title
stringlengths 7
246
| abstract
stringlengths 6
3.31k
|
|---|---|
Causal and anti-causal learning in pattern recognition for neuroimaging | Pattern recognition in neuroimaging distinguishes between two types of
models: encoding- and decoding models. This distinction is based on the insight
that brain state features, that are found to be relevant in an experimental
paradigm, carry a different meaning in encoding- than in decoding models. In
this paper, we argue that this distinction is not sufficient: Relevant features
in encoding- and decoding models carry a different meaning depending on whether
they represent causal- or anti-causal relations. We provide a theoretical
justification for this argument and conclude that causal inference is essential
for interpretation in neuroimaging.
|
Feature-Level Domain Adaptation | Domain adaptation is the supervised learning setting in which the training
and test data are sampled from different distributions: training data is
sampled from a source domain, whilst test data is sampled from a target domain.
This paper proposes and studies an approach, called feature-level domain
adaptation (FLDA), that models the dependence between the two domains by means
of a feature-level transfer model that is trained to describe the transfer from
source to target domain. Subsequently, we train a domain-adapted classifier by
minimizing the expected loss under the resulting transfer model. For linear
classifiers and a large family of loss functions and transfer models, this
expected loss can be computed or approximated analytically, and minimized
efficiently. Our empirical evaluation of FLDA focuses on problems comprising
binary and count data in which the transfer can be naturally modeled via a
dropout distribution, which allows the classifier to adapt to differences in
the marginal probability of features in the source and the target domain. Our
experiments on several real-world problems show that FLDA performs on par with
state-of-the-art domain-adaptation techniques.
|
Data Driven Resource Allocation for Distributed Learning | In distributed machine learning, data is dispatched to multiple machines for
processing. Motivated by the fact that similar data points often belong to the
same or similar classes, and more generally, classification rules of high
accuracy tend to be "locally simple but globally complex" (Vapnik & Bottou
1993), we propose data dependent dispatching that takes advantage of such
structure. We present an in-depth analysis of this model, providing new
algorithms with provable worst-case guarantees, analysis proving existing
scalable heuristics perform well in natural non worst-case conditions, and
techniques for extending a dispatching rule from a small sample to the entire
distribution. We overcome novel technical challenges to satisfy important
conditions for accurate distributed learning, including fault tolerance and
balancedness. We empirically compare our approach with baselines based on
random partitioning, balanced partition trees, and locality sensitive hashing,
showing that we achieve significantly higher accuracy on both synthetic and
real world image and advertising datasets. We also demonstrate that our
technique strongly scales with the available computing power.
|
Energy-Efficient Classification for Anomaly Detection: The Wireless
Channel as a Helper | Anomaly detection has various applications including condition monitoring and
fault diagnosis. The objective is to sense the environment, learn the normal
system state, and then periodically classify whether the instantaneous state
deviates from the normal one or not. A flexible and cost-effective way of
monitoring a system state is to use a wireless sensor network. In the
traditional approach, the sensors encode their observations and transmit them
to a fusion center by means of some interference avoiding channel access
method. The fusion center then decodes all the data and classifies the
corresponding system state. As this approach can be highly inefficient in terms
of energy consumption, in this paper we propose a transmission scheme that
exploits interference for carrying out the anomaly detection directly in the
air. In other words, the wireless channel helps the fusion center to retrieve
the sought classification outcome immediately from the channel output. To
achieve this, the chosen learning model is linear support vector machines.
After discussing the proposed scheme and proving its reliability, we present
numerical examples demonstrating that the scheme reduces the energy consumption
for anomaly detection by up to 53% compared to a strategy that uses time
division multiple-access.
|
Increasing the Action Gap: New Operators for Reinforcement Learning | This paper introduces new optimality-preserving operators on Q-functions. We
first describe an operator for tabular representations, the consistent Bellman
operator, which incorporates a notion of local policy consistency. We show that
this local consistency leads to an increase in the action gap at each state;
increasing this gap, we argue, mitigates the undesirable effects of
approximation and estimation errors on the induced greedy policies. This
operator can also be applied to discretized continuous space and time problems,
and we provide empirical results evidencing superior performance in this
context. Extending the idea of a locally consistent operator, we then derive
sufficient conditions for an operator to preserve optimality, leading to a
family of operators which includes our consistent Bellman operator. As
corollaries we provide a proof of optimality for Baird's advantage learning
algorithm and derive other gap-increasing operators with interesting
properties. We conclude with an empirical study on 60 Atari 2600 games
illustrating the strong potential of these new operators.
|
Strategies for Training Large Vocabulary Neural Language Models | Training neural network language models over large vocabularies is still
computationally very costly compared to count-based models such as Kneser-Ney.
At the same time, neural language models are gaining popularity for many
applications such as speech recognition and machine translation whose success
depends on scalability. We present a systematic comparison of strategies to
represent and train large vocabularies, including softmax, hierarchical
softmax, target sampling, noise contrastive estimation and self normalization.
We further extend self normalization to be a proper estimator of likelihood and
introduce an efficient variant of softmax. We evaluate each method on three
popular benchmarks, examining performance on rare words, the speed/accuracy
trade-off and complementarity to Kneser-Ney.
|
A Light Touch for Heavily Constrained SGD | Minimizing empirical risk subject to a set of constraints can be a useful
strategy for learning restricted classes of functions, such as monotonic
functions, submodular functions, classifiers that guarantee a certain class
label for some subset of examples, etc. However, these restrictions may result
in a very large number of constraints. Projected stochastic gradient descent
(SGD) is often the default choice for large-scale optimization in machine
learning, but requires a projection after each update. For heavily-constrained
objectives, we propose an efficient extension of SGD that stays close to the
feasible region while only applying constraints probabilistically at each
iteration. Theoretical analysis shows a compelling trade-off between
per-iteration work and the number of iterations needed on problems with a large
number of constraints.
|
Streaming Kernel Principal Component Analysis | Kernel principal component analysis (KPCA) provides a concise set of basis
vectors which capture non-linear structures within large data sets, and is a
central tool in data analysis and learning. To allow for non-linear relations,
typically a full $n \times n$ kernel matrix is constructed over $n$ data
points, but this requires too much space and time for large values of $n$.
Techniques such as the Nystr\"om method and random feature maps can help
towards this goal, but they do not explicitly maintain the basis vectors in a
stream and take more space than desired. We propose a new approach for
streaming KPCA which maintains a small set of basis elements in a stream,
requiring space only logarithmic in $n$, and also improves the dependence on
the error parameter. Our technique combines together random feature maps with
recent advances in matrix sketching, it has guaranteed spectral norm error
bounds with respect to the original kernel matrix, and it compares favorably in
practice to state-of-the-art approaches.
|
DNA-Level Splice Junction Prediction using Deep Recurrent Neural
Networks | A eukaryotic gene consists of multiple exons (protein coding regions) and
introns (non-coding regions), and a splice junction refers to the boundary
between a pair of exon and intron. Precise identification of spice junctions on
a gene is important for deciphering its primary structure, function, and
interaction. Experimental techniques for determining exon/intron boundaries
include RNA-seq, which is often accompanied by computational approaches.
Canonical splicing signals are known, but computational junction prediction
still remains challenging because of a large number of false positives and
other complications. In this paper, we exploit deep recurrent neural networks
(RNNs) to model DNA sequences and to detect splice junctions thereon. We test
various RNN units and architectures including long short-term memory units,
gated recurrent units, and recently proposed iRNN for in-depth design space
exploration. According to our experimental results, the proposed approach
significantly outperforms not only conventional machine learning-based methods
but also a recent state-of-the-art deep belief network-based technique in terms
of prediction accuracy.
|
Learning Games and Rademacher Observations Losses | It has recently been shown that supervised learning with the popular logistic
loss is equivalent to optimizing the exponential loss over sufficient
statistics about the class: Rademacher observations (rados). We first show that
this unexpected equivalence can actually be generalized to other example / rado
losses, with necessary and sufficient conditions for the equivalence,
exemplified on four losses that bear popular names in various fields:
exponential (boosting), mean-variance (finance), Linear Hinge (on-line
learning), ReLU (deep learning), and unhinged (statistics). Second, we show
that the generalization unveils a surprising new connection to regularized
learning, and in particular a sufficient condition under which regularizing the
loss over examples is equivalent to regularizing the rados (with Minkowski
sums) in the equivalent rado loss. This brings simple and powerful rado-based
learning algorithms for sparsity-controlling regularization, that we exemplify
on a boosting algorithm for the regularized exponential rado-loss, which
formally boosts over four types of regularization, including the popular ridge
and lasso, and the recently coined slope --- we obtain the first proven
boosting algorithm for this last regularization. Through our first contribution
on the equivalence of rado and example-based losses, Omega-R.AdaBoost~appears
to be an efficient proxy to boost the regularized logistic loss over examples
using whichever of the four regularizers. Experiments display that
regularization consistently improves performances of rado-based learning, and
may challenge or beat the state of the art of example-based learning even when
learning over small sets of rados. Finally, we connect regularization to
differential privacy, and display how tiny budgets can be afforded on big
domains while beating (protected) example-based learning.
|
Blockout: Dynamic Model Selection for Hierarchical Deep Networks | Most deep architectures for image classification--even those that are trained
to classify a large number of diverse categories--learn shared image
representations with a single model. Intuitively, however, categories that are
more similar should share more information than those that are very different.
While hierarchical deep networks address this problem by learning separate
features for subsets of related categories, current implementations require
simplified models using fixed architectures specified via heuristic clustering
methods. Instead, we propose Blockout, a method for regularization and model
selection that simultaneously learns both the model architecture and
parameters. A generalization of Dropout, our approach gives a novel
parametrization of hierarchical architectures that allows for structure
learning via back-propagation. To demonstrate its utility, we evaluate Blockout
on the CIFAR and ImageNet datasets, demonstrating improved classification
accuracy, better regularization performance, faster training, and the clear
emergence of hierarchical network structures.
|
Feature Representation for ICU Mortality | Good predictors of ICU Mortality have the potential to identify high-risk
patients earlier, improve ICU resource allocation, or create more accurate
population-level risk models. Machine learning practitioners typically make
choices about how to represent features in a particular model, but these
choices are seldom evaluated quantitatively. This study compares the
performance of different representations of clinical event data from MIMIC II
in a logistic regression model to predict 36-hour ICU mortality. The most
common representations are linear (normalized counts) and binary (yes/no).
These, along with a new representation termed "hill", are compared using both
L1 and L2 regularization. Results indicate that the introduced "hill"
representation outperforms both the binary and linear representations, the hill
representation thus has the potential to improve existing models of ICU
mortality.
|
Unsupervised Feature Construction for Improving Data Representation and
Semantics | Feature-based format is the main data representation format used by machine
learning algorithms. When the features do not properly describe the initial
data, performance starts to degrade. Some algorithms address this problem by
internally changing the representation space, but the newly-constructed
features are rarely comprehensible. We seek to construct, in an unsupervised
way, new features that are more appropriate for describing a given dataset and,
at the same time, comprehensible for a human user. We propose two algorithms
that construct the new features as conjunctions of the initial primitive
features or their negations. The generated feature sets have reduced
correlations between features and succeed in catching some of the hidden
relations between individuals in a dataset. For example, a feature like $sky
\wedge \neg building \wedge panorama$ would be true for non-urban images and is
more informative than simple features expressing the presence or the absence of
an object. The notion of Pareto optimality is used to evaluate feature sets and
to obtain a balance between total correlation and the complexity of the
resulted feature set. Statistical hypothesis testing is used in order to
automatically determine the values of the parameters used for constructing a
data-dependent feature set. We experimentally show that our approaches achieve
the construction of informative feature sets for multiple datasets.
|
An Empirical Comparison of Neural Architectures for Reinforcement
Learning in Partially Observable Environments | This paper explores the performance of fitted neural Q iteration for
reinforcement learning in several partially observable environments, using
three recurrent neural network architectures: Long Short-Term Memory, Gated
Recurrent Unit and MUT1, a recurrent neural architecture evolved from a pool of
several thousands candidate architectures. A variant of fitted Q iteration,
based on Advantage values instead of Q values, is also explored. The results
show that GRU performs significantly better than LSTM and MUT1 for most of the
problems considered, requiring less training episodes and less CPU time before
learning a very good policy. Advantage learning also tends to produce better
results.
|
Deep-Spying: Spying using Smartwatch and Deep Learning | Wearable technologies are today on the rise, becoming more common and broadly
available to mainstream users. In fact, wristband and armband devices such as
smartwatches and fitness trackers already took an important place in the
consumer electronics market and are becoming ubiquitous. By their very nature
of being wearable, these devices, however, provide a new pervasive attack
surface threatening users privacy, among others.
In the meantime, advances in machine learning are providing unprecedented
possibilities to process complex data efficiently. Allowing patterns to emerge
from high dimensional unavoidably noisy data.
The goal of this work is to raise awareness about the potential risks related
to motion sensors built-in wearable devices and to demonstrate abuse
opportunities leveraged by advanced neural network architectures.
The LSTM-based implementation presented in this research can perform
touchlogging and keylogging on 12-keys keypads with above-average accuracy even
when confronted with raw unprocessed data. Thus demonstrating that deep neural
networks are capable of making keystroke inference attacks based on motion
sensors easier to achieve by removing the need for non-trivial pre-processing
pipelines and carefully engineered feature extraction strategies. Our results
suggest that the complete technological ecosystem of a user can be compromised
when a wearable wristband device is worn.
|
Probabilistic Programming with Gaussian Process Memoization | Gaussian Processes (GPs) are widely used tools in statistics, machine
learning, robotics, computer vision, and scientific computation. However,
despite their popularity, they can be difficult to apply; all but the simplest
classification or regression applications require specification and inference
over complex covariance functions that do not admit simple analytical
posteriors. This paper shows how to embed Gaussian processes in any
higher-order probabilistic programming language, using an idiom based on
memoization, and demonstrates its utility by implementing and extending classic
and state-of-the-art GP applications. The interface to Gaussian processes,
called gpmem, takes an arbitrary real-valued computational process as input and
returns a statistical emulator that automatically improve as the original
process is invoked and its input-output behavior is recorded. The flexibility
of gpmem is illustrated via three applications: (i) robust GP regression with
hierarchical hyper-parameter learning, (ii) discovering symbolic expressions
from time-series data by fully Bayesian structure learning over kernels
generated by a stochastic grammar, and (iii) a bandit formulation of Bayesian
optimization with automatic inference and action selection. All applications
share a single 50-line Python library and require fewer than 20 lines of
probabilistic code each.
|
A Survey of Available Corpora for Building Data-Driven Dialogue Systems | During the past decade, several areas of speech and language understanding
have witnessed substantial breakthroughs from the use of data-driven models. In
the area of dialogue systems, the trend is less obvious, and most practical
systems are still built through significant engineering and expert knowledge.
Nevertheless, several recent results suggest that data-driven approaches are
feasible and quite promising. To facilitate research in this area, we have
carried out a wide survey of publicly available datasets suitable for
data-driven learning of dialogue systems. We discuss important characteristics
of these datasets, how they can be used to learn diverse dialogue strategies,
and their other potential uses. We also examine methods for transfer learning
between datasets and the use of external knowledge. Finally, we discuss
appropriate choice of evaluation metrics for the learning objective.
|
Successive Ray Refinement and Its Application to Coordinate Descent for
LASSO | Coordinate descent is one of the most popular approaches for solving Lasso
and its extensions due to its simplicity and efficiency. When applying
coordinate descent to solving Lasso, we update one coordinate at a time while
fixing the remaining coordinates. Such an update, which is usually easy to
compute, greedily decreases the objective function value. In this paper, we aim
to improve its computational efficiency by reducing the number of coordinate
descent iterations. To this end, we propose a novel technique called Successive
Ray Refinement (SRR). SRR makes use of the following ray continuation property
on the successive iterations: for a particular coordinate, the value obtained
in the next iteration almost always lies on a ray that starts at its previous
iteration and passes through the current iteration. Motivated by this
ray-continuation property, we propose that coordinate descent be performed not
directly on the previous iteration but on a refined search point that has the
following properties: on one hand, it lies on a ray that starts at a history
solution and passes through the previous iteration, and on the other hand, it
achieves the minimum objective function value among all the points on the ray.
We propose two schemes for defining the search point and show that the refined
search point can be efficiently obtained. Empirical results for real and
synthetic data sets show that the proposed SRR can significantly reduce the
number of coordinate descent iterations, especially for small Lasso
regularization parameters.
|
Relay Backpropagation for Effective Learning of Deep Convolutional
Neural Networks | Learning deeper convolutional neural networks becomes a tendency in recent
years. However, many empirical evidences suggest that performance improvement
cannot be gained by simply stacking more layers. In this paper, we consider the
issue from an information theoretical perspective, and propose a novel method
Relay Backpropagation, that encourages the propagation of effective information
through the network in training stage. By virtue of the method, we achieved the
first place in ILSVRC 2015 Scene Classification Challenge. Extensive
experiments on two challenging large scale datasets demonstrate the
effectiveness of our method is not restricted to a specific dataset or network
architecture. Our models will be available to the research community later.
|
Deep Poisson Factorization Machines: factor analysis for mapping
behaviors in journalist ecosystem | Newsroom in online ecosystem is difficult to untangle. With prevalence of
social media, interactions between journalists and individuals become visible,
but lack of understanding to inner processing of information feedback loop in
public sphere leave most journalists baffled. Can we provide an organized view
to characterize journalist behaviors on individual level to know better of the
ecosystem? To this end, I propose Poisson Factorization Machine (PFM), a
Bayesian analogue to matrix factorization that assumes Poisson distribution for
generative process. The model generalizes recent studies on Poisson Matrix
Factorization to account temporal interaction which involves tensor-like
structure, and label information. Two inference procedures are designed, one
based on batch variational EM and another stochastic variational inference
scheme that efficiently scales with data size. An important novelty in this
note is that I show how to stack layers of PFM to introduce a deep
architecture. This work discusses some potential results applying the model and
explains how such latent factors may be useful for analyzing latent behaviors
for data exploration.
|
Complexity and Approximation of the Fuzzy K-Means Problem | The fuzzy $K$-means problem is a generalization of the classical $K$-means
problem to soft clusterings, i.e. clusterings where each points belongs to each
cluster to some degree. Although popular in practice, prior to this work the
fuzzy $K$-means problem has not been studied from a complexity theoretic or
algorithmic perspective. We show that optimal solutions for fuzzy $K$-means
cannot, in general, be expressed by radicals over the input points.
Surprisingly, this already holds for very simple inputs in one-dimensional
space. Hence, one cannot expect to compute optimal solutions exactly. We give
the first $(1+\epsilon)$-approximation algorithms for the fuzzy $K$-means
problem. First, we present a deterministic approximation algorithm whose
runtime is polynomial in $N$ and linear in the dimension $D$ of the input set,
given that $K$ is constant, i.e. a polynomial time approximation algorithm
given a fixed $K$. We achieve this result by showing that for each soft
clustering there exists a hard clustering with comparable properties. Second,
by using techniques known from coreset constructions for the $K$-means problem,
we develop a deterministic approximation algorithm that runs in time almost
linear in $N$ but exponential in the dimension $D$. We complement these results
with a randomized algorithm which imposes some natural restrictions on the
input set and whose runtime is comparable to some of the most efficient
approximation algorithms for $K$-means, i.e. linear in the number of points and
the dimension, but exponential in the number of clusters.
|
Asymptotic Behavior of Mean Partitions in Consensus Clustering | Although consistency is a minimum requirement of any estimator, little is
known about consistency of the mean partition approach in consensus clustering.
This contribution studies the asymptotic behavior of mean partitions. We show
that under normal assumptions, the mean partition approach is consistent and
asymptotic normal. To derive both results, we represent partitions as points of
some geometric space, called orbit space. Then we draw on results from the
theory of Fr\'echet means and stochastic programming. The asymptotic properties
hold for continuous extensions of standard cluster criteria (indices). The
results justify consensus clustering using finite but sufficiently large sample
sizes. Furthermore, the orbit space framework provides a mathematical
foundation for studying further statistical, geometrical, and analytical
properties of sets of partitions.
|
Discriminative Subnetworks with Regularized Spectral Learning for
Global-state Network Data | Data mining practitioners are facing challenges from data with network
structure. In this paper, we address a specific class of global-state networks
which comprises of a set of network instances sharing a similar structure yet
having different values at local nodes. Each instance is associated with a
global state which indicates the occurrence of an event. The objective is to
uncover a small set of discriminative subnetworks that can optimally classify
global network values. Unlike most existing studies which explore an
exponential subnetwork space, we address this difficult problem by adopting a
space transformation approach. Specifically, we present an algorithm that
optimizes a constrained dual-objective function to learn a low-dimensional
subspace that is capable of discriminating networks labelled by different
global states, while reconciling with common network topology sharing across
instances. Our algorithm takes an appealing approach from spectral graph
learning and we show that the globally optimum solution can be achieved via
matrix eigen-decomposition.
|
Poseidon: A System Architecture for Efficient GPU-based Deep Learning on
Multiple Machines | Deep learning (DL) has achieved notable successes in many machine learning
tasks. A number of frameworks have been developed to expedite the process of
designing and training deep neural networks (DNNs), such as Caffe, Torch and
Theano. Currently they can harness multiple GPUs on a single machine, but are
unable to use GPUs that are distributed across multiple machines; as even
average-sized DNNs can take days to train on a single GPU with 100s of GBs to
TBs of data, distributed GPUs present a prime opportunity for scaling up DL.
However, the limited bandwidth available on commodity Ethernet networks
presents a bottleneck to distributed GPU training, and prevents its trivial
realization.
To investigate how to adapt existing frameworks to efficiently support
distributed GPUs, we propose Poseidon, a scalable system architecture for
distributed inter-machine communication in existing DL frameworks. We integrate
Poseidon with Caffe and evaluate its performance at training DNNs for object
recognition. Poseidon features three key contributions that accelerate DNN
training on clusters: (1) a three-level hybrid architecture that allows
Poseidon to support both CPU-only and GPU-equipped clusters, (2) a distributed
wait-free backpropagation (DWBP) algorithm to improve GPU utilization and to
balance communication, and (3) a structure-aware communication protocol (SACP)
to minimize communication overheads. We empirically show that Poseidon
converges to same objectives as a single machine, and achieves state-of-art
training speedup across multiple models and well-established datasets using a
commodity GPU cluster of 8 nodes (e.g. 4.5x speedup on AlexNet, 4x on
GoogLeNet, 4x on CIFAR-10). On the much larger ImageNet22K dataset, Poseidon
with 8 nodes achieves better speedup and competitive accuracy to recent
CPU-based distributed systems such as Adam and Le et al., which use 10s to
1000s of nodes.
|
A new robust adaptive algorithm for underwater acoustic channel
equalization | We introduce a novel family of adaptive robust equalizers for highly
challenging underwater acoustic (UWA) channel equalization. Since the
underwater environment is highly non-stationary and subjected to impulsive
noise, we use adaptive filtering techniques based on a relative logarithmic
cost function inspired by the competitive methods from the online learning
literature. To improve the convergence performance of the conventional linear
equalization methods, while mitigating the stability issues, we intrinsically
combine different norms of the error in the cost function, using logarithmic
functions. Hence, we achieve a comparable convergence performance to least mean
fourth (LMF) equalizer, while significantly enhancing the stability performance
in such an adverse communication medium. We demonstrate the performance of our
algorithms through highly realistic experiments performed on accurately
simulated underwater acoustic channels.
|
Using machine learning for medium frequency derivative portfolio trading | We use machine learning for designing a medium frequency trading strategy for
a portfolio of 5 year and 10 year US Treasury note futures. We formulate this
as a classification problem where we predict the weekly direction of movement
of the portfolio using features extracted from a deep belief network trained on
technical indicators of the portfolio constituents. The experimentation shows
that the resulting pipeline is effective in making a profitable trade.
|
The Limitations of Optimization from Samples | In this paper we consider the following question: can we optimize objective
functions from the training data we use to learn them? We formalize this
question through a novel framework we call optimization from samples (OPS). In
OPS, we are given sampled values of a function drawn from some distribution and
the objective is to optimize the function under some constraint.
While there are interesting classes of functions that can be optimized from
samples, our main result is an impossibility. We show that there are classes of
functions which are statistically learnable and optimizable, but for which no
reasonable approximation for optimization from samples is achievable. In
particular, our main result shows that there is no constant factor
approximation for maximizing coverage functions under a cardinality constraint
using polynomially-many samples drawn from any distribution.
We also show tight approximation guarantees for maximization under a
cardinality constraint of several interesting classes of functions including
unit-demand, additive, and general monotone submodular functions, as well as a
constant factor approximation for monotone submodular functions with bounded
curvature.
|
A Mathematical Theory of Deep Convolutional Neural Networks for Feature
Extraction | Deep convolutional neural networks have led to breakthrough results in
numerous practical machine learning tasks such as classification of images in
the ImageNet data set, control-policy-learning to play Atari games or the board
game Go, and image captioning. Many of these applications first perform feature
extraction and then feed the results thereof into a trainable classifier. The
mathematical analysis of deep convolutional neural networks for feature
extraction was initiated by Mallat, 2012. Specifically, Mallat considered
so-called scattering networks based on a wavelet transform followed by the
modulus non-linearity in each network layer, and proved translation invariance
(asymptotically in the wavelet scale parameter) and deformation stability of
the corresponding feature extractor. This paper complements Mallat's results by
developing a theory that encompasses general convolutional transforms, or in
more technical parlance, general semi-discrete frames (including
Weyl-Heisenberg filters, curvelets, shearlets, ridgelets, wavelets, and learned
filters), general Lipschitz-continuous non-linearities (e.g., rectified linear
units, shifted logistic sigmoids, hyperbolic tangents, and modulus functions),
and general Lipschitz-continuous pooling operators emulating, e.g.,
sub-sampling and averaging. In addition, all of these elements can be different
in different network layers. For the resulting feature extractor we prove a
translation invariance result of vertical nature in the sense of the features
becoming progressively more translation-invariant with increasing network
depth, and we establish deformation sensitivity bounds that apply to signal
classes such as, e.g., band-limited functions, cartoon functions, and Lipschitz
functions.
|
Kernel principal component analysis network for image classification | In order to classify the nonlinear feature with linear classifier and improve
the classification accuracy, a deep learning network named kernel principal
component analysis network (KPCANet) is proposed. First, mapping the data into
higher space with kernel principal component analysis to make the data linearly
separable. Then building a two-layer KPCANet to obtain the principal components
of image. Finally, classifying the principal components with linearly
classifier. Experimental results show that the proposed KPCANet is effective in
face recognition, object recognition and hand-writing digits recognition, it
also outperforms principal component analysis network (PCANet) generally as
well. Besides, KPCANet is invariant to illumination and stable to occlusion and
slight deformation.
|
Revisiting Differentially Private Regression: Lessons From Learning
Theory and their Consequences | Private regression has received attention from both database and security
communities. Recent work by Fredrikson et al. (USENIX Security 2014) analyzed
the functional mechanism (Zhang et al. VLDB 2012) for training linear
regression models over medical data. Unfortunately, they found that model
accuracy is already unacceptable with differential privacy when $\varepsilon =
5$. We address this issue, presenting an explicit connection between
differential privacy and stable learning theory through which a substantially
better privacy/utility tradeoff can be obtained. Perhaps more importantly, our
theory reveals that the most basic mechanism in differential privacy, output
perturbation, can be used to obtain a better tradeoff for all
convex-Lipschitz-bounded learning tasks. Since output perturbation is simple to
implement, it means that our approach is potentially widely applicable in
practice. We go on to apply it on the same medical data as used by Fredrikson
et al. Encouragingly, we achieve accurate models even for $\varepsilon = 0.1$.
In the last part of this paper, we study the impact of our improved
differentially private mechanisms on model inversion attacks, a privacy attack
introduced by Fredrikson et al. We observe that the improved tradeoff makes the
resulting differentially private model more susceptible to inversion attacks.
We analyze this phenomenon formally.
|
Behavioral Modeling for Churn Prediction: Early Indicators and Accurate
Predictors of Custom Defection and Loyalty | Churn prediction, or the task of identifying customers who are likely to
discontinue use of a service, is an important and lucrative concern of firms in
many different industries. As these firms collect an increasing amount of
large-scale, heterogeneous data on the characteristics and behaviors of
customers, new methods become possible for predicting churn. In this paper, we
present a unified analytic framework for detecting the early warning signs of
churn, and assigning a "Churn Score" to each customer that indicates the
likelihood that the particular individual will churn within a predefined amount
of time. This framework employs a brute force approach to feature engineering,
then winnows the set of relevant attributes via feature selection, before
feeding the final feature-set into a suite of supervised learning algorithms.
Using several terabytes of data from a large mobile phone network, our method
identifies several intuitive - and a few surprising - early warning signs of
churn, and our best model predicts whether a subscriber will churn with 89.4%
accuracy.
|
ATD: Anomalous Topic Discovery in High Dimensional Discrete Data | We propose an algorithm for detecting patterns exhibited by anomalous
clusters in high dimensional discrete data. Unlike most anomaly detection (AD)
methods, which detect individual anomalies, our proposed method detects groups
(clusters) of anomalies; i.e. sets of points which collectively exhibit
abnormal patterns. In many applications this can lead to better understanding
of the nature of the atypical behavior and to identifying the sources of the
anomalies. Moreover, we consider the case where the atypical patterns exhibit
on only a small (salient) subset of the very high dimensional feature space.
Individual AD techniques and techniques that detect anomalies using all the
features typically fail to detect such anomalies, but our method can detect
such instances collectively, discover the shared anomalous patterns exhibited
by them, and identify the subsets of salient features. In this paper, we focus
on detecting anomalous topics in a batch of text documents, developing our
algorithm based on topic models. Results of our experiments show that our
method can accurately detect anomalous topics and salient features (words)
under each such topic in a synthetic data set and two real-world text corpora
and achieves better performance compared to both standard group AD and
individual AD techniques. All required code to reproduce our experiments is
available from https://github.com/hsoleimani/ATD
|
Backward and Forward Language Modeling for Constrained Sentence
Generation | Recent language models, especially those based on recurrent neural networks
(RNNs), make it possible to generate natural language from a learned
probability. Language generation has wide applications including machine
translation, summarization, question answering, conversation systems, etc.
Existing methods typically learn a joint probability of words conditioned on
additional information, which is (either statically or dynamically) fed to
RNN's hidden layer. In many applications, we are likely to impose hard
constraints on the generated texts, i.e., a particular word must appear in the
sentence. Unfortunately, existing approaches could not solve this problem. In
this paper, we propose a novel backward and forward language model. Provided a
specific word, we use RNNs to generate previous words and future words, either
simultaneously or asynchronously, resulting in two model variants. In this way,
the given word could appear at any position in the sentence. Experimental
results show that the generated texts are comparable to sequential LMs in
quality.
|
Deep Learning for Surface Material Classification Using Haptic And
Visual Information | When a user scratches a hand-held rigid tool across an object surface, an
acceleration signal can be captured, which carries relevant information about
the surface. More importantly, such a haptic signal is complementary to the
visual appearance of the surface, which suggests the combination of both
modalities for the recognition of the surface material. In this paper, we
present a novel deep learning method dealing with the surface material
classification problem based on a Fully Convolutional Network (FCN), which
takes as input the aforementioned acceleration signal and a corresponding image
of the surface texture. Compared to previous surface material classification
solutions, which rely on a careful design of hand-crafted domain-specific
features, our method automatically extracts discriminative features utilizing
the advanced deep learning methodologies. Experiments performed on the TUM
surface material database demonstrate that our method achieves state-of-the-art
classification accuracy robustly and efficiently.
|
Multilinear Subspace Clustering | In this paper we present a new model and an algorithm for unsupervised
clustering of 2-D data such as images. We assume that the data comes from a
union of multilinear subspaces (UOMS) model, which is a specific structured
case of the much studied union of subspaces (UOS) model. For segmentation under
this model, we develop Multilinear Subspace Clustering (MSC) algorithm and
evaluate its performance on the YaleB and Olivietti image data sets. We show
that MSC is highly competitive with existing algorithms employing the UOS model
in terms of clustering performance while enjoying improvement in computational
complexity.
|
GraphConnect: A Regularization Framework for Neural Networks | Deep neural networks have proved very successful in domains where large
training sets are available, but when the number of training samples is small,
their performance suffers from overfitting. Prior methods of reducing
overfitting such as weight decay, Dropout and DropConnect are data-independent.
This paper proposes a new method, GraphConnect, that is data-dependent, and is
motivated by the observation that data of interest lie close to a manifold. The
new method encourages the relationships between the learned decisions to
resemble a graph representing the manifold structure. Essentially GraphConnect
is designed to learn attributes that are present in data samples in contrast to
weight decay, Dropout and DropConnect which are simply designed to make it more
difficult to fit to random error or noise. Empirical Rademacher complexity is
used to connect the generalization error of the neural network to spectral
properties of the graph learned from the input data. This framework is used to
show that GraphConnect is superior to weight decay. Experimental results on
several benchmark datasets validate the theoretical analysis, and show that
when the number of training samples is small, GraphConnect is able to
significantly improve performance over weight decay.
|
Predicting the Co-Evolution of Event and Knowledge Graphs | Embedding learning, a.k.a. representation learning, has been shown to be able
to model large-scale semantic knowledge graphs. A key concept is a mapping of
the knowledge graph to a tensor representation whose entries are predicted by
models using latent representations of generalized entities. Knowledge graphs
are typically treated as static: A knowledge graph grows more links when more
facts become available but the ground truth values associated with links is
considered time invariant. In this paper we address the issue of knowledge
graphs where triple states depend on time. We assume that changes in the
knowledge graph always arrive in form of events, in the sense that the events
are the gateway to the knowledge graph. We train an event prediction model
which uses both knowledge graph background information and information on
recent events. By predicting future events, we also predict likely changes in
the knowledge graph and thus obtain a model for the evolution of the knowledge
graph as well. Our experiments demonstrate that our approach performs well in a
clinical application, a recommendation engine and a sensor network application.
|
A C++ library for Multimodal Deep Learning | MDL, Multimodal Deep Learning Library, is a deep learning framework that
supports multiple models, and this document explains its philosophy and
functionality. MDL runs on Linux, Mac, and Unix platforms. It depends on
OpenCV.
|
On the Differential Privacy of Bayesian Inference | We study how to communicate findings of Bayesian inference to third parties,
while preserving the strong guarantee of differential privacy. Our main
contributions are four different algorithms for private Bayesian inference on
proba-bilistic graphical models. These include two mechanisms for adding noise
to the Bayesian updates, either directly to the posterior parameters, or to
their Fourier transform so as to preserve update consistency. We also utilise a
recently introduced posterior sampling mechanism, for which we prove bounds for
the specific but general case of discrete Bayesian networks; and we introduce a
maximum-a-posteriori private mechanism. Our analysis includes utility and
privacy bounds, with a novel focus on the influence of graph structure on
privacy. Worked examples and experiments with Bayesian na{\"i}ve Bayes and
Bayesian linear regression illustrate the application of our mechanisms.
|
FAASTA: A fast solver for total-variation regularization of
ill-conditioned problems with application to brain imaging | The total variation (TV) penalty, as many other analysis-sparsity problems,
does not lead to separable factors or a proximal operatorwith a closed-form
expression, such as soft thresholding for the $\ell\_1$ penalty. As a result,
in a variational formulation of an inverse problem or statisticallearning
estimation, it leads to challenging non-smooth optimization problemsthat are
often solved with elaborate single-step first-order methods. When thedata-fit
term arises from empirical measurements, as in brain imaging, it isoften very
ill-conditioned and without simple structure. In this situation, in proximal
splitting methods, the computation cost of thegradient step can easily dominate
each iteration. Thus it is beneficialto minimize the number of gradient
steps.We present fAASTA, a variant of FISTA, that relies on an internal solver
forthe TV proximal operator, and refines its tolerance to balance
computationalcost of the gradient and the proximal steps. We give benchmarks
andillustrations on "brain decoding": recovering brain maps from
noisymeasurements to predict observed behavior. The algorithm as well as
theempirical study of convergence speed are valuable for any non-exact
proximaloperator, in particular analysis-sparsity problems.
|
Implementation of deep learning algorithm for automatic detection of
brain tumors using intraoperative IR-thermal mapping data | The efficiency of deep machine learning for automatic delineation of tumor
areas has been demonstrated for intraoperative neuronavigation using active
IR-mapping with the use of the cold test. The proposed approach employs a
matrix IR-imager to remotely register the space-time distribution of surface
temperature pattern, which is determined by the dynamics of local cerebral
blood flow. The advantages of this technique are non-invasiveness, zero risks
for the health of patients and medical staff, low implementation and
operational costs, ease and speed of use. Traditional IR-diagnostic technique
has a crucial limitation - it involves a diagnostician who determines the
boundaries of tumor areas, which gives rise to considerable uncertainty, which
can lead to diagnosis errors that are difficult to control. The current study
demonstrates that implementing deep learning algorithms allows to eliminate the
explained drawback.
|
Move from Perturbed scheme to exponential weighting average | In an online decision problem, one makes decisions often with a pool of
decision sequence called experts but without knowledge of the future. After
each step, one pays a cost based on the decision and observed rate. One
reasonal goal would be to perform as well as the best expert in the pool. The
modern and well-known way to attain this goal is the algorithm of exponential
weighting. However, recently, another algorithm called follow the perturbed
leader is developed and achieved about the same performance. In our work, we
first show the properties shared in common by the two algorithms which explain
the similarities on the performance. Next we will show that for a specific
perturbation, the two algorithms are identical. Finally, we show with some
examples that follow-the-leader style algorithms extend naturally to a large
class of structured online problems for which the exponential algorithms are
inefficient.
|
Recent Advances in Convolutional Neural Networks | In the last few years, deep learning has led to very good performance on a
variety of problems, such as visual recognition, speech recognition and natural
language processing. Among different types of deep neural networks,
convolutional neural networks have been most extensively studied. Leveraging on
the rapid growth in the amount of the annotated data and the great improvements
in the strengths of graphics processor units, the research on convolutional
neural networks has been emerged swiftly and achieved state-of-the-art results
on various tasks. In this paper, we provide a broad survey of the recent
advances in convolutional neural networks. We detailize the improvements of CNN
on different aspects, including layer design, activation function, loss
function, regularization, optimization and fast computation. Besides, we also
introduce various applications of convolutional neural networks in computer
vision, speech and natural language processing.
|
Refined Error Bounds for Several Learning Algorithms | This article studies the achievable guarantees on the error rates of certain
learning algorithms, with particular focus on refining logarithmic factors.
Many of the results are based on a general technique for obtaining bounds on
the error rates of sample-consistent classifiers with monotonic error regions,
in the realizable case. We prove bounds of this type expressed in terms of
either the VC dimension or the sample compression size. This general technique
also enables us to derive several new bounds on the error rates of general
sample-consistent learning algorithms, as well as refined bounds on the label
complexity of the CAL active learning algorithm. Additionally, we establish a
simple necessary and sufficient condition for the existence of a
distribution-free bound on the error rates of all sample-consistent learning
rules, converging at a rate inversely proportional to the sample size. We also
study learning in the presence of classification noise, deriving a new excess
error rate guarantee for general VC classes under Tsybakov's noise condition,
and establishing a simple and general necessary and sufficient condition for
the minimax excess risk under bounded noise to converge at a rate inversely
proportional to the sample size.
|
Feature Selection for Classification under Anonymity Constraint | Over the last decade, proliferation of various online platforms and their
increasing adoption by billions of users have heightened the privacy risk of a
user enormously. In fact, security researchers have shown that sparse microdata
containing information about online activities of a user although anonymous,
can still be used to disclose the identity of the user by cross-referencing the
data with other data sources. To preserve the privacy of a user, in existing
works several methods (k-anonymity, l-diversity, differential privacy) are
proposed that ensure a dataset which is meant to share or publish bears small
identity disclosure risk. However, the majority of these methods modify the
data in isolation, without considering their utility in subsequent knowledge
discovery tasks, which makes these datasets less informative. In this work, we
consider labeled data that are generally used for classification, and propose
two methods for feature selection considering two goals: first, on the reduced
feature set the data has small disclosure risk, and second, the utility of the
data is preserved for performing a classification task. Experimental results on
various real-world datasets show that the method is effective and useful in
practice.
|
Latent Variable Modeling with Diversity-Inducing Mutual Angular
Regularization | Latent Variable Models (LVMs) are a large family of machine learning models
providing a principled and effective way to extract underlying patterns,
structure and knowledge from observed data. Due to the dramatic growth of
volume and complexity of data, several new challenges have emerged and cannot
be effectively addressed by existing LVMs: (1) How to capture long-tail
patterns that carry crucial information when the popularity of patterns is
distributed in a power-law fashion? (2) How to reduce model complexity and
computational cost without compromising the modeling power of LVMs? (3) How to
improve the interpretability and reduce the redundancy of discovered patterns?
To addresses the three challenges discussed above, we develop a novel
regularization technique for LVMs, which controls the geometry of the latent
space during learning to enable the learned latent components of LVMs to be
diverse in the sense that they are favored to be mutually different from each
other, to accomplish long-tail coverage, low redundancy, and better
interpretability. We propose a mutual angular regularizer (MAR) to encourage
the components in LVMs to have larger mutual angles. The MAR is non-convex and
non-smooth, entailing great challenges for optimization. To cope with this
issue, we derive a smooth lower bound of the MAR and optimize the lower bound
instead. We show that the monotonicity of the lower bound is closely aligned
with the MAR to qualify the lower bound as a desirable surrogate of the MAR.
Using neural network (NN) as an instance, we analyze how the MAR affects the
generalization performance of NN. On two popular latent variable models ---
restricted Boltzmann machine and distance metric learning, we demonstrate that
MAR can effectively capture long-tail patterns, reduce model complexity without
sacrificing expressivity and improve interpretability.
|
A Deep Generative Deconvolutional Image Model | A deep generative model is developed for representation and analysis of
images, based on a hierarchical convolutional dictionary-learning framework.
Stochastic {\em unpooling} is employed to link consecutive layers in the model,
yielding top-down image generation. A Bayesian support vector machine is linked
to the top-layer features, yielding max-margin discrimination. Deep
deconvolutional inference is employed when testing, to infer the latent
features, and the top-layer features are connected with the max-margin
classifier for discrimination tasks. The model is efficiently trained using a
Monte Carlo expectation-maximization (MCEM) algorithm, with implementation on
graphical processor units (GPUs) for efficient large-scale learning, and fast
testing. Excellent results are obtained on several benchmark datasets,
including ImageNet, demonstrating that the proposed model achieves results that
are highly competitive with similarly sized convolutional neural networks.
|
Adaptive Algorithms for Online Convex Optimization with Long-term
Constraints | We present an adaptive online gradient descent algorithm to solve online
convex optimization problems with long-term constraints , which are constraints
that need to be satisfied when accumulated over a finite number of rounds T ,
but can be violated in intermediate rounds. For some user-defined trade-off
parameter $\beta$ $\in$ (0, 1), the proposed algorithm achieves cumulative
regret bounds of O(T^max{$\beta$,1--$\beta$}) and O(T^(1--$\beta$/2)) for the
loss and the constraint violations respectively. Our results hold for convex
losses and can handle arbitrary convex constraints without requiring knowledge
of the number of rounds in advance. Our contributions improve over the best
known cumulative regret bounds by Mahdavi, et al. (2012) that are respectively
O(T^1/2) and O(T^3/4) for general convex domains, and respectively O(T^2/3) and
O(T^2/3) when further restricting to polyhedral domains. We supplement the
analysis with experiments validating the performance of our algorithm in
practice.
|
Adaptive Ensemble Learning with Confidence Bounds | Extracting actionable intelligence from distributed, heterogeneous,
correlated and high-dimensional data sources requires run-time processing and
learning both locally and globally. In the last decade, a large number of
meta-learning techniques have been proposed in which local learners make online
predictions based on their locally-collected data instances, and feed these
predictions to an ensemble learner, which fuses them and issues a global
prediction. However, most of these works do not provide performance guarantees
or, when they do, these guarantees are asymptotic. None of these existing works
provide confidence estimates about the issued predictions or rate of learning
guarantees for the ensemble learner. In this paper, we provide a systematic
ensemble learning method called Hedged Bandits, which comes with both long run
(asymptotic) and short run (rate of learning) performance guarantees. Moreover,
our approach yields performance guarantees with respect to the optimal local
prediction strategy, and is also able to adapt its predictions in a data-driven
manner. We illustrate the performance of Hedged Bandits in the context of
medical informatics and show that it outperforms numerous online and offline
ensemble learning methods.
|
A Latent-Variable Lattice Model | Markov random field (MRF) learning is intractable, and its approximation
algorithms are computationally expensive. We target a small subset of MRF that
is used frequently in computer vision. We characterize this subset with three
concepts: Lattice, Homogeneity, and Inertia; and design a non-markov model as
an alternative. Our goal is robust learning from small datasets. Our learning
algorithm uses vector quantization and, at time complexity O(U log U) for a
dataset of U pixels, is much faster than that of general-purpose MRF.
|
Satisficing in multi-armed bandit problems | Satisficing is a relaxation of maximizing and allows for less risky decision
making in the face of uncertainty. We propose two sets of satisficing
objectives for the multi-armed bandit problem, where the objective is to
achieve reward-based decision-making performance above a given threshold. We
show that these new problems are equivalent to various standard multi-armed
bandit problems with maximizing objectives and use the equivalence to find
bounds on performance. The different objectives can result in qualitatively
different behavior; for example, agents explore their options continually in
one case and only a finite number of times in another. For the case of Gaussian
rewards we show an additional equivalence between the two sets of satisficing
objectives that allows algorithms developed for one set to be applied to the
other. We then develop variants of the Upper Credible Limit (UCL) algorithm
that solve the problems with satisficing objectives and show that these
modified UCL algorithms achieve efficient satisficing performance.
|
The Max $K$-Armed Bandit: PAC Lower Bounds and Efficient Algorithms | We consider the Max $K$-Armed Bandit problem, where a learning agent is faced
with several stochastic arms, each a source of i.i.d. rewards of unknown
distribution. At each time step the agent chooses an arm, and observes the
reward of the obtained sample. Each sample is considered here as a separate
item with the reward designating its value, and the goal is to find an item
with the highest possible value. Our basic assumption is a known lower bound on
the {\em tail function} of the reward distributions. Under the PAC framework,
we provide a lower bound on the sample complexity of any
$(\epsilon,\delta)$-correct algorithm, and propose an algorithm that attains
this bound up to logarithmic factors. We analyze the robustness of the proposed
algorithm and in addition, we compare the performance of this algorithm to the
variant in which the arms are not distinguishable by the agent and are chosen
randomly at each stage. Interestingly, when the maximal rewards of the arms
happen to be similar, the latter approach may provide better performance.
|
Deep Reinforcement Learning in Large Discrete Action Spaces | Being able to reason in an environment with a large number of discrete
actions is essential to bringing reinforcement learning to a larger class of
problems. Recommender systems, industrial plants and language models are only
some of the many real-world tasks involving large numbers of discrete actions
for which current methods are difficult or even often impossible to apply. An
ability to generalize over the set of actions as well as sub-linear complexity
relative to the size of the set are both necessary to handle such tasks.
Current approaches are not able to provide both of these, which motivates the
work in this paper. Our proposed approach leverages prior information about the
actions to embed them in a continuous space upon which it can generalize.
Additionally, approximate nearest-neighbor methods allow for logarithmic-time
lookup complexity relative to the number of actions, which is necessary for
time-wise tractable training. This combined approach allows reinforcement
learning methods to be applied to large-scale learning problems previously
intractable with current methods. We demonstrate our algorithm's abilities on a
series of tasks having up to one million actions.
|
Fast Parallel SVM using Data Augmentation | As one of the most popular classifiers, linear SVMs still have challenges in
dealing with very large-scale problems, even though linear or sub-linear
algorithms have been developed recently on single machines. Parallel computing
methods have been developed for learning large-scale SVMs. However, existing
methods rely on solving local sub-optimization problems. In this paper, we
develop a novel parallel algorithm for learning large-scale linear SVM. Our
approach is based on a data augmentation equivalent formulation, which casts
the problem of learning SVM as a Bayesian inference problem, for which we can
develop very efficient parallel sampling methods. We provide empirical results
for this parallel sampling SVM, and provide extensions for SVR, non-linear
kernels, and provide a parallel implementation of the Crammer and Singer model.
This approach is very promising in its own right, and further is a very useful
technique to parallelize a broader family of general maximum-margin models.
|
Real-Time Audio-to-Score Alignment of Music Performances Containing
Errors and Arbitrary Repeats and Skips | This paper discusses real-time alignment of audio signals of music
performance to the corresponding score (a.k.a. score following) which can
handle tempo changes, errors and arbitrary repeats and/or skips (repeats/skips)
in performances. This type of score following is particularly useful in
automatic accompaniment for practices and rehearsals, where errors and
repeats/skips are often made. Simple extensions of the algorithms previously
proposed in the literature are not applicable in these situations for scores of
practical length due to the problem of large computational complexity. To cope
with this problem, we present two hidden Markov models of monophonic
performance with errors and arbitrary repeats/skips, and derive efficient
score-following algorithms with an assumption that the prior probability
distributions of score positions before and after repeats/skips are independent
from each other. We confirmed real-time operation of the algorithms with music
scores of practical length (around 10000 notes) on a modern laptop and their
tracking ability to the input performance within 0.7 s on average after
repeats/skips in clarinet performance data. Further improvements and extension
for polyphonic signals are also discussed.
|
The Lov\'asz Hinge: A Novel Convex Surrogate for Submodular Losses | Learning with non-modular losses is an important problem when sets of
predictions are made simultaneously. The main tools for constructing convex
surrogate loss functions for set prediction are margin rescaling and slack
rescaling. In this work, we show that these strategies lead to tight convex
surrogates iff the underlying loss function is increasing in the number of
incorrect predictions. However, gradient or cutting-plane computation for these
functions is NP-hard for non-supermodular loss functions. We propose instead a
novel surrogate loss function for submodular losses, the Lov\'asz hinge, which
leads to O(p log p) complexity with O(p) oracle accesses to the loss function
to compute a gradient or cutting-plane. We prove that the Lov\'asz hinge is
convex and yields an extension. As a result, we have developed the first
tractable convex surrogates in the literature for submodular losses. We
demonstrate the utility of this novel convex surrogate through several set
prediction tasks, including on the PASCAL VOC and Microsoft COCO datasets.
|
Visualizations Relevant to The User By Multi-View Latent Variable
Factorization | A main goal of data visualization is to find, from among all the available
alternatives, mappings to the 2D/3D display which are relevant to the user.
Assuming user interaction data, or other auxiliary data about the items or
their relationships, the goal is to identify which aspects in the primary data
support the user\'s input and, equally importantly, which aspects of the
user\'s potentially noisy input have support in the primary data. For solving
the problem, we introduce a multi-view embedding in which a latent
factorization identifies which aspects in the two data views (primary data and
user data) are related and which are specific to only one of them. The
factorization is a generative model in which the display is parameterized as a
part of the factorization and the other factors explain away the aspects not
expressible in a two-dimensional display. Functioning of the model is
demonstrated on several data sets.
|
Implementing a Bayes Filter in a Neural Circuit: The Case of Unknown
Stimulus Dynamics | In order to interact intelligently with objects in the world, animals must
first transform neural population responses into estimates of the dynamic,
unknown stimuli which caused them. The Bayesian solution to this problem is
known as a Bayes filter, which applies Bayes' rule to combine population
responses with the predictions of an internal model. In this paper we present a
method for learning to approximate a Bayes filter when the stimulus dynamics
are unknown. To do this we use the inferential properties of probabilistic
population codes to compute Bayes' rule, and train a neural network to compute
approximate predictions by the method of maximum likelihood. In particular, we
perform stochastic gradient descent on the negative log-likelihood with a novel
approximation of the gradient. We demonstrate our methods on a finite-state, a
linear, and a nonlinear filtering problem, and show how the hidden layer of the
neural network develops tuning curves which are consistent with findings in
experimental neuroscience.
|
Context-Based Prediction of App Usage | There are around a hundred installed apps on an average smartphone. The high
number of apps and the limited number of app icons that can be displayed on the
device's screen requires a new paradigm to address their visibility to the
user. In this paper we propose a new online algorithm for dynamically
predicting a set of apps that the user is likely to use. The algorithm runs on
the user's device and constantly learns the user's habits at a given time,
location, and device state. It is designed to actively help the user to
navigate to the desired app as well as to provide a personalized feeling, and
hence is aimed at maximizing the AUC. We show both theoretically and
empirically that the algorithm maximizes the AUC, and yields good results on a
set of 1,000 devices.
|
An unsupervised spatiotemporal graphical modeling approach to anomaly
detection in distributed CPS | Modern distributed cyber-physical systems (CPSs) encounter a large variety of
physical faults and cyber anomalies and in many cases, they are vulnerable to
catastrophic fault propagation scenarios due to strong connectivity among the
sub-systems. This paper presents a new data-driven framework for system-wide
anomaly detection for addressing such issues. The framework is based on a
spatiotemporal feature extraction scheme built on the concept of symbolic
dynamics for discovering and representing causal interactions among the
subsystems of a CPS. The extracted spatiotemporal features are then used to
learn system-wide patterns via a Restricted Boltzmann Machine (RBM). The
results show that: (1) the RBM free energy in the off-nominal conditions is
different from that in the nominal conditions and can be used for anomaly
detection; (2) the framework can capture multiple nominal modes with one
graphical model; (3) the case studies with simulated data and an integrated
building system validate the proposed approach.
|
Bridging the Gap between Stochastic Gradient MCMC and Stochastic
Optimization | Stochastic gradient Markov chain Monte Carlo (SG-MCMC) methods are Bayesian
analogs to popular stochastic optimization methods; however, this connection is
not well studied. We explore this relationship by applying simulated annealing
to an SGMCMC algorithm. Furthermore, we extend recent SG-MCMC methods with two
key components: i) adaptive preconditioners (as in ADAgrad or RMSprop), and ii)
adaptive element-wise momentum weights. The zero-temperature limit gives a
novel stochastic optimization method with adaptive element-wise momentum
weights, while conventional optimization methods only have a shared, static
momentum weight. Under certain assumptions, our theoretical analysis suggests
the proposed simulated annealing approach converges close to the global optima.
Experiments on several deep neural network models show state-of-the-art results
compared to related stochastic optimization algorithms.
|
Inducing Generalized Multi-Label Rules with Learning Classifier Systems | In recent years, multi-label classification has attracted a significant body
of research, motivated by real-life applications, such as text classification
and medical diagnoses. Although sparsely studied in this context, Learning
Classifier Systems are naturally well-suited to multi-label classification
problems, whose search space typically involves multiple highly specific
niches. This is the motivation behind our current work that introduces a
generalized multi-label rule format -- allowing for flexible label-dependency
modeling, with no need for explicit knowledge of which correlations to search
for -- and uses it as a guide for further adapting the general Michigan-style
supervised Learning Classifier System framework. The integration of the
aforementioned rule format and framework adaptations results in a novel
algorithm for multi-label classification whose behavior is studied through a
set of properly defined artificial problems. The proposed algorithm is also
thoroughly evaluated on a set of multi-label datasets and found competitive to
other state-of-the-art multi-label classification methods.
|
Discovering topic structures of a temporally evolving document corpus | In this paper we describe a novel framework for the discovery of the topical
content of a data corpus, and the tracking of its complex structural changes
across the temporal dimension. In contrast to previous work our model does not
impose a prior on the rate at which documents are added to the corpus nor does
it adopt the Markovian assumption which overly restricts the type of changes
that the model can capture. Our key technical contribution is a framework based
on (i) discretization of time into epochs, (ii) epoch-wise topic discovery
using a hierarchical Dirichlet process-based model, and (iii) a temporal
similarity graph which allows for the modelling of complex topic changes:
emergence and disappearance, evolution, splitting, and merging. The power of
the proposed framework is demonstrated on two medical literature corpora
concerned with the autism spectrum disorder (ASD) and the metabolic syndrome
(MetS) -- both increasingly important research subjects with significant social
and healthcare consequences. In addition to the collected ASD and metabolic
syndrome literature corpora which we made freely available, our contribution
also includes an extensive empirical analysis of the proposed framework. We
describe a detailed and careful examination of the effects that our
algorithms's free parameters have on its output, and discuss the significance
of the findings both in the context of the practical application of our
algorithm as well as in the context of the existing body of work on temporal
topic analysis. Our quantitative analysis is followed by several qualitative
case studies highly relevant to the current research on ASD and MetS, on which
our algorithm is shown to capture well the actual developments in these fields.
|
Statistical Learning under Nonstationary Mixing Processes | We study a special case of the problem of statistical learning without the
i.i.d. assumption. Specifically, we suppose a learning method is presented with
a sequence of data points, and required to make a prediction (e.g., a
classification) for each one, and can then observe the loss incurred by this
prediction. We go beyond traditional analyses, which have focused on stationary
mixing processes or nonstationary product processes, by combining these two
relaxations to allow nonstationary mixing processes. We are particularly
interested in the case of $\beta$-mixing processes, with the sum of changes in
marginal distributions growing sublinearly in the number of samples. Under
these conditions, we propose a learning method, and establish that for bounded
VC subgraph classes, the cumulative excess risk grows sublinearly in the number
of predictions, at a quantified rate.
|
Inverse Reinforcement Learning via Deep Gaussian Process | We propose a new approach to inverse reinforcement learning (IRL) based on
the deep Gaussian process (deep GP) model, which is capable of learning
complicated reward structures with few demonstrations. Our model stacks
multiple latent GP layers to learn abstract representations of the state
feature space, which is linked to the demonstrations through the Maximum
Entropy learning framework. Incorporating the IRL engine into the nonlinear
latent structure renders existing deep GP inference approaches intractable. To
tackle this, we develop a non-standard variational approximation framework
which extends previous inference schemes. This allows for approximate Bayesian
treatment of the feature space and guards against overfitting. Carrying out
representation and inverse reinforcement learning simultaneously within our
model outperforms state-of-the-art approaches, as we demonstrate with
experiments on standard benchmarks ("object world","highway driving") and a new
benchmark ("binary world").
|
Regularized Orthogonal Tensor Decompositions for Multi-Relational
Learning | Multi-relational learning has received lots of attention from researchers in
various research communities. Most existing methods either suffer from
superlinear per-iteration cost, or are sensitive to the given ranks. To address
both issues, we propose a scalable core tensor trace norm Regularized
Orthogonal Iteration Decomposition (ROID) method for full or incomplete tensor
analytics, which can be generalized as a graph Laplacian regularized version by
using auxiliary information or a sparse higher-order orthogonal iteration
(SHOOI) version. We first induce the equivalence relation of the Schatten
p-norm (0<p<\infty) of a low multi-linear rank tensor and its core tensor. Then
we achieve a much smaller matrix trace norm minimization problem. Finally, we
develop two efficient augmented Lagrange multiplier algorithms to solve our
problems with convergence guarantees. Extensive experiments using both real and
synthetic datasets, even though with only a few observations, verified both the
efficiency and effectiveness of our methods.
|
The Utility of Abstaining in Binary Classification | We explore the problem of binary classification in machine learning, with a
twist - the classifier is allowed to abstain on any datum, professing ignorance
about the true class label without committing to any prediction. This is
directly motivated by applications like medical diagnosis and fraud risk
assessment, in which incorrect predictions have potentially calamitous
consequences. We focus on a recent spate of theoretically driven work in this
area that characterizes how allowing abstentions can lead to fewer errors in
very general settings. Two areas are highlighted: the surprising possibility of
zero-error learning, and the fundamental tradeoff between predicting
sufficiently often and avoiding incorrect predictions. We review efficient
algorithms with provable guarantees for each of these areas. We also discuss
connections to other scenarios, notably active learning, as they suggest
promising directions of further inquiry in this emerging field.
|
Self-Excitation: An Enabler for Online Thermal Estimation and Model
Predictive Control of Buildings | This paper investigates a method to improve buildings' thermal predictive
control performance via online identification and excitation (active learning
process) that minimally disrupts normal operations. In previous studies we have
demonstrated scalable methods to acquire multi-zone thermal models of passive
buildings using a gray-box approach that leverages building topology and
measurement data. Here we extend the method to multi-zone actively controlled
buildings and examine how to improve the thermal model estimation by using the
controller to excite unknown portions of the building's dynamics. Comparing
against a baseline thermostat controller, we demonstrate the utility of both
the initially acquired and improved thermal models within a Model Predictive
Control (MPC) framework, which anticipates weather uncertainty and time-varying
temperature set-points. A simulation study demonstrates self-excitation
improves model estimation, which corresponds to improved MPC energy savings and
occupant comfort. By coupling building topology, estimation, and control
routines into a single online framework, we have demonstrated the potential for
low-cost scalable methods to actively learn and control buildings to ensure
occupant comfort and minimize energy usage, all while using the existing
building's HVAC sensors and hardware.
|
Electricity Demand Forecasting by Multi-Task Learning | We explore the application of kernel-based multi-task learning techniques to
forecast the demand of electricity in multiple nodes of a distribution network.
We show that recently developed output kernel learning techniques are
particularly well suited to solve this problem, as they allow to flexibly model
the complex seasonal effects that characterize electricity demand data, while
learning and exploiting correlations between multiple demand profiles. We also
demonstrate that kernels with a multiplicative structure yield superior
predictive performance with respect to the widely adopted (generalized)
additive models. Our study is based on residential and industrial smart meter
data provided by the Irish Commission for Energy Regulation (CER).
|
New Perspectives on $k$-Support and Cluster Norms | We study a regularizer which is defined as a parameterized infimum of
quadratics, and which we call the box-norm. We show that the k-support norm, a
regularizer proposed by [Argyriou et al, 2012] for sparse vector prediction
problems, belongs to this family, and the box-norm can be generated as a
perturbation of the former. We derive an improved algorithm to compute the
proximity operator of the squared box-norm, and we provide a method to compute
the norm. We extend the norms to matrices, introducing the spectral k-support
norm and spectral box-norm. We note that the spectral box-norm is essentially
equivalent to the cluster norm, a multitask learning regularizer introduced by
[Jacob et al. 2009a], and which in turn can be interpreted as a perturbation of
the spectral k-support norm. Centering the norm is important for multitask
learning and we also provide a method to use centered versions of the norms as
regularizers. Numerical experiments indicate that the spectral k-support and
box-norms and their centered variants provide state of the art performance in
matrix completion and multitask learning problems respectively.
|
Robust Semi-supervised Least Squares Classification by Implicit
Constraints | We introduce the implicitly constrained least squares (ICLS) classifier, a
novel semi-supervised version of the least squares classifier. This classifier
minimizes the squared loss on the labeled data among the set of parameters
implied by all possible labelings of the unlabeled data. Unlike other
discriminative semi-supervised methods, this approach does not introduce
explicit additional assumptions into the objective function, but leverages
implicit assumptions already present in the choice of the supervised least
squares classifier. This method can be formulated as a quadratic programming
problem and its solution can be found using a simple gradient descent
procedure. We prove that, in a limited 1-dimensional setting, this approach
never leads to performance worse than the supervised classifier. Experimental
results show that also in the general multidimensional case performance
improvements can be expected, both in terms of the squared loss that is
intrinsic to the classifier, as well as in terms of the expected classification
error.
|
Using Causal Discovery to Track Information Flow in Spatio-Temporal Data
- A Testbed and Experimental Results Using Advection-Diffusion Simulations | Causal discovery algorithms based on probabilistic graphical models have
emerged in geoscience applications for the identification and visualization of
dynamical processes. The key idea is to learn the structure of a graphical
model from observed spatio-temporal data, which indicates information flow,
thus pathways of interactions, in the observed physical system. Studying those
pathways allows geoscientists to learn subtle details about the underlying
dynamical mechanisms governing our planet. Initial studies using this approach
on real-world atmospheric data have shown great potential for scientific
discovery. However, in these initial studies no ground truth was available, so
that the resulting graphs have been evaluated only by whether a domain expert
thinks they seemed physically plausible. This paper seeks to fill this gap. We
develop a testbed that emulates two dynamical processes dominant in many
geoscience applications, namely advection and diffusion, in a 2D grid. Then we
apply the causal discovery based information tracking algorithms to the
simulation data to study how well the algorithms work for different scenarios
and to gain a better understanding of the physical meaning of the graph
results, in particular of instantaneous connections. We make all data sets used
in this study available to the community as a benchmark.
Keywords: Information flow, graphical model, structure learning, causal
discovery, geoscience.
|
Natural Language Inference by Tree-Based Convolution and Heuristic
Matching | In this paper, we propose the TBCNN-pair model to recognize entailment and
contradiction between two sentences. In our model, a tree-based convolutional
neural network (TBCNN) captures sentence-level semantics; then heuristic
matching layers like concatenation, element-wise product/difference combine the
information in individual sentences. Experimental results show that our model
outperforms existing sentence encoding-based approaches by a large margin.
|
Convexified Modularity Maximization for Degree-corrected Stochastic
Block Models | The stochastic block model (SBM) is a popular framework for studying
community detection in networks. This model is limited by the assumption that
all nodes in the same community are statistically equivalent and have equal
expected degrees. The degree-corrected stochastic block model (DCSBM) is a
natural extension of SBM that allows for degree heterogeneity within
communities. This paper proposes a convexified modularity maximization approach
for estimating the hidden communities under DCSBM. Our approach is based on a
convex programming relaxation of the classical (generalized) modularity
maximization formulation, followed by a novel doubly-weighted $ \ell_1 $-norm $
k $-median procedure. We establish non-asymptotic theoretical guarantees for
both approximate clustering and perfect clustering. Our approximate clustering
results are insensitive to the minimum degree, and hold even in sparse regime
with bounded average degrees. In the special case of SBM, these theoretical
results match the best-known performance guarantees of computationally feasible
algorithms. Numerically, we provide an efficient implementation of our
algorithm, which is applied to both synthetic and real-world networks.
Experiment results show that our method enjoys competitive performance compared
to the state of the art in the literature.
|
Visually Indicated Sounds | Objects make distinctive sounds when they are hit or scratched. These sounds
reveal aspects of an object's material properties, as well as the actions that
produced them. In this paper, we propose the task of predicting what sound an
object makes when struck as a way of studying physical interactions within a
visual scene. We present an algorithm that synthesizes sound from silent videos
of people hitting and scratching objects with a drumstick. This algorithm uses
a recurrent neural network to predict sound features from videos and then
produces a waveform from these features with an example-based synthesis
procedure. We show that the sounds predicted by our model are realistic enough
to fool participants in a "real or fake" psychophysical experiment, and that
they convey significant information about material properties and physical
interactions.
|
Taming the Noise in Reinforcement Learning via Soft Updates | Model-free reinforcement learning algorithms, such as Q-learning, perform
poorly in the early stages of learning in noisy environments, because much
effort is spent unlearning biased estimates of the state-action value function.
The bias results from selecting, among several noisy estimates, the apparent
optimum, which may actually be suboptimal. We propose G-learning, a new
off-policy learning algorithm that regularizes the value estimates by
penalizing deterministic policies in the beginning of the learning process. We
show that this method reduces the bias of the value-function estimation,
leading to faster convergence to the optimal value and the optimal policy.
Moreover, G-learning enables the natural incorporation of prior domain
knowledge, when available. The stochastic nature of G-learning also makes it
avoid some exploration costs, a property usually attributed only to on-policy
algorithms. We illustrate these ideas in several examples, where G-learning
results in significant improvements of the convergence rate and the cost of the
learning process.
|
Structured Pruning of Deep Convolutional Neural Networks | Real time application of deep learning algorithms is often hindered by high
computational complexity and frequent memory accesses. Network pruning is a
promising technique to solve this problem. However, pruning usually results in
irregular network connections that not only demand extra representation efforts
but also do not fit well on parallel computation. We introduce structured
sparsity at various scales for convolutional neural networks, which are channel
wise, kernel wise and intra kernel strided sparsity. This structured sparsity
is very advantageous for direct computational resource savings on embedded
computers, parallel computing environments and hardware based systems. To
decide the importance of network connections and paths, the proposed method
uses a particle filtering approach. The importance weight of each particle is
assigned by computing the misclassification rate with corresponding
connectivity pattern. The pruned network is re-trained to compensate for the
losses due to pruning. While implementing convolutions as matrix products, we
particularly show that intra kernel strided sparsity with a simple constraint
can significantly reduce the size of kernel and feature map matrices. The
pruned network is finally fixed point optimized with reduced word length
precision. This results in significant reduction in the total storage size
providing advantages for on-chip memory based implementations of deep neural
networks.
|
Optimal Selective Attention in Reactive Agents | In POMDPs, information about the hidden state, delivered through
observations, is both valuable to the agent, allowing it to base its actions on
better informed internal states, and a "curse", exploding the size and
diversity of the internal state space. One attempt to deal with this is to
focus on reactive policies, that only base their actions on the most recent
observation. However, even reactive policies can be demanding on resources, and
agents need to pay selective attention to only some of the information
available to them in observations. In this report we present the
minimum-information principle for selective attention in reactive agents. We
further motivate this approach by reducing the general problem of optimal
control in POMDPs, to reactive control with complex observations. Lastly, we
explore a newly discovered phenomenon of this optimization process - period
doubling bifurcations. This necessitates periodic policies, and raises many
more questions regarding stability, periodicity and chaos in optimal control.
|
A Simple Baseline for Travel Time Estimation using Large-Scale Trip Data | The increased availability of large-scale trajectory data around the world
provides rich information for the study of urban dynamics. For example, New
York City Taxi Limousine Commission regularly releases source-destination
information about trips in the taxis they regulate. Taxi data provide
information about traffic patterns, and thus enable the study of urban flow --
what will traffic between two locations look like at a certain date and time in
the future? Existing big data methods try to outdo each other in terms of
complexity and algorithmic sophistication. In the spirit of "big data beats
algorithms", we present a very simple baseline which outperforms
state-of-the-art approaches, including Bing Maps and Baidu Maps (whose APIs
permit large scale experimentation). Such a travel time estimation baseline has
several important uses, such as navigation (fast travel time estimates can
serve as approximate heuristics for A search variants for path finding) and
trip planning (which uses operating hours for popular destinations along with
travel time estimates to create an itinerary).
|
Tight Bounds for Approximate Carath\'eodory and Beyond | We give a deterministic nearly-linear time algorithm for approximating any
point inside a convex polytope with a sparse convex combination of the
polytope's vertices. Our result provides a constructive proof for the
Approximate Carath\'{e}odory Problem, which states that any point inside a
polytope contained in the $\ell_p$ ball of radius $D$ can be approximated to
within $\epsilon$ in $\ell_p$ norm by a convex combination of only $O\left(D^2
p/\epsilon^2\right)$ vertices of the polytope for $p \geq 2$. We also show that
this bound is tight, using an argument based on anti-concentration for the
binomial distribution.
Along the way of establishing the upper bound, we develop a technique for
minimizing norms over convex sets with complicated geometry; this is achieved
by running Mirror Descent on a dual convex function obtained via Sion's
Theorem.
As simple extensions of our method, we then provide new algorithms for
submodular function minimization and SVM training. For submodular function
minimization we obtain a simplification and (provable) speed-up over Wolfe's
algorithm, the method commonly found to be the fastest in practice. For SVM
training, we obtain $O(1/\epsilon^2)$ convergence for arbitrary kernels; each
iteration only requires matrix-vector operations involving the kernel matrix,
so we overcome the obstacle of having to explicitly store the kernel or compute
its Cholesky factorization.
|
Feed-Forward Networks with Attention Can Solve Some Long-Term Memory
Problems | We propose a simplified model of attention which is applicable to
feed-forward neural networks and demonstrate that the resulting model can solve
the synthetic "addition" and "multiplication" long-term memory problems for
sequence lengths which are both longer and more widely varying than the best
published results for these tasks.
|
Matrix Completion Under Monotonic Single Index Models | Most recent results in matrix completion assume that the matrix under
consideration is low-rank or that the columns are in a union of low-rank
subspaces. In real-world settings, however, the linear structure underlying
these models is distorted by a (typically unknown) nonlinear transformation.
This paper addresses the challenge of matrix completion in the face of such
nonlinearities. Given a few observations of a matrix that are obtained by
applying a Lipschitz, monotonic function to a low rank matrix, our task is to
estimate the remaining unobserved entries. We propose a novel matrix completion
method that alternates between low-rank matrix estimation and monotonic
function estimation to estimate the missing matrix elements. Mean squared error
bounds provide insight into how well the matrix can be estimated based on the
size, rank of the matrix and properties of the nonlinear transformation.
Empirical results on synthetic and real-world datasets demonstrate the
competitiveness of the proposed approach.
|
Common Variable Learning and Invariant Representation Learning using
Siamese Neural Networks | We consider the statistical problem of learning common source of variability
in data which are synchronously captured by multiple sensors, and demonstrate
that Siamese neural networks can be naturally applied to this problem. This
approach is useful in particular in exploratory, data-driven applications,
where neither a model nor label information is available. In recent years, many
researchers have successfully applied Siamese neural networks to obtain an
embedding of data which corresponds to a "semantic similarity". We present an
interpretation of this "semantic similarity" as learning of equivalence
classes. We discuss properties of the embedding obtained by Siamese networks
and provide empirical results that demonstrate the ability of Siamese networks
to learn common variability.
|
Sparse group factor analysis for biclustering of multiple data sources | Motivation: Modelling methods that find structure in data are necessary with
the current large volumes of genomic data, and there have been various efforts
to find subsets of genes exhibiting consistent patterns over subsets of
treatments. These biclustering techniques have focused on one data source,
often gene expression data. We present a Bayesian approach for joint
biclustering of multiple data sources, extending a recent method Group Factor
Analysis (GFA) to have a biclustering interpretation with additional sparsity
assumptions. The resulting method enables data-driven detection of linear
structure present in parts of the data sources. Results: Our simulation studies
show that the proposed method reliably infers bi-clusters from heterogeneous
data sources. We tested the method on data from the NCI-DREAM drug sensitivity
prediction challenge, resulting in an excellent prediction accuracy. Moreover,
the predictions are based on several biclusters which provide insight into the
data sources, in this case on gene expression, DNA methylation, protein
abundance, exome sequence, functional connectivity fingerprints and drug
sensitivity.
|
Learning to Filter with Predictive State Inference Machines | Latent state space models are a fundamental and widely used tool for modeling
dynamical systems. However, they are difficult to learn from data and learned
models often lack performance guarantees on inference tasks such as filtering
and prediction. In this work, we present the PREDICTIVE STATE INFERENCE MACHINE
(PSIM), a data-driven method that considers the inference procedure on a
dynamical system as a composition of predictors. The key idea is that rather
than first learning a latent state space model, and then using the learned
model for inference, PSIM directly learns predictors for inference in
predictive state space. We provide theoretical guarantees for inference, in
both realizable and agnostic settings, and showcase practical performance on a
variety of simulated and real world robotics benchmarks.
|
Estimation of the sample covariance matrix from compressive measurements | This paper focuses on the estimation of the sample covariance matrix from
low-dimensional random projections of data known as compressive measurements.
In particular, we present an unbiased estimator to extract the covariance
structure from compressive measurements obtained by a general class of random
projection matrices consisting of i.i.d. zero-mean entries and finite first
four moments. In contrast to previous works, we make no structural assumptions
about the underlying covariance matrix such as being low-rank. In fact, our
analysis is based on a non-Bayesian data setting which requires no
distributional assumptions on the set of data samples. Furthermore, inspired by
the generality of the projection matrices, we propose an approach to covariance
estimation that utilizes sparse Rademacher matrices. Therefore, our algorithm
can be used to estimate the covariance matrix in applications with limited
memory and computation power at the acquisition devices. Experimental results
demonstrate that our approach allows for accurate estimation of the sample
covariance matrix on several real-world data sets, including video data.
|
Online Keyword Spotting with a Character-Level Recurrent Neural Network | In this paper, we propose a context-aware keyword spotting model employing a
character-level recurrent neural network (RNN) for spoken term detection in
continuous speech. The RNN is end-to-end trained with connectionist temporal
classification (CTC) to generate the probabilities of character and
word-boundary labels. There is no need for the phonetic transcription, senone
modeling, or system dictionary in training and testing. Also, keywords can
easily be added and modified by editing the text based keyword list without
retraining the RNN. Moreover, the unidirectional RNN processes an infinitely
long input audio streams without pre-segmentation and keywords are detected
with low-latency before the utterance is finished. Experimental results show
that the proposed keyword spotter significantly outperforms the deep neural
network (DNN) and hidden Markov model (HMM) based keyword-filler model even
with less computations.
|
Simple, Robust and Optimal Ranking from Pairwise Comparisons | We consider data in the form of pairwise comparisons of n items, with the
goal of precisely identifying the top k items for some value of k < n, or
alternatively, recovering a ranking of all the items. We analyze the Copeland
counting algorithm that ranks the items in order of the number of pairwise
comparisons won, and show it has three attractive features: (a) its
computational efficiency leads to speed-ups of several orders of magnitude in
computation time as compared to prior work; (b) it is robust in that
theoretical guarantees impose no conditions on the underlying matrix of
pairwise-comparison probabilities, in contrast to some prior work that applies
only to the BTL parametric model; and (c) it is an optimal method up to
constant factors, meaning that it achieves the information-theoretic limits for
recovering the top k-subset. We extend our results to obtain sharp guarantees
for approximate recovery under the Hamming distortion metric, and more
generally, to any arbitrary error requirement that satisfies a simple and
natural monotonicity condition.
|
Detection in the stochastic block model with multiple clusters: proof of
the achievability conjectures, acyclic BP, and the information-computation
gap | In a paper that initiated the modern study of the stochastic block model,
Decelle et al., backed by Mossel et al., made the following conjecture: Denote
by $k$ the number of balanced communities, $a/n$ the probability of connecting
inside communities and $b/n$ across, and set
$\mathrm{SNR}=(a-b)^2/(k(a+(k-1)b)$; for any $k \geq 2$, it is possible to
detect communities efficiently whenever $\mathrm{SNR}>1$ (the KS threshold),
whereas for $k\geq 4$, it is possible to detect communities
information-theoretically for some $\mathrm{SNR}<1$. Massouli\'e, Mossel et
al.\ and Bordenave et al.\ succeeded in proving that the KS threshold is
efficiently achievable for $k=2$, while Mossel et al.\ proved that it cannot be
crossed information-theoretically for $k=2$. The above conjecture remained open
for $k \geq 3$.
This paper proves this conjecture, further extending the efficient detection
to non-symmetrical SBMs with a generalized notion of detection and KS
threshold. For the efficient part, a linearized acyclic belief propagation
(ABP) algorithm is developed and proved to detect communities for any $k$ down
to the KS threshold in time $O(n \log n)$. Achieving this requires showing
optimality of ABP in the presence of cycles, a challenge for message passing
algorithms. The paper further connects ABP to a power iteration method with a
nonbacktracking operator of generalized order, formalizing the interplay
between message passing and spectral methods. For the information-theoretic
(IT) part, a non-efficient algorithm sampling a typical clustering is shown to
break down the KS threshold at $k=4$. The emerging gap is shown to be large in
some cases; if $a=0$, the KS threshold reads $b \gtrsim k^2$ whereas the IT
bound reads $b \gtrsim k \ln(k)$, making the SBM a good study-case for
information-computation gaps.
|
Low rank approximation and decomposition of large matrices using error
correcting codes | Low rank approximation is an important tool used in many applications of
signal processing and machine learning. Recently, randomized sketching
algorithms were proposed to effectively construct low rank approximations and
obtain approximate singular value decompositions of large matrices. Similar
ideas were used to solve least squares regression problems. In this paper, we
show how matrices from error correcting codes can be used to find such low rank
approximations and matrix decompositions, and extend the framework to linear
least squares regression problems. The benefits of using these code matrices
are the following: (i) They are easy to generate and they reduce randomness
significantly. (ii) Code matrices with mild properties satisfy the subspace
embedding property, and have a better chance of preserving the geometry of an
entire subspace of vectors. (iii) For parallel and distributed applications,
code matrices have significant advantages over structured random matrices and
Gaussian random matrices. (iv) Unlike Fourier or Hadamard transform matrices,
which require sampling $O(k\log k)$ columns for a rank-$k$ approximation, the
log factor is not necessary for certain types of code matrices. That is,
$(1+\epsilon)$ optimal Frobenius norm error can be achieved for a rank-$k$
approximation with $O(k/\epsilon)$ samples. (v) Fast multiplication is possible
with structured code matrices, so fast approximations can be achieved for
general dense input matrices. (vi) For least squares regression problem
$\min\|Ax-b\|_2$ where $A\in \mathbb{R}^{n\times d}$, the $(1+\epsilon)$
relative error approximation can be achieved with $O(d/\epsilon)$ samples, with
high probability, when certain code matrices are used.
|
Statistical Query Algorithms for Mean Vector Estimation and Stochastic
Convex Optimization | Stochastic convex optimization, where the objective is the expectation of a
random convex function, is an important and widely used method with numerous
applications in machine learning, statistics, operations research and other
areas. We study the complexity of stochastic convex optimization given only
statistical query (SQ) access to the objective function. We show that
well-known and popular first-order iterative methods can be implemented using
only statistical queries. For many cases of interest we derive nearly matching
upper and lower bounds on the estimation (sample) complexity including linear
optimization in the most general setting. We then present several consequences
for machine learning, differential privacy and proving concrete lower bounds on
the power of convex optimization based methods.
The key ingredient of our work is SQ algorithms and lower bounds for
estimating the mean vector of a distribution over vectors supported on a convex
body in $\mathbb{R}^d$. This natural problem has not been previously studied
and we show that our solutions can be used to get substantially improved SQ
versions of Perceptron and other online algorithms for learning halfspaces.
|
Personalized Course Sequence Recommendations | Given the variability in student learning it is becoming increasingly
important to tailor courses as well as course sequences to student needs. This
paper presents a systematic methodology for offering personalized course
sequence recommendations to students. First, a forward-search
backward-induction algorithm is developed that can optimally select course
sequences to decrease the time required for a student to graduate. The
algorithm accounts for prerequisite requirements (typically present in higher
level education) and course availability. Second, using the tools of
multi-armed bandits, an algorithm is developed that can optimally recommend a
course sequence that both reduces the time to graduate while also increasing
the overall GPA of the student. The algorithm dynamically learns how students
with different contextual backgrounds perform for given course sequences and
then recommends an optimal course sequence for new students. Using real-world
student data from the UCLA Mechanical and Aerospace Engineering department, we
illustrate how the proposed algorithms outperform other methods that do not
include student contextual information when making course sequence
recommendations.
|
Bayes-Optimal Effort Allocation in Crowdsourcing: Bounds and Index
Policies | We consider effort allocation in crowdsourcing, where we wish to assign
labeling tasks to imperfect homogeneous crowd workers to maximize overall
accuracy in a continuous-time Bayesian setting, subject to budget and time
constraints. The Bayes-optimal policy for this problem is the solution to a
partially observable Markov decision process, but the curse of dimensionality
renders the computation infeasible. Based on the Lagrangian Relaxation
technique in Adelman & Mersereau (2008), we provide a computationally tractable
instance-specific upper bound on the value of this Bayes-optimal policy, which
can in turn be used to bound the optimality gap of any other sub-optimal
policy. In an approach similar in spirit to the Whittle index for restless
multiarmed bandits, we provide an index policy for effort allocation in
crowdsourcing and demonstrate numerically that it outperforms other stateof-
arts and performs close to optimal solution.
|
Denoising and Completion of 3D Data via Multidimensional Dictionary
Learning | In this paper a new dictionary learning algorithm for multidimensional data
is proposed. Unlike most conventional dictionary learning methods which are
derived for dealing with vectors or matrices, our algorithm, named KTSVD,
learns a multidimensional dictionary directly via a novel algebraic approach
for tensor factorization as proposed in [3, 12, 13]. Using this approach one
can define a tensor-SVD and we propose to extend K-SVD algorithm used for 1-D
data to a K-TSVD algorithm for handling 2-D and 3-D data. Our algorithm, based
on the idea of sparse coding (using group-sparsity over multidimensional
coefficient vectors), alternates between estimating a compact representation
and dictionary learning. We analyze our KTSVD algorithm and demonstrate its
result on video completion and multispectral image denoising.
|
Strategies and Principles of Distributed Machine Learning on Big Data | The rise of Big Data has led to new demands for Machine Learning (ML) systems
to learn complex models with millions to billions of parameters, that promise
adequate capacity to digest massive datasets and offer powerful predictive
analytics thereupon. In order to run ML algorithms at such scales, on a
distributed cluster with 10s to 1000s of machines, it is often the case that
significant engineering efforts are required --- and one might fairly ask if
such engineering truly falls within the domain of ML research or not. Taking
the view that Big ML systems can benefit greatly from ML-rooted statistical and
algorithmic insights --- and that ML researchers should therefore not shy away
from such systems design --- we discuss a series of principles and strategies
distilled from our recent efforts on industrial-scale ML solutions. These
principles and strategies span a continuum from application, to engineering,
and to theoretical research and development of Big ML systems and
architectures, with the goal of understanding how to make them efficient,
generally-applicable, and supported with convergence and scaling guarantees.
They concern four key questions which traditionally receive little attention in
ML research: How to distribute an ML program over a cluster? How to bridge ML
computation with inter-machine communication? How to perform such
communication? What should be communicated between machines? By exposing
underlying statistical and algorithmic characteristics unique to ML programs
but not typically seen in traditional computer programs, and by dissecting
successful cases to reveal how we have harnessed these principles to design and
develop both high-performance distributed ML software as well as
general-purpose ML frameworks, we present opportunities for ML researchers and
practitioners to further shape and grow the area that lies between ML and
systems.
|
Autoencoding beyond pixels using a learned similarity metric | We present an autoencoder that leverages learned representations to better
measure similarities in data space. By combining a variational autoencoder with
a generative adversarial network we can use learned feature representations in
the GAN discriminator as basis for the VAE reconstruction objective. Thereby,
we replace element-wise errors with feature-wise errors to better capture the
data distribution while offering invariance towards e.g. translation. We apply
our method to images of faces and show that it outperforms VAEs with
element-wise similarity measures in terms of visual fidelity. Moreover, we show
that the method learns an embedding in which high-level abstract visual
features (e.g. wearing glasses) can be modified using simple arithmetic.
|
Distributed Bayesian Learning with Stochastic Natural-gradient
Expectation Propagation and the Posterior Server | This paper makes two contributions to Bayesian machine learning algorithms.
Firstly, we propose stochastic natural gradient expectation propagation (SNEP),
a novel alternative to expectation propagation (EP), a popular variational
inference algorithm. SNEP is a black box variational algorithm, in that it does
not require any simplifying assumptions on the distribution of interest, beyond
the existence of some Monte Carlo sampler for estimating the moments of the EP
tilted distributions. Further, as opposed to EP which has no guarantee of
convergence, SNEP can be shown to be convergent, even when using Monte Carlo
moment estimates. Secondly, we propose a novel architecture for distributed
Bayesian learning which we call the posterior server. The posterior server
allows scalable and robust Bayesian learning in cases where a data set is
stored in a distributed manner across a cluster, with each compute node
containing a disjoint subset of data. An independent Monte Carlo sampler is run
on each compute node, with direct access only to the local data subset, but
which targets an approximation to the global posterior distribution given all
data across the whole cluster. This is achieved by using a distributed
asynchronous implementation of SNEP to pass messages across the cluster. We
demonstrate SNEP and the posterior server on distributed Bayesian learning of
logistic regression and neural networks.
Keywords: Distributed Learning, Large Scale Learning, Deep Learning, Bayesian
Learn- ing, Variational Inference, Expectation Propagation, Stochastic
Approximation, Natural Gradient, Markov chain Monte Carlo, Parameter Server,
Posterior Server.
|
Homology Computation of Large Point Clouds using Quantum Annealing | Homology is a tool in topological data analysis which measures the shape of
the data. In many cases, these measurements translate into new insights which
are not available by other means. To compute homology, we rely on mathematical
constructions which scale exponentially with the size of the data. Therefore,
for large point clouds, the computation is infeasible using classical
computers. In this paper, we present a quantum annealing pipeline for
computation of homology of large point clouds. The pipeline takes as input a
graph approximating the given point cloud. It uses quantum annealing to compute
a clique covering of the graph and then uses this cover to construct a
Mayer-Vietoris complex. The pipeline terminates by performing a simplified
homology computation of the Mayer-Vietoris complex. We have introduced three
different clique coverings and their quantum annealing formulation. Our
pipeline scales polynomially in the size of the data, once the covering step is
solved. To prove correctness of our algorithm, we have also included tests
using D-Wave 2X quantum processor.
|
Selecting Near-Optimal Learners via Incremental Data Allocation | We study a novel machine learning (ML) problem setting of sequentially
allocating small subsets of training data amongst a large set of classifiers.
The goal is to select a classifier that will give near-optimal accuracy when
trained on all data, while also minimizing the cost of misallocated samples.
This is motivated by large modern datasets and ML toolkits with many
combinations of learning algorithms and hyper-parameters. Inspired by the
principle of "optimism under uncertainty," we propose an innovative strategy,
Data Allocation using Upper Bounds (DAUB), which robustly achieves these
objectives across a variety of real-world datasets.
We further develop substantial theoretical support for DAUB in an idealized
setting where the expected accuracy of a classifier trained on $n$ samples can
be known exactly. Under these conditions we establish a rigorous sub-linear
bound on the regret of the approach (in terms of misallocated data), as well as
a rigorous bound on suboptimality of the selected classifier. Our accuracy
estimates using real-world datasets only entail mild violations of the
theoretical scenario, suggesting that the practical behavior of DAUB is likely
to approach the idealized behavior.
|
Write a Classifier: Predicting Visual Classifiers from Unstructured Text | People typically learn through exposure to visual concepts associated with
linguistic descriptions. For instance, teaching visual object categories to
children is often accompanied by descriptions in text or speech. In a machine
learning context, these observations motivates us to ask whether this learning
process could be computationally modeled to learn visual classifiers. More
specifically, the main question of this work is how to utilize purely textual
description of visual classes with no training images, to learn explicit visual
classifiers for them. We propose and investigate two baseline formulations,
based on regression and domain transfer, that predict a linear classifier.
Then, we propose a new constrained optimization formulation that combines a
regression function and a knowledge transfer function with additional
constraints to predict the parameters of a linear classifier. We also propose a
generic kernelized models where a kernel classifier is predicted in the form
defined by the representer theorem. The kernelized models allow defining and
utilizing any two RKHS (Reproducing Kernel Hilbert Space) kernel functions in
the visual space and text space, respectively. We finally propose a kernel
function between unstructured text descriptions that builds on distributional
semantics, which shows an advantage in our setting and could be useful for
other applications. We applied all the studied models to predict visual
classifiers on two fine-grained and challenging categorization datasets (CU
Birds and Flower Datasets), and the results indicate successful predictions of
our final model over several baselines that we designed.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.