropout as
regularization, we develop a natural semi-supervised algorithm that uses
unlabeled data to create a better adaptive regularizer. We apply this idea to
document classification tasks, and show that it consistently boosts the
performance of dropout training, improving on state-of-the-art results on the
IMDB reviews dataset.
|
Comparing Data-mining Algorithms Developed for Longitudinal
Observational Databases | Longitudinal observational databases have become a recent interest in the
post marketing drug surveillance community due to their ability of presenting a
new perspective for detecting negative side effects. Algorithms mining
longitudinal observation databases are not restricted by many of the
limitations associated with the more conventional methods that have been
developed for spontaneous reporting system databases. In this paper we
investigate the robustness of four recently developed algorithms that mine
longitudinal observational databases by applying them to The Health Improvement
Network (THIN) for six drugs with well document known negative side effects.
Our results show that none of the existing algorithms was able to consistently
identify known adverse drug reactions above events related to the cause of the
drug and no algorithm was superior.
|
Supervised Learning and Anti-learning of Colorectal Cancer Classes and
Survival Rates from Cellular Biology Parameters | In this paper, we describe a dataset relating to cellular and physical
conditions of patients who are operated upon to remove colorectal tumours. This
data provides a unique insight into immunological status at the point of tumour
removal, tumour classification and post-operative survival. Attempts are made
to learn relationships between attributes (physical and immunological) and the
resulting tumour stage and survival. Results for conventional machine learning
approaches can be considered poor, especially for predicting tumour stages for
the most important types of cancer. This poor performance is further
investigated and compared with a synthetic, dataset based on the logical
exclusive-OR function and it is shown that there is a significant level of
'anti-learning' present in all supervised methods used and this can be
explained by the highly dimensional, complex and sparsely representative
dataset. For predicting the stage of cancer from the immunological attributes,
anti-learning approaches outperform a range of popular algorithms.
|
Biomarker Clustering of Colorectal Cancer Data to Complement Clinical
Classification | In this paper, we describe a dataset relating to cellular and physical
conditions of patients who are operated upon to remove colorectal tumours. This
data provides a unique insight into immunological status at the point of tumour
removal, tumour classification and post-operative survival. Attempts are made
to cluster this dataset and important subsets of it in an effort to
characterize the data and validate existing standards for tumour
classification. It is apparent from optimal clustering that existing tumour
classification is largely unrelated to immunological factors within a patient
and that there may be scope for re-evaluating treatment options and survival
estimates based on a combination of tumour physiology and patient
histochemistry.
|
Polyglot: Distributed Word Representations for Multilingual NLP | Distributed word representations (word embeddings) have recently contributed
to competitive performance in language modeling and several NLP tasks. In this
work, we train word embeddings for more than 100 languages using their
corresponding Wikipedias. We quantitatively demonstrate the utility of our word
embeddings by using them as the sole features for training a part of speech
tagger for a subset of these languages. We find their performance to be
competitive with near state-of-art methods in English, Danish and Swedish.
Moreover, we investigate the semantic features captured by these embeddings
through the proximity of word groupings. We will release these embeddings
publicly to help researchers in the development and enhancement of multilingual
applications.
|
Stochastic Optimization of PCA with Capped MSG | We study PCA as a stochastic optimization problem and propose a novel
stochastic approximation algorithm which we refer to as "Matrix Stochastic
Gradient" (MSG), as well as a practical variant, Capped MSG. We study the
method both theoretically and empirically.
|
Approximate dynamic programming using fluid and diffusion approximations
with applications to power management | Neuro-dynamic programming is a class of powerful techniques for approximating
the solution to dynamic programming equations. In their most computationally
attractive formulations, these techniques provide the approximate solution only
within a prescribed finite-dimensional function class. Thus, the question that
always arises is how should the function class be chosen? The goal of this
paper is to propose an approach using the solutions to associated fluid and
diffusion approximations. In order to illustrate this approach, the paper
focuses on an application to dynamic speed scaling for power management in
computer processors.
|
Ensemble Methods for Multi-label Classification | Ensemble methods have been shown to be an effective tool for solving
multi-label classification tasks. In the RAndom k-labELsets (RAKEL) algorithm,
each member of the ensemble is associated with a small randomly-selected subset
of k labels. Then, a single label classifier is trained according to each
combination of elements in the subset. In this paper we adopt a similar
approach, however, instead of randomly choosing subsets, we select the minimum
required subsets of k labels that cover all labels and meet additional
constraints such as coverage of inter-label correlations. Construction of the
cover is achieved by formulating the subset selection as a minimum set covering
problem (SCP) and solving it by using approximation algorithms. Every cover
needs only to be prepared once by offline algorithms. Once prepared, a cover
may be applied to the classification of any given multi-label dataset whose
properties conform with those of the cover. The contribution of this paper is
two-fold. First, we introduce SCP as a general framework for constructing label
covers while allowing the user to incorporate cover construction constraints.
We demonstrate the effectiveness of this framework by proposing two
construction constraints whose enforcement produces covers that improve the
prediction performance of random selection. Second, we provide theoretical
bounds that quantify the probabilities of random selection to produce covers
that meet the proposed construction criteria. The experimental results indicate
that the proposed methods improve multi-label classification accuracy and
stability compared with the RAKEL algorithm and to other state-of-the-art
algorithms.
|
Loss minimization and parameter estimation with heavy tails | This work studies applications and generalizations of a simple estimation
technique that provides exponential concentration under heavy-tailed
distributions, assuming only bounded low-order moments. We show that the
technique can be used for approximate minimization of smooth and strongly
convex losses, and specifically for least squares linear regression. For
instance, our $d$-dimensional estimator requires just
$\tilde{O}(d\log(1/\delta))$ random samples to obtain a constant factor
approximation to the optimal least squares loss with probability $1-\delta$,
without requiring the covariates or noise to be bounded or subgaussian. We
provide further applications to sparse linear regression and low-rank
covariance matrix estimation with similar allowances on the noise and covariate
distributions. The core technique is a generalization of the median-of-means
estimator to arbitrary metric spaces.
|
B-tests: Low Variance Kernel Two-Sample Tests | A family of maximum mean discrepancy (MMD) kernel two-sample tests is
introduced. Members of the test family are called Block-tests or B-tests, since
the test statistic is an average over MMDs computed on subsets of the samples.
The choice of block size allows control over the tradeoff between test power
and computation time. In this respect, the $B$-test family combines favorable
properties of previously proposed MMD two-sample tests: B-tests are more
powerful than a linear time test where blocks are just pairs of samples, yet
they are more computationally efficient than a quadratic time test where a
single large block incorporating all the samples is used to compute a
U-statistic. A further important advantage of the B-tests is their
asymptotically Normal null distribution: this is by contrast with the
U-statistic, which is degenerate under the null hypothesis, and for which
estimates of the null distribution are computationally demanding. Recent
results on kernel selection for hypothesis testing transfer seamlessly to the
B-tests, yielding a means to optimize test power via kernel choice.
|
Using Clustering to extract Personality Information from socio economic
data | It has become apparent that models that have been applied widely in
economics, including Machine Learning techniques and Data Mining methods,
should take into consideration principles that derive from the theories of
Personality Psychology in order to discover more comprehensive knowledge
regarding complicated economic behaviours. In this work, we present a method to
extract Behavioural Groups by using simple clustering techniques that can
potentially reveal aspects of the Personalities for their members. We believe
that this is very important because the psychological information regarding the
Personalities of individuals is limited in real world applications and because
it can become a useful tool in improving the traditional models of Knowledge
Economy.
|
Finding the creatures of habit; Clustering households based on their
flexibility in using electricity | Changes in the UK electricity market, particularly with the roll out of smart
meters, will provide greatly increased opportunities for initiatives intended
to change households' electricity usage patterns for the benefit of the overall
system. Users show differences in their regular behaviours and clustering
households into similar groupings based on this variability provides for
efficient targeting of initiatives. Those people who are stuck into a regular
pattern of activity may be the least receptive to an initiative to change
behaviour. A sample of 180 households from the UK are clustered into four
groups as an initial test of the concept and useful, actionable groupings are
found.
|
A PAC-Bayesian Tutorial with A Dropout Bound | This tutorial gives a concise overview of existing PAC-Bayesian theory
focusing on three generalization bounds. The first is an Occam bound which
handles rules with finite precision parameters and which states that
generalization loss is near training loss when the number of bits needed to
write the rule is small compared to the sample size. The second is a
PAC-Bayesian bound providing a generalization guarantee for posterior
distributions rather than for individual rules. The PAC-Bayesian bound
naturally handles infinite precision rule parameters, $L_2$ regularization,
{\em provides a bound for dropout training}, and defines a natural notion of a
single distinguished PAC-Bayesian posterior distribution. The third bound is a
training-variance bound --- a kind of bias-variance analysis but with bias
replaced by expected training loss. The training-variance bound dominates the
other bounds but is more difficult to interpret. It seems to suggest variance
reduction methods such as bagging and may ultimately provide a more meaningful
analysis of dropouts.
|
Transmodal Analysis of Neural Signals | Localizing neuronal activity in the brain, both in time and in space, is a
central challenge to advance the understanding of brain function. Because of
the inability of any single neuroimaging techniques to cover all aspects at
once, there is a growing interest to combine signals from multiple modalities
in order to benefit from the advantages of each acquisition method. Due to the
complexity and unknown parameterization of any suggested complete model of BOLD
response in functional magnetic resonance imaging (fMRI), the development of a
reliable ultimate fusion approach remains difficult. But besides the primary
goal of superior temporal and spatial resolution, conjoint analysis of data
from multiple imaging modalities can alternatively be used to segregate neural
information from physiological and acquisition noise. In this paper we suggest
a novel methodology which relies on constructing a quantifiable mapping of data
from one modality (electroencephalography; EEG) into another (fMRI), called
transmodal analysis of neural signals (TRANSfusion). TRANSfusion attempts to
map neural data embedded within the EEG signal into its reflection in fMRI
data. Assessing the mapping performance on unseen data allows to localize brain
areas where a significant portion of the signal could be reliably
reconstructed, hence the areas neural activity of which is reflected in both
EEG and fMRI data. Consecutive analysis of the learnt model allows to localize
areas associated with specific frequency bands of EEG, or areas functionally
related (connected or coherent) to any given EEG sensor. We demonstrate the
performance of TRANSfusion on artificial and real data from an auditory
experiment. We further speculate on possible alternative uses: cross-modal data
filtering and EEG-driven interpolation of fMRI signals to obtain arbitrarily
high temporal sampling of BOLD.
|
Bridging Information Criteria and Parameter Shrinkage for Model
Selection | Model selection based on classical information criteria, such as BIC, is
generally computationally demanding, but its properties are well studied. On
the other hand, model selection based on parameter shrinkage by $\ell_1$-type
penalties is computationally efficient. In this paper we make an attempt to
combine their strengths, and propose a simple approach that penalizes the
likelihood with data-dependent $\ell_1$ penalties as in adaptive Lasso and
exploits a fixed penalization parameter. Even for finite samples, its model
selection results approximately coincide with those based on information
criteria; in particular, we show that in some special cases, this approach and
the corresponding information criterion produce exactly the same model. One can
also consider this approach as a way to directly determine the penalization
parameter in adaptive Lasso to achieve information criteria-like model
selection. As extensions, we apply this idea to complex models including
Gaussian mixture model and mixture of factor analyzers, whose model selection
is traditionally difficult to do; by adopting suitable penalties, we provide
continuous approximators to the corresponding information criteria, which are
easy to optimize and enable efficient model selection.
|
Bayesian Discovery of Multiple Bayesian Networks via Transfer Learning | Bayesian network structure learning algorithms with limited data are being
used in domains such as systems biology and neuroscience to gain insight into
the underlying processes that produce observed data. Learning reliable networks
from limited data is difficult, therefore transfer learning can improve the
robustness of learned networks by leveraging data from related tasks. Existing
transfer learning algorithms for Bayesian network structure learning give a
single maximum a posteriori estimate of network models. Yet, many other models
may be equally likely, and so a more informative result is provided by Bayesian
structure discovery. Bayesian structure discovery algorithms estimate posterior
probabilities of structural features, such as edges. We present transfer
learning for Bayesian structure discovery which allows us to explore the shared
and unique structural features among related tasks. Efficient computation
requires that our transfer learning objective factors into local calculations,
which we prove is given by a broad class of transfer biases. Theoretically, we
show the efficiency of our approach. Empirically, we show that compared to
single task learning, transfer learning is better able to positively identify
true edges. We apply the method to whole-brain neuroimaging data.
|
Tuned Models of Peer Assessment in MOOCs | In massive open online courses (MOOCs), peer grading serves as a critical
tool for scaling the grading of complex, open-ended assignments to courses with
tens or hundreds of thousands of students. But despite promising initial
trials, it does not always deliver accurate results compared to human experts.
In this paper, we develop algorithms for estimating and correcting for grader
biases and reliabilities, showing significant improvement in peer grading
accuracy on real data with 63,199 peer grades from Coursera's HCI course
offerings --- the largest peer grading networks analysed to date. We relate
grader biases and reliabilities to other student factors such as student
engagement, performance as well as commenting style. We also show that our
model can lead to more intelligent assignment of graders to gradees.
|
Controlling the Precision-Recall Tradeoff in Differential Dependency
Network Analysis | Graphical models have gained a lot of attention recently as a tool for
learning and representing dependencies among variables in multivariate data.
Often, domain scientists are looking specifically for differences among the
dependency networks of different conditions or populations (e.g. differences
between regulatory networks of different species, or differences between
dependency networks of diseased versus healthy populations). The standard
method for finding these differences is to learn the dependency networks for
each condition independently and compare them. We show that this approach is
prone to high false discovery rates (low precision) that can render the
analysis useless. We then show that by imposing a bias towards learning similar
dependency networks for each condition the false discovery rates can be reduced
to acceptable levels, at the cost of finding a reduced number of differences.
Algorithms developed in the transfer learning literature can be used to vary
the strength of the imposed similarity bias and provide a natural mechanism to
smoothly adjust this differential precision-recall tradeoff to cater to the
requirements of the analysis conducted. We present real case studies
(oncological and neurological) where domain experts use the proposed technique
to extract useful differential networks that shed light on the biological
processes involved in cancer and brain function.
|
Error Rate Bounds in Crowdsourcing Models | Crowdsourcing is an effective tool for human-powered computation on many
tasks challenging for computers. In this paper, we provide finite-sample
exponential bounds on the error rate (in probability and in expectation) of
hyperplane binary labeling rules under the Dawid-Skene crowdsourcing model. The
bounds can be applied to analyze many common prediction methods, including the
majority voting and weighted majority voting. These bound results could be
useful for controlling the error rate and designing better algorithms. We show
that the oracle Maximum A Posterior (MAP) rule approximately optimizes our
upper bound on the mean error rate for any hyperplane binary labeling rule, and
propose a simple data-driven weighted majority voting (WMV) rule (called
one-step WMV) that attempts to approximate the oracle MAP and has a provable
theoretical guarantee on the error rate. Moreover, we use simulated and real
data to demonstrate that the data-driven EM-MAP rule is a good approximation to
the oracle MAP rule, and to demonstrate that the mean error rate of the
data-driven EM-MAP rule is also bounded by the mean error rate bound of the
oracle MAP rule with estimated parameters plugging into the bound.
|
Flow-Based Algorithms for Local Graph Clustering | Given a subset S of vertices of an undirected graph G, the cut-improvement
problem asks us to find a subset S that is similar to A but has smaller
conductance. A very elegant algorithm for this problem has been given by
Andersen and Lang [AL08] and requires solving a small number of
single-commodity maximum flow computations over the whole graph G. In this
paper, we introduce LocalImprove, the first cut-improvement algorithm that is
local, i.e. that runs in time dependent on the size of the input set A rather
than on the size of the entire graph. Moreover, LocalImprove achieves this
local behaviour while essentially matching the same theoretical guarantee as
the global algorithm of Andersen and Lang.
The main application of LocalImprove is to the design of better
local-graph-partitioning algorithms. All previously known local algorithms for
graph partitioning are random-walk based and can only guarantee an output
conductance of O(\sqrt{OPT}) when the target set has conductance OPT \in [0,1].
Very recently, Zhu, Lattanzi and Mirrokni [ZLM13] improved this to O(OPT /
\sqrt{CONN}) where the internal connectivity parameter CONN \in [0,1] is
defined as the reciprocal of the mixing time of the random walk over the graph
induced by the target set. In this work, we show how to use LocalImprove to
obtain a constant approximation O(OPT) as long as CONN/OPT = Omega(1). This
yields the first flow-based algorithm. Moreover, its performance strictly
outperforms the ones based on random walks and surprisingly matches that of the
best known global algorithm, which is SDP-based, in this parameter regime
[MMV12].
Finally, our results show that spectral methods are not the only viable
approach to the construction of local graph partitioning algorithm and open
door to the study of algorithms with even better approximation and locality
guarantees.
|
Semantic Context Forests for Learning-Based Knee Cartilage Segmentation
in 3D MR Images | The automatic segmentation of human knee cartilage from 3D MR images is a
useful yet challenging task due to the thin sheet structure of the cartilage
with diffuse boundaries and inhomogeneous intensities. In this paper, we
present an iterative multi-class learning method to segment the femoral, tibial
and patellar cartilage simultaneously, which effectively exploits the spatial
contextual constraints between bone and cartilage, and also between different
cartilages. First, based on the fact that the cartilage grows in only certain
area of the corresponding bone surface, we extract the distance features of not
only to the surface of the bone, but more informatively, to the densely
registered anatomical landmarks on the bone surface. Second, we introduce a set
of iterative discriminative classifiers that at each iteration, probability
comparison features are constructed from the class confidence maps derived by
previously learned classifiers. These features automatically embed the semantic
context information between different cartilages of interest. Validated on a
total of 176 volumes from the Osteoarthritis Initiative (OAI) dataset, the
proposed approach demonstrates high robustness and accuracy of segmentation in
comparison with existing state-of-the-art MR cartilage segmentation methods.
|
Accuracy of MAP segmentation with hidden Potts and Markov mesh prior
models via Path Constrained Viterbi Training, Iterated Conditional Modes and
Graph Cut based algorithms | In this paper, we study statistical classification accuracy of two different
Markov field environments for pixelwise image segmentation, considering the
labels of the image as hidden states and solving the estimation of such labels
as a solution of the MAP equation. The emission distribution is assumed the
same in all models, and the difference lays in the Markovian prior hypothesis
made over the labeling random field. The a priori labeling knowledge will be
modeled with a) a second order anisotropic Markov Mesh and b) a classical
isotropic Potts model. Under such models, we will consider three different
segmentation procedures, 2D Path Constrained Viterbi training for the Hidden
Markov Mesh, a Graph Cut based segmentation for the first order isotropic Potts
model, and ICM (Iterated Conditional Modes) for the second order isotropic
Potts model.
We provide a unified view of all three methods, and investigate goodness of
fit for classification, studying the influence of parameter estimation,
computational gain, and extent of automation in the statistical measures
Overall Accuracy, Relative Improvement and Kappa coefficient, allowing robust
and accurate statistical analysis on synthetic and real-life experimental data
coming from the field of Dental Diagnostic Radiography. All algorithms, using
the learned parameters, generate good segmentations with little interaction
when the images have a clear multimodal histogram. Suboptimal learning proves
to be frail in the case of non-distinctive modes, which limits the complexity
of usable models, and hence the achievable error rate as well.
All Matlab code written is provided in a toolbox available for download from
our website, following the Reproducible Research Paradigm.
|
Statistical Active Learning Algorithms for Noise Tolerance and
Differential Privacy | We describe a framework for designing efficient active learning algorithms
that are tolerant to random classification noise and are
differentially-private. The framework is based on active learning algorithms
that are statistical in the sense that they rely on estimates of expectations
of functions of filtered random examples. It builds on the powerful statistical
query framework of Kearns (1993).
We show that any efficient active statistical learning algorithm can be
automatically converted to an efficient active learning algorithm which is
tolerant to random classification noise as well as other forms of
"uncorrelated" noise. The complexity of the resulting algorithms has
information-theoretically optimal quadratic dependence on $1/(1-2\eta)$, where
$\eta$ is the noise rate.
We show that commonly studied concept classes including thresholds,
rectangles, and linear separators can be efficiently actively learned in our
framework. These results combined with our generic conversion lead to the first
computationally-efficient algorithms for actively learning some of these
concept classes in the presence of random classification noise that provide
exponential improvement in the dependence on the error $\epsilon$ over their
passive counterparts. In addition, we show that our algorithms can be
automatically converted to efficient active differentially-private algorithms.
This leads to the first differentially-private active learning algorithms with
exponential label savings over the passive case.
|
Fast gradient descent for drifting least squares regression, with
application to bandits | Online learning algorithms require to often recompute least squares
regression estimates of parameters. We study improving the computational
complexity of such algorithms by using stochastic gradient descent (SGD) type
schemes in place of classic regression solvers. We show that SGD schemes
efficiently track the true solutions of the regression problems, even in the
presence of a drift. This finding coupled with an $O(d)$ improvement in
complexity, where $d$ is the dimension of the data, make them attractive for
implementation in the big data settings. In the case when strong convexity in
the regression problem is guaranteed, we provide bounds on the error both in
expectation and high probability (the latter is often needed to provide
theoretical guarantees for higher level algorithms), despite the drifting least
squares solution. As an example of this case we prove that the regret
performance of an SGD version of the PEGE linear bandit algorithm
[Rusmevichientong and Tsitsiklis 2010] is worse that that of PEGE itself only
by a factor of $O(\log^4 n)$. When strong convexity of the regression problem
cannot be guaranteed, we investigate using an adaptive regularisation. We make
an empirical study of an adaptively regularised, SGD version of LinUCB [Li et
al. 2010] in a news article recommendation application, which uses the large
scale news recommendation dataset from Yahoo! front page. These experiments
show a large gain in computational complexity, with a consistently low tracking
error and click-through-rate (CTR) performance that is $75\%$ close.
|
Optimal Bounds on Approximation of Submodular and XOS Functions by
Juntas | We investigate the approximability of several classes of real-valued
functions by functions of a small number of variables ({\em juntas}). Our main
results are tight bounds on the number of variables required to approximate a
function $f:\{0,1\}^n \rightarrow [0,1]$ within $\ell_2$-error $\epsilon$ over
the uniform distribution: 1. If $f$ is submodular, then it is $\epsilon$-close
to a function of $O(\frac{1}{\epsilon^2} \log \frac{1}{\epsilon})$ variables.
This is an exponential improvement over previously known results. We note that
$\Omega(\frac{1}{\epsilon^2})$ variables are necessary even for linear
functions. 2. If $f$ is fractionally subadditive (XOS) it is $\epsilon$-close
to a function of $2^{O(1/\epsilon^2)}$ variables. This result holds for all
functions with low total $\ell_1$-influence and is a real-valued analogue of
Friedgut's theorem for boolean functions. We show that $2^{\Omega(1/\epsilon)}$
variables are necessary even for XOS functions.
As applications of these results, we provide learning algorithms over the
uniform distribution. For XOS functions, we give a PAC learning algorithm that
runs in time $2^{poly(1/\epsilon)} poly(n)$. For submodular functions we give
an algorithm in the more demanding PMAC learning model (Balcan and Harvey,
2011) which requires a multiplicative $1+\gamma$ factor approximation with
probability at least $1-\epsilon$ over the target distribution. Our uniform
distribution algorithm runs in time $2^{poly(1/(\gamma\epsilon))} poly(n)$.
This is the first algorithm in the PMAC model that over the uniform
distribution can achieve a constant approximation factor arbitrarily close to 1
for all submodular functions. As follows from the lower bounds in (Feldman et
al., 2013) both of these algorithms are close to optimal. We also give
applications for proper learning, testing and agnostic learning with value
queries of these classes.
|
Unsupervised Gene Expression Data using Enhanced Clustering Method | Microarrays are made it possible to simultaneously monitor the expression
profiles of thousands of genes under various experimental conditions.
Identification of co-expressed genes and coherent patterns is the central goal
in microarray or gene expression data analysis and is an important task in
bioinformatics research. Feature selection is a process to select features
which are more informative. It is one of the important steps in knowledge
discovery. The problem is that not all features are important. Some of the
features may be redundant, and others may be irrelevant and noisy. In this work
the unsupervised Gene selection method and Enhanced Center Initialization
Algorithm (ECIA) with K-Means algorithms have been applied for clustering of
Gene Expression Data. This proposed clustering algorithm overcomes the
drawbacks in terms of specifying the optimal number of clusters and
initialization of good cluster centroids. Gene Expression Data show that could
identify compact clusters with performs well in terms of the Silhouette
Coefficients cluster measure.
|
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.