title
stringlengths 7
246
| abstract
stringlengths 6
3.31k
|
|---|---|
Science Question Answering using Instructional Materials | We provide a solution for elementary science test using instructional
materials. We posit that there is a hidden structure that explains the
correctness of an answer given the question and instructional materials and
present a unified max-margin framework that learns to find these hidden
structures (given a corpus of question-answer pairs and instructional
materials), and uses what it learns to answer novel elementary science
questions. Our evaluation shows that our framework outperforms several strong
baselines.
|
Joint Dimensionality Reduction for Two Feature Vectors | Many machine learning problems, especially multi-modal learning problems,
have two sets of distinct features (e.g., image and text features in news story
classification, or neuroimaging data and neurocognitive data in cognitive
science research). This paper addresses the joint dimensionality reduction of
two feature vectors in supervised learning problems. In particular, we assume a
discriminative model where low-dimensional linear embeddings of the two feature
vectors are sufficient statistics for predicting a dependent variable. We show
that a simple algorithm involving singular value decomposition can accurately
estimate the embeddings provided that certain sample complexities are
satisfied, without specifying the nonlinear link function (regressor or
classifier). The main results establish sample complexities under multiple
settings. Sample complexities for different link functions only differ by
constant factors.
|
Convex Optimization For Non-Convex Problems via Column Generation | We apply column generation to approximating complex structured objects via a
set of primitive structured objects under either the cross entropy or L2 loss.
We use L1 regularization to encourage the use of few structured primitive
objects. We attack approximation using convex optimization over an infinite
number of variables each corresponding to a primitive structured object that
are generated on demand by easy inference in the Lagrangian dual. We apply our
approach to producing low rank approximations to large 3-way tensors.
|
Identifiability Assumptions and Algorithm for Directed Graphical Models
with Feedback | Directed graphical models provide a useful framework for modeling causal or
directional relationships for multivariate data. Prior work has largely focused
on identifiability and search algorithms for directed acyclic graphical (DAG)
models. In many applications, feedback naturally arises and directed graphical
models that permit cycles occur. In this paper we address the issue of
identifiability for general directed cyclic graphical (DCG) models satisfying
the Markov assumption. In particular, in addition to the faithfulness
assumption which has already been introduced for cyclic models, we introduce
two new identifiability assumptions, one based on selecting the model with the
fewest edges and the other based on selecting the DCG model that entails the
maximum number of d-separation rules. We provide theoretical results comparing
these assumptions which show that: (1) selecting models with the largest number
of d-separation rules is strictly weaker than the faithfulness assumption; (2)
unlike for DAG models, selecting models with the fewest edges does not
necessarily result in a milder assumption than the faithfulness assumption. We
also provide connections between our two new principles and minimality
assumptions. We use our identifiability assumptions to develop search
algorithms for small-scale DCG models. Our simulation study supports our
theoretical results, showing that the algorithms based on our two new
principles generally out-perform algorithms based on the faithfulness
assumption in terms of selecting the true skeleton for DCG models.
|
Unsupervised Domain Adaptation with Residual Transfer Networks | The recent success of deep neural networks relies on massive amounts of
labeled data. For a target task where labeled data is unavailable, domain
adaptation can transfer a learner from a different source domain. In this
paper, we propose a new approach to domain adaptation in deep networks that can
jointly learn adaptive classifiers and transferable features from labeled data
in the source domain and unlabeled data in the target domain. We relax a
shared-classifier assumption made by previous methods and assume that the
source classifier and target classifier differ by a residual function. We
enable classifier adaptation by plugging several layers into deep network to
explicitly learn the residual function with reference to the target classifier.
We fuse features of multiple layers with tensor product and embed them into
reproducing kernel Hilbert spaces to match distributions for feature
adaptation. The adaptation can be achieved in most feed-forward models by
extending them with new residual layers and loss functions, which can be
trained efficiently via back-propagation. Empirical evidence shows that the new
approach outperforms state of the art methods on standard domain adaptation
benchmarks.
|
Frequency Analysis of Temporal Graph Signals | This letter extends the concept of graph-frequency to graph signals that
evolve with time. Our goal is to generalize and, in fact, unify the familiar
concepts from time- and graph-frequency analysis. To this end, we study a joint
temporal and graph Fourier transform (JFT) and demonstrate its attractive
properties. We build on our results to create filters which act on the joint
(temporal and graph) frequency domain, and show how these can be used to
perform interference cancellation. The proposed algorithms are distributed,
have linear complexity, and can approximate any desired joint filtering
objective.
|
Random Forest Based Approach for Concept Drift Handling | Concept drift has potential in smart grid analysis because the socio-economic
behaviour of consumers is not governed by the laws of physics. Likewise there
are also applications in wind power forecasting. In this paper we present
decision tree ensemble classification method based on the Random Forest
algorithm for concept drift. The weighted majority voting ensemble aggregation
rule is employed based on the ideas of Accuracy Weighted Ensemble (AWE) method.
Base learner weight in our case is computed for each sample evaluation using
base learners accuracy and intrinsic proximity measure of Random Forest. Our
algorithm exploits both temporal weighting of samples and ensemble pruning as a
forgetting strategy. We present results of empirical comparison of our method
with original random forest with incorporated "replace-the-looser" forgetting
andother state-of-the-art concept-drfit classifiers like AWE2.
|
Autoregressive Moving Average Graph Filtering | One of the cornerstones of the field of signal processing on graphs are graph
filters, direct analogues of classical filters, but intended for signals
defined on graphs. This work brings forth new insights on the distributed graph
filtering problem. We design a family of autoregressive moving average (ARMA)
recursions, which (i) are able to approximate any desired graph frequency
response, and (ii) give exact solutions for tasks such as graph signal
denoising and interpolation. The design philosophy, which allows us to design
the ARMA coefficients independently from the underlying graph, renders the ARMA
graph filters suitable in static and, particularly, time-varying settings. The
latter occur when the graph signal and/or graph are changing over time. We show
that in case of a time-varying graph signal our approach extends naturally to a
two-dimensional filter, operating concurrently in the graph and regular time
domains. We also derive sufficient conditions for filter stability when the
graph and signal are time-varying. The analytical and numerical results
presented in this paper illustrate that ARMA graph filters are practically
appealing for static and time-varying settings, as predicted by theoretical
derivations.
|
Bayesian Optimization with Safety Constraints: Safe and Automatic
Parameter Tuning in Robotics | Robotic algorithms typically depend on various parameters, the choice of
which significantly affects the robot's performance. While an initial guess for
the parameters may be obtained from dynamic models of the robot, parameters are
usually tuned manually on the real system to achieve the best performance.
Optimization algorithms, such as Bayesian optimization, have been used to
automate this process. However, these methods may evaluate unsafe parameters
during the optimization process that lead to safety-critical system failures.
Recently, a safe Bayesian optimization algorithm, called SafeOpt, has been
developed, which guarantees that the performance of the system never falls
below a critical value; that is, safety is defined based on the performance
function. However, coupling performance and safety is often not desirable in
robotics. For example, high-gain controllers might achieve low average tracking
error (performance), but can overshoot and violate input constraints. In this
paper, we present a generalized algorithm that allows for multiple safety
constraints separate from the objective. Given an initial set of safe
parameters, the algorithm maximizes performance but only evaluates parameters
that satisfy safety for all constraints with high probability. To this end, it
carefully explores the parameter space by exploiting regularity assumptions in
terms of a Gaussian process prior. Moreover, we show how context variables can
be used to safely transfer knowledge to new situations and tasks. We provide a
theoretical analysis and demonstrate that the proposed algorithm enables fast,
automatic, and safe optimization of tuning parameters in experiments on a
quadrotor vehicle.
|
Generalization Properties of Learning with Random Features | We study the generalization properties of ridge regression with random
features in the statistical learning framework. We show for the first time that
$O(1/\sqrt{n})$ learning bounds can be achieved with only $O(\sqrt{n}\log n)$
random features rather than $O({n})$ as suggested by previous results. Further,
we prove faster learning rates and show that they might require more random
features, unless they are sampled according to a possibly problem dependent
distribution. Our results shed light on the statistical computational
trade-offs in large scale kernelized learning, showing the potential
effectiveness of random features in reducing the computational complexity while
keeping optimal generalization properties.
|
Surprising properties of dropout in deep networks | We analyze dropout in deep networks with rectified linear units and the
quadratic loss. Our results expose surprising differences between the behavior
of dropout and more traditional regularizers like weight decay. For example, on
some simple data sets dropout training produces negative weights even though
the output is the sum of the inputs. This provides a counterpoint to the
suggestion that dropout discourages co-adaptation of weights. We also show that
the dropout penalty can grow exponentially in the depth of the network while
the weight-decay penalty remains essentially linear, and that dropout is
insensitive to various re-scalings of the input features, outputs, and network
weights. This last insensitivity implies that there are no isolated local
minima of the dropout training criterion. Our work uncovers new properties of
dropout, extends our understanding of why dropout succeeds, and lays the
foundation for further progress.
|
Benefits of depth in neural networks | For any positive integer $k$, there exist neural networks with $\Theta(k^3)$
layers, $\Theta(1)$ nodes per layer, and $\Theta(1)$ distinct parameters which
can not be approximated by networks with $\mathcal{O}(k)$ layers unless they
are exponentially large --- they must possess $\Omega(2^k)$ nodes. This result
is proved here for a class of nodes termed "semi-algebraic gates" which
includes the common choices of ReLU, maximum, indicator, and piecewise
polynomial functions, therefore establishing benefits of depth against not just
standard networks with ReLU gates, but also convolutional networks with ReLU
and maximization gates, sum-product networks, and boosted decision trees (in
this last case with a stronger separation: $\Omega(2^{k^3})$ total tree nodes
are required).
|
Convolutional Tables Ensemble: classification in microseconds | We study classifiers operating under severe classification time constraints,
corresponding to 1-1000 CPU microseconds, using Convolutional Tables Ensemble
(CTE), an inherently fast architecture for object category recognition. The
architecture is based on convolutionally-applied sparse feature extraction,
using trees or ferns, and a linear voting layer. Several structure and
optimization variants are considered, including novel decision functions, tree
learning algorithm, and distillation from CNN to CTE architecture. Accuracy
improvements of 24-45% over related art of similar speed are demonstrated on
standard object recognition benchmarks. Using Pareto speed-accuracy curves, we
show that CTE can provide better accuracy than Convolutional Neural Networks
(CNN) for a certain range of classification time constraints, or alternatively
provide similar error rates with 5-200X speedup.
|
Learning Granger Causality for Hawkes Processes | Learning Granger causality for general point processes is a very challenging
task. In this paper, we propose an effective method, learning Granger
causality, for a special but significant type of point processes --- Hawkes
process. We reveal the relationship between Hawkes process's impact function
and its Granger causality graph. Specifically, our model represents impact
functions using a series of basis functions and recovers the Granger causality
graph via group sparsity of the impact functions' coefficients. We propose an
effective learning algorithm combining a maximum likelihood estimator (MLE)
with a sparse-group-lasso (SGL) regularizer. Additionally, the flexibility of
our model allows to incorporate the clustering structure event types into
learning framework. We analyze our learning algorithm and propose an adaptive
procedure to select basis functions. Experiments on both synthetic and
real-world data show that our method can learn the Granger causality graph and
the triggering patterns of the Hawkes processes simultaneously.
|
Adversarial Top-$K$ Ranking | We study the top-$K$ ranking problem where the goal is to recover the set of
top-$K$ ranked items out of a large collection of items based on partially
revealed preferences. We consider an adversarial crowdsourced setting where
there are two population sets, and pairwise comparison samples drawn from one
of the populations follow the standard Bradley-Terry-Luce model (i.e., the
chance of item $i$ beating item $j$ is proportional to the relative score of
item $i$ to item $j$), while in the other population, the corresponding chance
is inversely proportional to the relative score. When the relative size of the
two populations is known, we characterize the minimax limit on the sample size
required (up to a constant) for reliably identifying the top-$K$ items, and
demonstrate how it scales with the relative size. Moreover, by leveraging a
tensor decomposition method for disambiguating mixture distributions, we extend
our result to the more realistic scenario in which the relative population size
is unknown, thus establishing an upper bound on the fundamental limit of the
sample size for recovering the top-$K$ set.
|
Personalized Expertise Search at LinkedIn | LinkedIn is the largest professional network with more than 350 million
members. As the member base increases, searching for experts becomes more and
more challenging. In this paper, we propose an approach to address the problem
of personalized expertise search on LinkedIn, particularly for exploratory
search queries containing {\it skills}. In the offline phase, we introduce a
collaborative filtering approach based on matrix factorization. Our approach
estimates expertise scores for both the skills that members list on their
profiles as well as the skills they are likely to have but do not explicitly
list. In the online phase (at query time) we use expertise scores on these
skills as a feature in combination with other features to rank the results. To
learn the personalized ranking function, we propose a heuristic to extract
training data from search logs while handling position and sample selection
biases. We tested our models on two products - LinkedIn homepage and LinkedIn
recruiter. A/B tests showed significant improvements in click through rates -
31% for CTR@1 for recruiter (18% for homepage) as well as downstream messages
sent from search - 37% for recruiter (20% for homepage). As of writing this
paper, these models serve nearly all live traffic for skills search on LinkedIn
homepage as well as LinkedIn recruiter.
|
Secure Approximation Guarantee for Cryptographically Private Empirical
Risk Minimization | Privacy concern has been increasingly important in many machine learning (ML)
problems. We study empirical risk minimization (ERM) problems under secure
multi-party computation (MPC) frameworks. Main technical tools for MPC have
been developed based on cryptography. One of limitations in current
cryptographically private ML is that it is computationally intractable to
evaluate non-linear functions such as logarithmic functions or exponential
functions. Therefore, for a class of ERM problems such as logistic regression
in which non-linear function evaluations are required, one can only obtain
approximate solutions. In this paper, we introduce a novel cryptographically
private tool called secure approximation guarantee (SAG) method. The key
property of SAG method is that, given an arbitrary approximate solution, it can
provide a non-probabilistic assumption-free bound on the approximation quality
under cryptographically secure computation framework. We demonstrate the
benefit of the SAG method by applying it to several problems including a
practical privacy-preserving data analysis task on genomic and clinical
information.
|
Optimal Best Arm Identification with Fixed Confidence | We give a complete characterization of the complexity of best-arm
identification in one-parameter bandit problems. We prove a new, tight lower
bound on the sample complexity. We propose the `Track-and-Stop' strategy, which
we prove to be asymptotically optimal. It consists in a new sampling rule
(which tracks the optimal proportions of arm draws highlighted by the lower
bound) and in a stopping rule named after Chernoff, for which we give a new
analysis.
|
Distributed Information-Theoretic Clustering | We study a novel multi-terminal source coding setup motivated by the
biclustering problem. Two separate encoders observe two i.i.d. sequences $X^n$
and $Y^n$, respectively. The goal is to find rate-limited encodings $f(x^n)$
and $g(z^n)$ that maximize the mutual information $I(f(X^n); g(Y^n))/n$. We
discuss connections of this problem with hypothesis testing against
independence, pattern recognition, and the information bottleneck method.
Improving previous cardinality bounds for the inner and outer bounds allows us
to thoroughly study the special case of a binary symmetric source and to
quantify the gap between the inner and the outer bound in this special case.
Furthermore, we investigate a multiple description (MD) extension of the Chief
Operating Officer (CEO) problem with mutual information constraint.
Surprisingly, this MD-CEO problem permits a tight single-letter
characterization of the achievable region.
|
Deep Exploration via Bootstrapped DQN | Efficient exploration in complex environments remains a major challenge for
reinforcement learning. We propose bootstrapped DQN, a simple algorithm that
explores in a computationally and statistically efficient manner through use of
randomized value functions. Unlike dithering strategies such as epsilon-greedy
exploration, bootstrapped DQN carries out temporally-extended (or deep)
exploration; this can lead to exponentially faster learning. We demonstrate
these benefits in complex stochastic MDPs and in the large-scale Arcade
Learning Environment. Bootstrapped DQN substantially improves learning times
and performance across most Atari games.
|
Efficient Representation of Low-Dimensional Manifolds using Deep
Networks | We consider the ability of deep neural networks to represent data that lies
near a low-dimensional manifold in a high-dimensional space. We show that deep
networks can efficiently extract the intrinsic, low-dimensional coordinates of
such data. We first show that the first two layers of a deep network can
exactly embed points lying on a monotonic chain, a special type of piecewise
linear manifold, mapping them to a low-dimensional Euclidean space. Remarkably,
the network can do this using an almost optimal number of parameters. We also
show that this network projects nearby points onto the manifold and then embeds
them with little error. We then extend these results to more general manifolds.
|
Delay and Cooperation in Nonstochastic Bandits | We study networks of communicating learning agents that cooperate to solve a
common nonstochastic bandit problem. Agents use an underlying communication
network to get messages about actions selected by other agents, and drop
messages that took more than $d$ hops to arrive, where $d$ is a delay
parameter. We introduce \textsc{Exp3-Coop}, a cooperative version of the {\sc
Exp3} algorithm and prove that with $K$ actions and $N$ agents the average
per-agent regret after $T$ rounds is at most of order $\sqrt{\bigl(d+1 +
\tfrac{K}{N}\alpha_{\le d}\bigr)(T\ln K)}$, where $\alpha_{\le d}$ is the
independence number of the $d$-th power of the connected communication graph
$G$. We then show that for any connected graph, for $d=\sqrt{K}$ the regret
bound is $K^{1/4}\sqrt{T}$, strictly better than the minimax regret $\sqrt{KT}$
for noncooperating agents. More informed choices of $d$ lead to bounds which
are arbitrarily close to the full information minimax regret $\sqrt{T\ln K}$
when $G$ is dense. When $G$ has sparse components, we show that a variant of
\textsc{Exp3-Coop}, allowing agents to choose their parameters according to
their centrality in $G$, strictly improves the regret. Finally, as a by-product
of our analysis, we provide the first characterization of the minimax regret
for bandit learning with delay.
|
Quantum Perceptron Models | We demonstrate how quantum computation can provide non-trivial improvements
in the computational and statistical complexity of the perceptron model. We
develop two quantum algorithms for perceptron learning. The first algorithm
exploits quantum information processing to determine a separating hyperplane
using a number of steps sublinear in the number of data points $N$, namely
$O(\sqrt{N})$. The second algorithm illustrates how the classical mistake bound
of $O(\frac{1}{\gamma^2})$ can be further improved to
$O(\frac{1}{\sqrt{\gamma}})$ through quantum means, where $\gamma$ denotes the
margin. Such improvements are achieved through the application of quantum
amplitude amplification to the version space interpretation of the perceptron
model.
|
DR-ABC: Approximate Bayesian Computation with Kernel-Based Distribution
Regression | Performing exact posterior inference in complex generative models is often
difficult or impossible due to an expensive to evaluate or intractable
likelihood function. Approximate Bayesian computation (ABC) is an inference
framework that constructs an approximation to the true likelihood based on the
similarity between the observed and simulated data as measured by a predefined
set of summary statistics. Although the choice of appropriate problem-specific
summary statistics crucially influences the quality of the likelihood
approximation and hence also the quality of the posterior sample in ABC, there
are only few principled general-purpose approaches to the selection or
construction of such summary statistics. In this paper, we develop a novel
framework for this task using kernel-based distribution regression. We model
the functional relationship between data distributions and the optimal choice
(with respect to a loss function) of summary statistics using kernel-based
distribution regression. We show that our approach can be implemented in a
computationally and statistically efficient way using the random Fourier
features framework for large-scale kernel learning. In addition to that, our
framework shows superior performance when compared to related methods on toy
and real-world problems.
|
Black-box optimization with a politician | We propose a new framework for black-box convex optimization which is
well-suited for situations where gradient computations are expensive. We derive
a new method for this framework which leverages several concepts from convex
optimization, from standard first-order methods (e.g. gradient descent or
quasi-Newton methods) to analytical centers (i.e. minimizers of self-concordant
barriers). We demonstrate empirically that our new technique compares favorably
with state of the art algorithms (such as BFGS).
|
Bi-directional LSTM Recurrent Neural Network for Chinese Word
Segmentation | Recurrent neural network(RNN) has been broadly applied to natural language
processing(NLP) problems. This kind of neural network is designed for modeling
sequential data and has been testified to be quite efficient in sequential
tagging tasks. In this paper, we propose to use bi-directional RNN with long
short-term memory(LSTM) units for Chinese word segmentation, which is a crucial
preprocess task for modeling Chinese sentences and articles. Classical methods
focus on designing and combining hand-craft features from context, whereas
bi-directional LSTM network(BLSTM) does not need any prior knowledge or
pre-designing, and it is expert in keeping the contextual information in both
directions. Experiment result shows that our approach gets state-of-the-art
performance in word segmentation on both traditional Chinese datasets and
simplified Chinese datasets.
|
Unsupervised Domain Adaptation Using Approximate Label Matching | Domain adaptation addresses the problem created when training data is
generated by a so-called source distribution, but test data is generated by a
significantly different target distribution. In this work, we present
approximate label matching (ALM), a new unsupervised domain adaptation
technique that creates and leverages a rough labeling on the test samples, then
uses these noisy labels to learn a transformation that aligns the source and
target samples. We show that the transformation estimated by ALM has favorable
properties compared to transformations estimated by other methods, which do not
use any kind of target labeling. Our model is regularized by requiring that a
classifier trained to discriminate source from transformed target samples
cannot distinguish between the two. We experiment with ALM on simulated and
real data, and show that it outperforms techniques commonly used in the field.
|
Segmentation Rectification for Video Cutout via One-Class Structured
Learning | Recent works on interactive video object cutout mainly focus on designing
dynamic foreground-background (FB) classifiers for segmentation propagation.
However, the research on optimally removing errors from the FB classification
is sparse, and the errors often accumulate rapidly, causing significant errors
in the propagated frames. In this work, we take the initial steps to addressing
this problem, and we call this new task \emph{segmentation rectification}. Our
key observation is that the possibly asymmetrically distributed false positive
and false negative errors were handled equally in the conventional methods. We,
alternatively, propose to optimally remove these two types of errors. To this
effect, we propose a novel bilayer Markov Random Field (MRF) model for this new
task. We also adopt the well-established structured learning framework to learn
the optimal model from data. Additionally, we propose a novel one-class
structured SVM (OSSVM) which greatly speeds up the structured learning process.
Our method naturally extends to RGB-D videos as well. Comprehensive experiments
on both RGB and RGB-D data demonstrate that our simple and effective method
significantly outperforms the segmentation propagation methods adopted in the
state-of-the-art video cutout systems, and the results also suggest the
potential usefulness of our method in image cutout system.
|
Gradient Descent Converges to Minimizers | We show that gradient descent converges to a local minimizer, almost surely
with random initialization. This is proved by applying the Stable Manifold
Theorem from dynamical systems theory.
|
Personalized Federated Search at LinkedIn | LinkedIn has grown to become a platform hosting diverse sources of
information ranging from member profiles, jobs, professional groups, slideshows
etc. Given the existence of multiple sources, when a member issues a query like
"software engineer", the member could look for software engineer profiles, jobs
or professional groups. To tackle this problem, we exploit a data-driven
approach that extracts searcher intents from their profile data and recent
activities at a large scale. The intents such as job seeking, hiring, content
consuming are used to construct features to personalize federated search
experience. We tested the approach on the LinkedIn homepage and A/B tests show
significant improvements in member engagement. As of writing this paper, the
approach powers all of federated search on LinkedIn homepage.
|
"Why Should I Trust You?": Explaining the Predictions of Any Classifier | Despite widespread adoption, machine learning models remain mostly black
boxes. Understanding the reasons behind predictions is, however, quite
important in assessing trust, which is fundamental if one plans to take action
based on a prediction, or when choosing whether to deploy a new model. Such
understanding also provides insights into the model, which can be used to
transform an untrustworthy model or prediction into a trustworthy one. In this
work, we propose LIME, a novel explanation technique that explains the
predictions of any classifier in an interpretable and faithful manner, by
learning an interpretable model locally around the prediction. We also propose
a method to explain models by presenting representative individual predictions
and their explanations in a non-redundant way, framing the task as a submodular
optimization problem. We demonstrate the flexibility of these methods by
explaining different models for text (e.g. random forests) and image
classification (e.g. neural networks). We show the utility of explanations via
novel experiments, both simulated and with human subjects, on various scenarios
that require trust: deciding if one should trust a prediction, choosing between
models, improving an untrustworthy classifier, and identifying why a classifier
should not be trusted.
|
Q($\lambda$) with Off-Policy Corrections | We propose and analyze an alternate approach to off-policy multi-step
temporal difference learning, in which off-policy returns are corrected with
the current Q-function in terms of rewards, rather than with the target policy
in terms of transition probabilities. We prove that such approximate
corrections are sufficient for off-policy convergence both in policy evaluation
and control, provided certain conditions. These conditions relate the distance
between the target and behavior policies, the eligibility trace parameter and
the discount factor, and formalize an underlying tradeoff in off-policy
TD($\lambda$). We illustrate this theoretical relationship empirically on a
continuous-state control task.
|
Stochastic Process Bandits: Upper Confidence Bounds Algorithms via
Generic Chaining | The paper considers the problem of global optimization in the setup of
stochastic process bandits. We introduce an UCB algorithm which builds a
cascade of discretization trees based on generic chaining in order to render
possible his operability over a continuous domain. The theoretical framework
applies to functions under weak probabilistic smoothness assumptions and also
extends significantly the spectrum of application of UCB strategies. Moreover
generic regret bounds are derived which are then specialized to Gaussian
processes indexed on infinite-dimensional spaces as well as to quadratic forms
of Gaussian processes. Lower bounds are also proved in the case of Gaussian
processes to assess the optimality of the proposed algorithm.
|
A Subsequence Interleaving Model for Sequential Pattern Mining | Recent sequential pattern mining methods have used the minimum description
length (MDL) principle to define an encoding scheme which describes an
algorithm for mining the most compressing patterns in a database. We present a
novel subsequence interleaving model based on a probabilistic model of the
sequence database, which allows us to search for the most compressing set of
patterns without designing a specific encoding scheme. Our proposed algorithm
is able to efficiently mine the most relevant sequential patterns and rank them
using an associated measure of interestingness. The efficient inference in our
model is a direct result of our use of a structural expectation-maximization
framework, in which the expectation-step takes the form of a submodular
optimization problem subject to a coverage constraint. We show on both
synthetic and real world datasets that our model mines a set of sequential
patterns with low spuriousness and redundancy, high interpretability and
usefulness in real-world applications. Furthermore, we demonstrate that the
quality of the patterns from our approach is comparable to, if not better than,
existing state of the art sequential pattern mining algorithms.
|
Generating images with recurrent adversarial networks | Gatys et al. (2015) showed that optimizing pixels to match features in a
convolutional network with respect reference image features is a way to render
images of high visual quality. We show that unrolling this gradient-based
optimization yields a recurrent computation that creates images by
incrementally adding onto a visual "canvas". We propose a recurrent generative
model inspired by this view, and show that it can be trained using adversarial
training to generate very good image samples. We also propose a way to
quantitatively compare adversarial networks by having the generators and
discriminators of these networks compete against each other.
|
Patient Flow Prediction via Discriminative Learning of
Mutually-Correcting Processes | Over the past decade the rate of care unit (CU) use in the United States has
been increasing. With an aging population and ever-growing demand for medical
care, effective management of patients' transitions among different care
facilities will prove indispensible for shortening the length of hospital
stays, improving patient outcomes, allocating critical care resources, and
reducing preventable re-admissions. In this paper, we focus on an important
problem of predicting the so-called "patient flow" from longitudinal electronic
health records (EHRs), which has not been explored via existing machine
learning techniques. By treating a sequence of transition events as a point
process, we develop a novel framework for modeling patient flow through various
CUs and jointly predicting patients' destination CUs and duration days. Instead
of learning a generative point process model via maximum likelihood estimation,
we propose a novel discriminative learning algorithm aiming at improving the
prediction of transition events in the case of sparse data. By parameterizing
the proposed model as a mutually-correcting process, we formulate the
estimation problem via generalized linear models, which lends itself to
efficient learning based on alternating direction method of multipliers (ADMM).
Furthermore, we achieve simultaneous feature selection and learning by adding a
group-lasso regularizer to the ADMM algorithm. Additionally, for suppressing
the negative influence of data imbalance on the learning of model, we
synthesize auxiliary training data for the classes with extremely few samples,
and improve the robustness of our learning method accordingly. Testing on
real-world data, we show that our method obtains superior performance in terms
of accuracy of predicting the destination CU transition and duration of each CU
occupancy.
|
Practical Introduction to Clustering Data | Data clustering is an approach to seek for structure in sets of complex data,
i.e., sets of "objects". The main objective is to identify groups of objects
which are similar to each other, e.g., for classification. Here, an
introduction to clustering is given and three basic approaches are introduced:
the k-means algorithm, neighbour-based clustering, and an agglomerative
clustering method. For all cases, C source code examples are given, allowing
for an easy implementation.
|
A Harmonic Extension Approach for Collaborative Ranking | We present a new perspective on graph-based methods for collaborative ranking
for recommender systems. Unlike user-based or item-based methods that compute a
weighted average of ratings given by the nearest neighbors, or low-rank
approximation methods using convex optimization and the nuclear norm, we
formulate matrix completion as a series of semi-supervised learning problems,
and propagate the known ratings to the missing ones on the user-user or
item-item graph globally. The semi-supervised learning problems are expressed
as Laplace-Beltrami equations on a manifold, or namely, harmonic extension, and
can be discretized by a point integral method. We show that our approach does
not impose a low-rank Euclidean subspace on the data points, but instead
minimizes the dimension of the underlying manifold. Our method, named LDM (low
dimensional manifold), turns out to be particularly effective in generating
rankings of items, showing decent computational efficiency and robust ranking
quality compared to state-of-the-art methods.
|
Fast Learning Requires Good Memory: A Time-Space Lower Bound for Parity
Learning | We prove that any algorithm for learning parities requires either a memory of
quadratic size or an exponential number of samples. This proves a recent
conjecture of Steinhardt, Valiant and Wager and shows that for some learning
problems a large storage space is crucial.
More formally, in the problem of parity learning, an unknown string $x \in
\{0,1\}^n$ was chosen uniformly at random. A learner tries to learn $x$ from a
stream of samples $(a_1, b_1), (a_2, b_2) \ldots$, where each~$a_t$ is
uniformly distributed over $\{0,1\}^n$ and $b_t$ is the inner product of $a_t$
and $x$, modulo~2. We show that any algorithm for parity learning, that uses
less than $\frac{n^2}{25}$ bits of memory, requires an exponential number of
samples.
Previously, there was no non-trivial lower bound on the number of samples
needed, for any learning problem, even if the allowed memory size is $O(n)$
(where $n$ is the space needed to store one sample).
We also give an application of our result in the field of bounded-storage
cryptography. We show an encryption scheme that requires a private key of
length $n$, as well as time complexity of $n$ per encryption/decription of each
bit, and is provenly and unconditionally secure as long as the attacker uses
less than $\frac{n^2}{25}$ memory bits and the scheme is used at most an
exponential number of times. Previous works on bounded-storage cryptography
assumed that the memory size used by the attacker is at most linear in the time
needed for encryption/decription.
|
Equilibrium Propagation: Bridging the Gap Between Energy-Based Models
and Backpropagation | We introduce Equilibrium Propagation, a learning framework for energy-based
models. It involves only one kind of neural computation, performed in both the
first phase (when the prediction is made) and the second phase of training
(after the target or prediction error is revealed). Although this algorithm
computes the gradient of an objective function just like Backpropagation, it
does not need a special computation or circuit for the second phase, where
errors are implicitly propagated. Equilibrium Propagation shares similarities
with Contrastive Hebbian Learning and Contrastive Divergence while solving the
theoretical issues of both algorithms: our algorithm computes the gradient of a
well defined objective function. Because the objective function is defined in
terms of local perturbations, the second phase of Equilibrium Propagation
corresponds to only nudging the prediction (fixed point, or stationary
distribution) towards a configuration that reduces prediction error. In the
case of a recurrent multi-layer supervised network, the output units are
slightly nudged towards their target in the second phase, and the perturbation
introduced at the output layer propagates backward in the hidden layers. We
show that the signal 'back-propagated' during this second phase corresponds to
the propagation of error derivatives and encodes the gradient of the objective
function, when the synaptic update corresponds to a standard form of
spike-timing dependent plasticity. This work makes it more plausible that a
mechanism similar to Backpropagation could be implemented by brains, since
leaky integrator neural computation performs both inference and error
back-propagation in our model. The only local difference between the two phases
is whether synaptic changes are allowed or not.
|
Primal-Dual Rates and Certificates | We propose an algorithm-independent framework to equip existing optimization
methods with primal-dual certificates. Such certificates and corresponding rate
of convergence guarantees are important for practitioners to diagnose progress,
in particular in machine learning applications. We obtain new primal-dual
convergence rates, e.g., for the Lasso as well as many L1, Elastic Net, group
Lasso and TV-regularized problems. The theory applies to any norm-regularized
generalized linear model. Our approach provides efficiently computable duality
gaps which are globally defined, without modifying the original problems in the
region of interest.
|
Monte Carlo Markov Chain Algorithms for Sampling Strongly Rayleigh
Distributions and Determinantal Point Processes | Strongly Rayleigh distributions are natural generalizations of product and
determinantal probability distributions and satisfy strongest form of negative
dependence properties. We show that the "natural" Monte Carlo Markov Chain
(MCMC) is rapidly mixing in the support of a {\em homogeneous} strongly
Rayleigh distribution. As a byproduct, our proof implies Markov chains can be
used to efficiently generate approximate samples of a $k$-determinantal point
process. This answers an open question raised by Deshpande and Rademacher.
|
Peak Criterion for Choosing Gaussian Kernel Bandwidth in Support Vector
Data Description | Support Vector Data Description (SVDD) is a machine-learning technique used
for single class classification and outlier detection. SVDD formulation with
kernel function provides a flexible boundary around data. The value of kernel
function parameters affects the nature of the data boundary. For example, it is
observed that with a Gaussian kernel, as the value of kernel bandwidth is
lowered, the data boundary changes from spherical to wiggly. The spherical data
boundary leads to underfitting, and an extremely wiggly data boundary leads to
overfitting. In this paper, we propose empirical criterion to obtain good
values of the Gaussian kernel bandwidth parameter. This criterion provides a
smooth boundary that captures the essential geometric features of the data.
|
Anomaly Detection in Clutter using Spectrally Enhanced Ladar | Discrete return (DR) Laser Detection and Ranging (Ladar) systems provide a
series of echoes that reflect from objects in a scene. These can be first, last
or multi-echo returns. In contrast, Full-Waveform (FW)-Ladar systems measure
the intensity of light reflected from objects continuously over a period of
time. In a camouflaged scenario, e.g., objects hidden behind dense foliage, a
FW-Ladar penetrates such foliage and returns a sequence of echoes including
buried faint echoes. The aim of this paper is to learn local-patterns of
co-occurring echoes characterised by their measured spectra. A deviation from
such patterns defines an abnormal event in a forest/tree depth profile. As far
as the authors know, neither DR or FW-Ladar, along with several spectral
measurements, has not been applied to anomaly detection. This work presents an
algorithm that allows detection of spectral and temporal anomalies in FW-Multi
Spectral Ladar (FW-MSL) data samples. An anomaly is defined as a full waveform
temporal and spectral signature that does not conform to a prior expectation,
represented using a learnt subspace (dictionary) and set of coefficients that
capture co-occurring local-patterns using an overlapping temporal window. A
modified optimization scheme is proposed for subspace learning based on
stochastic approximations. The objective function is augmented with a
discriminative term that represents the subspace's separability properties and
supports anomaly characterisation. The algorithm detects several man-made
objects and anomalous spectra hidden in a dense clutter of vegetation and also
allows tree species classification.
|
Choice by Elimination via Deep Neural Networks | We introduce Neural Choice by Elimination, a new framework that integrates
deep neural networks into probabilistic sequential choice models for learning
to rank. Given a set of items to chose from, the elimination strategy starts
with the whole item set and iteratively eliminates the least worthy item in the
remaining subset. We prove that the choice by elimination is equivalent to
marginalizing out the random Gompertz latent utilities. Coupled with the choice
model is the recently introduced Neural Highway Networks for approximating
arbitrarily complex rank functions. We evaluate the proposed framework on a
large-scale public dataset with over 425K items, drawn from the Yahoo! learning
to rank challenge. It is demonstrated that the proposed method is competitive
against state-of-the-art learning to rank methods.
|
Label Noise Reduction in Entity Typing by Heterogeneous Partial-Label
Embedding | Current systems of fine-grained entity typing use distant supervision in
conjunction with existing knowledge bases to assign categories (type labels) to
entity mentions. However, the type labels so obtained from knowledge bases are
often noisy (i.e., incorrect for the entity mention's local context). We define
a new task, Label Noise Reduction in Entity Typing (LNR), to be the automatic
identification of correct type labels (type-paths) for training examples, given
the set of candidate type labels obtained by distant supervision with a given
type hierarchy. The unknown type labels for individual entity mentions and the
semantic similarity between entity types pose unique challenges for solving the
LNR task. We propose a general framework, called PLE, to jointly embed entity
mentions, text features and entity types into the same low-dimensional space
where, in that space, objects whose types are semantically close have similar
representations. Then we estimate the type-path for each training example in a
top-down manner using the learned embeddings. We formulate a global objective
for learning the embeddings from text corpora and knowledge bases, which adopts
a novel margin-based loss that is robust to noisy labels and faithfully models
type correlation derived from knowledge bases. Our experiments on three public
typing datasets demonstrate the effectiveness and robustness of PLE, with an
average of 25% improvement in accuracy compared to next best method.
|
Large Scale Kernel Learning using Block Coordinate Descent | We demonstrate that distributed block coordinate descent can quickly solve
kernel regression and classification problems with millions of data points.
Armed with this capability, we conduct a thorough comparison between the full
kernel, the Nystr\"om method, and random features on three large classification
tasks from various domains. Our results suggest that the Nystr\"om method
generally achieves better statistical accuracy than random features, but can
require significantly more iterations of optimization. Lastly, we derive new
rates for block coordinate descent which support our experimental findings when
specialized to kernel methods.
|
Relative Error Embeddings for the Gaussian Kernel Distance | A reproducing kernel can define an embedding of a data point into an infinite
dimensional reproducing kernel Hilbert space (RKHS). The norm in this space
describes a distance, which we call the kernel distance. The random Fourier
features (of Rahimi and Recht) describe an oblivious approximate mapping into
finite dimensional Euclidean space that behaves similar to the RKHS. We show in
this paper that for the Gaussian kernel the Euclidean norm between these mapped
to features has $(1+\epsilon)$-relative error with respect to the kernel
distance. When there are $n$ data points, we show that $O((1/\epsilon^2)
\log(n))$ dimensions of the approximate feature space are sufficient and
necessary.
Without a bound on $n$, but when the original points lie in $\mathbb{R}^d$
and have diameter bounded by $\mathcal{M}$, then we show that $O((d/\epsilon^2)
\log(\mathcal{M}))$ dimensions are sufficient, and that this many are required,
up to $\log(1/\epsilon)$ factors.
|
Recommendations as Treatments: Debiasing Learning and Evaluation | Most data for evaluating and training recommender systems is subject to
selection biases, either through self-selection by the users or through the
actions of the recommendation system itself. In this paper, we provide a
principled approach to handling selection biases, adapting models and
estimation techniques from causal inference. The approach leads to unbiased
performance estimators despite biased data, and to a matrix factorization
method that provides substantially improved prediction performance on
real-world data. We theoretically and empirically characterize the robustness
of the approach, finding that it is highly practical and scalable.
|
Online optimization and regret guarantees for non-additive long-term
constraints | We consider online optimization in the 1-lookahead setting, where the
objective does not decompose additively over the rounds of the online game. The
resulting formulation enables us to deal with non-stationary and/or long-term
constraints , which arise, for example, in online display advertising problems.
We propose an on-line primal-dual algorithm for which we obtain dynamic
cumulative regret guarantees. They depend on the convexity and the smoothness
of the non-additive penalty, as well as terms capturing the smoothness with
which the residuals of the non-stationary and long-term constraints vary over
the rounds. We conduct experiments on synthetic data to illustrate the benefits
of the non-additive penalty and show vanishing regret convergence on live
traffic data collected by a display advertising platform in production.
|
Harder, Better, Faster, Stronger Convergence Rates for Least-Squares
Regression | We consider the optimization of a quadratic objective function whose
gradients are only accessible through a stochastic oracle that returns the
gradient at any given point plus a zero-mean finite variance random error. We
present the first algorithm that achieves jointly the optimal prediction error
rates for least-squares regression, both in terms of forgetting of initial
conditions in O(1/n 2), and in terms of dependence on the noise and dimension d
of the problem, as O(d/n). Our new algorithm is based on averaged accelerated
regularized gradient descent, and may also be analyzed through finer
assumptions on initial conditions and the Hessian matrix, leading to
dimension-free quantities that may still be small while the " optimal " terms
above are large. In order to characterize the tightness of these new bounds, we
consider an application to non-parametric regression and use the known lower
bounds on the statistical performance (without computational limits), which
happen to match our bounds obtained from a single pass on the data and thus
show optimality of our algorithm in a wide variety of particular trade-offs
between bias and variance.
|
Low-Rank Factorization of Determinantal Point Processes for
Recommendation | Determinantal point processes (DPPs) have garnered attention as an elegant
probabilistic model of set diversity. They are useful for a number of subset
selection tasks, including product recommendation. DPPs are parametrized by a
positive semi-definite kernel matrix. In this work we present a new method for
learning the DPP kernel from observed data using a low-rank factorization of
this kernel. We show that this low-rank factorization enables a learning
algorithm that is nearly an order of magnitude faster than previous approaches,
while also providing for a method for computing product recommendation
predictions that is far faster (up to 20x faster or more for large item
catalogs) than previous techniques that involve a full-rank DPP kernel.
Furthermore, we show that our method provides equivalent or sometimes better
predictive performance than prior full-rank DPP approaches, and better
performance than several other competing recommendation methods in many cases.
We conduct an extensive experimental evaluation using several real-world
datasets in the domain of product recommendation to demonstrate the utility of
our method, along with its limitations.
|
Cell segmentation with random ferns and graph-cuts | The progress in imaging techniques have allowed the study of various aspect
of cellular mechanisms. To isolate individual cells in live imaging data, we
introduce an elegant image segmentation framework that effectively extracts
cell boundaries, even in the presence of poor edge details. Our approach works
in two stages. First, we estimate pixel interior/border/exterior class
probabilities using random ferns. Then, we use an energy minimization framework
to compute boundaries whose localization is compliant with the pixel class
probabilities. We validate our approach on a manually annotated dataset.
|
Auxiliary Deep Generative Models | Deep generative models parameterized by neural networks have recently
achieved state-of-the-art performance in unsupervised and semi-supervised
learning. We extend deep generative models with auxiliary variables which
improves the variational approximation. The auxiliary variables leave the
generative model unchanged but make the variational distribution more
expressive. Inspired by the structure of the auxiliary variable we also propose
a model with two stochastic layers and skip connections. Our findings suggest
that more expressive and properly specified deep generative models converge
faster with better results. We show state-of-the-art performance within
semi-supervised learning on MNIST, SVHN and NORB datasets.
|
Multi-layer Representation Learning for Medical Concepts | Learning efficient representations for concepts has been proven to be an
important basis for many applications such as machine translation or document
classification. Proper representations of medical concepts such as diagnosis,
medication, procedure codes and visits will have broad applications in
healthcare analytics. However, in Electronic Health Records (EHR) the visit
sequences of patients include multiple concepts (diagnosis, procedure, and
medication codes) per visit. This structure provides two types of relational
information, namely sequential order of visits and co-occurrence of the codes
within each visit. In this work, we propose Med2Vec, which not only learns
distributed representations for both medical codes and visits from a large EHR
dataset with over 3 million visits, but also allows us to interpret the learned
representations confirmed positively by clinical experts. In the experiments,
Med2Vec displays significant improvement in key medical applications compared
to popular baselines such as Skip-gram, GloVe and stacked autoencoder, while
providing clinically meaningful interpretation.
|
Communication-Efficient Learning of Deep Networks from Decentralized
Data | Modern mobile devices have access to a wealth of data suitable for learning
models, which in turn can greatly improve the user experience on the device.
For example, language models can improve speech recognition and text entry, and
image models can automatically select good photos. However, this rich data is
often privacy sensitive, large in quantity, or both, which may preclude logging
to the data center and training there using conventional approaches. We
advocate an alternative that leaves the training data distributed on the mobile
devices, and learns a shared model by aggregating locally-computed updates. We
term this decentralized approach Federated Learning.
We present a practical method for the federated learning of deep networks
based on iterative model averaging, and conduct an extensive empirical
evaluation, considering five different model architectures and four datasets.
These experiments demonstrate the approach is robust to the unbalanced and
non-IID data distributions that are a defining characteristic of this setting.
Communication costs are the principal constraint, and we show a reduction in
required communication rounds by 10-100x as compared to synchronized stochastic
gradient descent.
|
Boost Picking: A Universal Method on Converting Supervised
Classification to Semi-supervised Classification | This paper proposes a universal method, Boost Picking, to train supervised
classification models mainly by un-labeled data. Boost Picking only adopts two
weak classifiers to estimate and correct the error. It is theoretically proved
that Boost Picking could train a supervised model mainly by un-labeled data as
effectively as the same model trained by 100% labeled data, only if recalls of
the two weak classifiers are all greater than zero and the sum of precisions is
greater than one. Based on Boost Picking, we present "Test along with Training
(TawT)" to improve the generalization of supervised models. Both Boost Picking
and TawT are successfully tested in varied little data sets.
|
Audio Recording Device Identification Based on Deep Learning | In this paper we present a research on identification of audio recording
devices from background noise, thus providing a method for forensics. The audio
signal is the sum of speech signal and noise signal. Usually, people pay more
attention to speech signal, because it carries the information to deliver. So a
great amount of researches have been dedicated to getting higher
Signal-Noise-Ratio (SNR). There are many speech enhancement algorithms to
improve the quality of the speech, which can be seen as reducing the noise.
However, noises can be regarded as the intrinsic fingerprint traces of an audio
recording device. These digital traces can be characterized and identified by
new machine learning techniques. Therefore, in our research, we use the noise
as the intrinsic features. As for the identification, multiple classifiers of
deep learning methods are used and compared. The identification result shows
that the method of getting feature vector from the noise of each device and
identifying them with deep learning techniques is viable, and well-preformed.
|
Adaptive Least Mean Squares Estimation of Graph Signals | The aim of this paper is to propose a least mean squares (LMS) strategy for
adaptive estimation of signals defined over graphs. Assuming the graph signal
to be band-limited, over a known bandwidth, the method enables reconstruction,
with guaranteed performance in terms of mean-square error, and tracking from a
limited number of observations over a subset of vertices. A detailed mean
square analysis provides the performance of the proposed method, and leads to
several insights for designing useful sampling strategies for graph signals.
Numerical results validate our theoretical findings, and illustrate the
performance of the proposed method. Furthermore, to cope with the case where
the bandwidth is not known beforehand, we propose a method that performs a
sparse online estimation of the signal support in the (graph) frequency domain,
which enables online adaptation of the graph sampling strategy. Finally, we
apply the proposed method to build the power spatial density cartography of a
given operational region in a cognitive network environment.
|
An improved analysis of the ER-SpUD dictionary learning algorithm | In "dictionary learning" we observe $Y = AX + E$ for some
$Y\in\mathbb{R}^{n\times p}$, $A \in\mathbb{R}^{m\times n}$, and
$X\in\mathbb{R}^{m\times p}$. The matrix $Y$ is observed, and $A, X, E$ are
unknown. Here $E$ is "noise" of small norm, and $X$ is column-wise sparse. The
matrix $A$ is referred to as a {\em dictionary}, and its columns as {\em
atoms}. Then, given some small number $p$ of samples, i.e.\ columns of $Y$, the
goal is to learn the dictionary $A$ up to small error, as well as $X$. The
motivation is that in many applications data is expected to sparse when
represented by atoms in the "right" dictionary $A$ (e.g.\ images in the Haar
wavelet basis), and the goal is to learn $A$ from the data to then use it for
other applications.
Recently, [SWW12] proposed the dictionary learning algorithm ER-SpUD with
provable guarantees when $E = 0$ and $m = n$. They showed if $X$ has
independent entries with an expected $s$ non-zeroes per column for $1 \lesssim
s \lesssim \sqrt{n}$, and with non-zero entries being subgaussian, then for
$p\gtrsim n^2\log^2 n$ with high probability ER-SpUD outputs matrices $A', X'$
which equal $A, X$ up to permuting and scaling columns (resp.\ rows) of $A$
(resp.\ $X$). They conjectured $p\gtrsim n\log n$ suffices, which they showed
was information theoretically necessary for {\em any} algorithm to succeed when
$s \simeq 1$. Significant progress was later obtained in [LV15].
We show that for a slight variant of ER-SpUD, $p\gtrsim n\log(n/\delta)$
samples suffice for successful recovery with probability $1-\delta$. We also
show that for the unmodified ER-SpUD, $p\gtrsim n^{1.99}$ samples are required
even to learn $A, X$ with polynomially small success probability. This resolves
the main conjecture of [SWW12], and contradicts the main result of [LV15],
which claimed that $p\gtrsim n\log^4 n$ guarantees success whp.
|
Toward Deeper Understanding of Neural Networks: The Power of
Initialization and a Dual View on Expressivity | We develop a general duality between neural networks and compositional
kernels, striving towards a better understanding of deep learning. We show that
initial representations generated by common random initializations are
sufficiently rich to express all functions in the dual kernel space. Hence,
though the training objective is hard to optimize in the worst case, the
initial weights form a good starting point for optimization. Our dual view also
reveals a pragmatic and aesthetic perspective of neural networks and
underscores their expressive power.
|
Efficient approaches for escaping higher order saddle points in
non-convex optimization | Local search heuristics for non-convex optimizations are popular in applied
machine learning. However, in general it is hard to guarantee that such
algorithms even converge to a local minimum, due to the existence of
complicated saddle point structures in high dimensions. Many functions have
degenerate saddle points such that the first and second order derivatives
cannot distinguish them with local optima. In this paper we use higher order
derivatives to escape these saddle points: we design the first efficient
algorithm guaranteed to converge to a third order local optimum (while existing
techniques are at most second order). We also show that it is NP-hard to extend
this further to finding fourth order local optima.
|
Local Rademacher Complexity-based Learning Guarantees for Multi-Task
Learning | We show a Talagrand-type concentration inequality for Multi-Task Learning
(MTL), using which we establish sharp excess risk bounds for MTL in terms of
distribution- and data-dependent versions of the Local Rademacher Complexity
(LRC). We also give a new bound on the LRC for norm regularized as well as
strongly convex hypothesis classes, which applies not only to MTL but also to
the standard i.i.d. setting. Combining both results, one can now easily derive
fast-rate bounds on the excess risk for many prominent MTL methods,
including---as we demonstrate---Schatten-norm, group-norm, and
graph-regularized MTL. The derived bounds reflect a relationship akeen to a
conservation law of asymptotic convergence rates. This very relationship allows
for trading off slower rates w.r.t. the number of tasks for faster rates with
respect to the number of available samples per task, when compared to the rates
obtained via a traditional, global Rademacher analysis.
|
Weighted Unsupervised Learning for 3D Object Detection | This paper introduces a novel weighted unsupervised learning for object
detection using an RGB-D camera. This technique is feasible for detecting the
moving objects in the noisy environments that are captured by an RGB-D camera.
The main contribution of this paper is a real-time algorithm for detecting each
object using weighted clustering as a separate cluster. In a preprocessing
step, the algorithm calculates the pose 3D position X, Y, Z and RGB color of
each data point and then it calculates each data point's normal vector using
the point's neighbor. After preprocessing, our algorithm calculates k-weights
for each data point; each weight indicates membership. Resulting in clustered
objects of the scene.
|
Revise Saturated Activation Functions | In this paper, we revise two commonly used saturated functions, the logistic
sigmoid and the hyperbolic tangent (tanh).
We point out that, besides the well-known non-zero centered property, slope
of the activation function near the origin is another possible reason making
training deep networks with the logistic function difficult to train. We
demonstrate that, with proper rescaling, the logistic sigmoid achieves
comparable results with tanh.
Then following the same argument, we improve tahn by penalizing in the
negative part. We show that "penalized tanh" is comparable and even outperforms
the state-of-the-art non-saturated functions including ReLU and leaky ReLU on
deep convolution neural networks.
Our results contradict to the conclusion of previous works that the
saturation property causes the slow convergence. It suggests further
investigation is necessary to better understand activation functions in deep
architectures.
|
Spectral Learning for Supervised Topic Models | Supervised topic models simultaneously model the latent topic structure of
large collections of documents and a response variable associated with each
document. Existing inference methods are based on variational approximation or
Monte Carlo sampling, which often suffers from the local minimum defect.
Spectral methods have been applied to learn unsupervised topic models, such as
latent Dirichlet allocation (LDA), with provable guarantees. This paper
investigates the possibility of applying spectral methods to recover the
parameters of supervised LDA (sLDA). We first present a two-stage spectral
method, which recovers the parameters of LDA followed by a power update method
to recover the regression model parameters. Then, we further present a
single-phase spectral algorithm to jointly recover the topic distribution
matrix as well as the regression weights. Our spectral algorithms are provably
correct and computationally efficient. We prove a sample complexity bound for
each algorithm and subsequently derive a sufficient condition for the
identifiability of sLDA. Thorough experiments on synthetic and real-world
datasets verify the theory and demonstrate the practical effectiveness of the
spectral algorithms. In fact, our results on a large-scale review rating
dataset demonstrate that our single-phase spectral algorithm alone gets
comparable or even better performance than state-of-the-art methods, while
previous work on spectral methods has rarely reported such promising
performance.
|
Structured Sparse Regression via Greedy Hard-Thresholding | Several learning applications require solving high-dimensional regression
problems where the relevant features belong to a small number of (overlapping)
groups. For very large datasets and under standard sparsity constraints, hard
thresholding methods have proven to be extremely efficient, but such methods
require NP hard projections when dealing with overlapping groups. In this
paper, we show that such NP-hard projections can not only be avoided by
appealing to submodular optimization, but such methods come with strong
theoretical guarantees even in the presence of poorly conditioned data (i.e.
say when two features have correlation $\geq 0.99$), which existing analyses
cannot handle. These methods exhibit an interesting computation-accuracy
trade-off and can be extended to significantly harder problems such as sparse
overlapping groups. Experiments on both real and synthetic data validate our
claims and demonstrate that the proposed methods are orders of magnitude faster
than other greedy and convex relaxation techniques for learning with
group-structured sparsity.
|
First-order Methods for Geodesically Convex Optimization | Geodesic convexity generalizes the notion of (vector space) convexity to
nonlinear metric spaces. But unlike convex optimization, geodesically convex
(g-convex) optimization is much less developed. In this paper we contribute to
the understanding of g-convex optimization by developing iteration complexity
analysis for several first-order algorithms on Hadamard manifolds.
Specifically, we prove upper bounds for the global complexity of deterministic
and stochastic (sub)gradient methods for optimizing smooth and nonsmooth
g-convex functions, both with and without strong g-convexity. Our analysis also
reveals how the manifold geometry, especially \emph{sectional curvature},
impacts convergence rates. To the best of our knowledge, our work is the first
to provide global complexity analysis for first-order algorithms for general
g-convex optimization.
|
Node-By-Node Greedy Deep Learning for Interpretable Features | Multilayer networks have seen a resurgence under the umbrella of deep
learning. Current deep learning algorithms train the layers of the network
sequentially, improving algorithmic performance as well as providing some
regularization. We present a new training algorithm for deep networks which
trains \emph{each node in the network} sequentially. Our algorithm is orders of
magnitude faster, creates more interpretable internal representations at the
node level, while not sacrificing on the ultimate out-of-sample performance.
|
GAP Safe Screening Rules for Sparse-Group-Lasso | In high dimensional settings, sparse structures are crucial for efficiency,
either in term of memory, computation or performance. In some contexts, it is
natural to handle more refined structures than pure sparsity, such as for
instance group sparsity. Sparse-Group Lasso has recently been introduced in the
context of linear regression to enforce sparsity both at the feature level and
at the group level. We adapt to the case of Sparse-Group Lasso recent safe
screening rules that discard early in the solver irrelevant features/groups.
Such rules have led to important speed-ups for a wide range of iterative
methods. Thanks to dual gap computations, we provide new safe screening rules
for Sparse-Group Lasso and show significant gains in term of computing time for
a coordinate descent implementation.
|
Stacking for machine learning redshifts applied to SDSS galaxies | We present an analysis of a general machine learning technique called
'stacking' for the estimation of photometric redshifts. Stacking techniques can
feed the photometric redshift estimate, as output by a base algorithm, back
into the same algorithm as an additional input feature in a subsequent learning
round. We shown how all tested base algorithms benefit from at least one
additional stacking round (or layer). To demonstrate the benefit of stacking,
we apply the method to both unsupervised machine learning techniques based on
self-organising maps (SOMs), and supervised machine learning methods based on
decision trees. We explore a range of stacking architectures, such as the
number of layers and the number of base learners per layer. Finally we explore
the effectiveness of stacking even when using a successful algorithm such as
AdaBoost. We observe a significant improvement of between 1.9% and 21% on all
computed metrics when stacking is applied to weak learners (such as SOMs and
decision trees). When applied to strong learning algorithms (such as AdaBoost)
the ratio of improvement shrinks, but still remains positive and is between
0.4% and 2.5% for the explored metrics and comes at almost no additional
computational cost.
|
Policy Error Bounds for Model-Based Reinforcement Learning with Factored
Linear Models | In this paper we study a model-based approach to calculating approximately
optimal policies in Markovian Decision Processes. In particular, we derive
novel bounds on the loss of using a policy derived from a factored linear
model, a class of models which generalize numerous previous models out of those
that come with strong computational guarantees. For the first time in the
literature, we derive performance bounds for model-based techniques where the
model inaccuracy is measured in weighted norms. Moreover, our bounds show a
decreased sensitivity to the discount factor and, unlike similar bounds derived
for other approaches, they are insensitive to measure mismatch. Similarly to
previous works, our proofs are also based on contraction arguments, but with
the main differences that we use carefully constructed norms building on Banach
lattices, and the contraction property is only assumed for operators acting on
"compressed" spaces, thus weakening previous assumptions, while strengthening
previous results.
|
FLASH: Fast Bayesian Optimization for Data Analytic Pipelines | Modern data science relies on data analytic pipelines to organize
interdependent computational steps. Such analytic pipelines often involve
different algorithms across multiple steps, each with its own hyperparameters.
To achieve the best performance, it is often critical to select optimal
algorithms and to set appropriate hyperparameters, which requires large
computational efforts. Bayesian optimization provides a principled way for
searching optimal hyperparameters for a single algorithm. However, many
challenges remain in solving pipeline optimization problems with
high-dimensional and highly conditional search space. In this work, we propose
Fast LineAr SearcH (FLASH), an efficient method for tuning analytic pipelines.
FLASH is a two-layer Bayesian optimization framework, which firstly uses a
parametric model to select promising algorithms, then computes a nonparametric
model to fine-tune hyperparameters of the promising algorithms. FLASH also
includes an effective caching algorithm which can further accelerate the search
process. Extensive experiments on a number of benchmark datasets have
demonstrated that FLASH significantly outperforms previous state-of-the-art
methods in both search speed and accuracy. Using 50% of the time budget, FLASH
achieves up to 20% improvement on test error rate compared to the baselines.
FLASH also yields state-of-the-art performance on a real-world application for
healthcare predictive modeling.
|
Distributed Private Online Learning for Social Big Data Computing over
Data Center Networks | With the rapid growth of Internet technologies, cloud computing and social
networks have become ubiquitous. An increasing number of people participate in
social networks and massive online social data are obtained. In order to
exploit knowledge from copious amounts of data obtained and predict social
behavior of users, we urge to realize data mining in social networks. Almost
all online websites use cloud services to effectively process the large scale
of social data, which are gathered from distributed data centers. These data
are so large-scale, high-dimension and widely distributed that we propose a
distributed sparse online algorithm to handle them. Additionally,
privacy-protection is an important point in social networks. We should not
compromise the privacy of individuals in networks, while these social data are
being learned for data mining. Thus we also consider the privacy problem in
this article. Our simulations shows that the appropriate sparsity of data would
enhance the performance of our algorithm and the privacy-preserving method does
not significantly hurt the performance of the proposed algorithm.
|
Uniform Hypergraph Partitioning: Provable Tensor Methods and Sampling
Techniques | In a series of recent works, we have generalised the consistency results in
the stochastic block model literature to the case of uniform and non-uniform
hypergraphs. The present paper continues the same line of study, where we focus
on partitioning weighted uniform hypergraphs---a problem often encountered in
computer vision. This work is motivated by two issues that arise when a
hypergraph partitioning approach is used to tackle computer vision problems:
(i) The uniform hypergraphs constructed for higher-order learning contain all
edges, but most have negligible weights. Thus, the adjacency tensor is nearly
sparse, and yet, not binary. (ii) A more serious concern is that standard
partitioning algorithms need to compute all edge weights, which is
computationally expensive for hypergraphs. This is usually resolved in practice
by merging the clustering algorithm with a tensor sampling strategy---an
approach that is yet to be analysed rigorously. We build on our earlier work on
partitioning dense unweighted uniform hypergraphs (Ghoshdastidar and Dukkipati,
ICML, 2015), and address the aforementioned issues by proposing provable and
efficient partitioning algorithms. Our analysis justifies the empirical success
of practical sampling techniques. We also complement our theoretical findings
by elaborate empirical comparison of various hypergraph partitioning schemes.
|
Multi-Task Learning with Labeled and Unlabeled Tasks | In multi-task learning, a learner is given a collection of prediction tasks
and needs to solve all of them. In contrast to previous work, which required
that annotated training data is available for all tasks, we consider a new
setting, in which for some tasks, potentially most of them, only unlabeled
training data is provided. Consequently, to solve all tasks, information must
be transferred between tasks with labels and tasks without labels. Focusing on
an instance-based transfer method we analyze two variants of this setting: when
the set of labeled tasks is fixed, and when it can be actively selected by the
learner. We state and prove a generalization bound that covers both scenarios
and derive from it an algorithm for making the choice of labeled tasks (in the
active case) and for transferring information between the tasks in a principled
way. We also illustrate the effectiveness of the algorithm by experiments on
synthetic and real data.
|
Machine learning meets network science: dimensionality reduction for
fast and efficient embedding of networks in the hyperbolic space | Complex network topologies and hyperbolic geometry seem specularly connected,
and one of the most fascinating and challenging problems of recent complex
network theory is to map a given network to its hyperbolic space. The
Popularity Similarity Optimization (PSO) model represents - at the moment - the
climax of this theory. It suggests that the trade-off between node popularity
and similarity is a mechanism to explain how complex network topologies emerge
- as discrete samples - from the continuous world of hyperbolic geometry. The
hyperbolic space seems appropriate to represent real complex networks. In fact,
it preserves many of their fundamental topological properties, and can be
exploited for real applications such as, among others, link prediction and
community detection. Here, we observe for the first time that a
topological-based machine learning class of algorithms - for nonlinear
unsupervised dimensionality reduction - can directly approximate the network's
node angular coordinates of the hyperbolic model into a two-dimensional space,
according to a similar topological organization that we named angular
coalescence. On the basis of this phenomenon, we propose a new class of
algorithms that offers fast and accurate coalescent embedding of networks in
the hyperbolic space even for graphs with thousands of nodes.
|
Multi-task and Lifelong Learning of Kernels | We consider a problem of learning kernels for use in SVM classification in
the multi-task and lifelong scenarios and provide generalization bounds on the
error of a large margin classifier. Our results show that, under mild
conditions on the family of kernels used for learning, solving several related
tasks simultaneously is beneficial over single task learning. In particular, as
the number of observed tasks grows, assuming that in the considered family of
kernels there exists one that yields low approximation error on all tasks, the
overhead associated with learning such a kernel vanishes and the complexity
converges to that of learning when this good kernel is given to the learner.
|
Determining the best attributes for surveillance video keywords
generation | Automatic video keyword generation is one of the key ingredients in reducing
the burden of security officers in analyzing surveillance videos. Keywords or
attributes are generally chosen manually based on expert knowledge of
surveillance. Most existing works primarily aim at either supervised learning
approaches relying on extensive manual labelling or hierarchical probabilistic
models that assume the features are extracted using the bag-of-words approach;
thus limiting the utilization of the other features. To address this, we turn
our attention to automatic attribute discovery approaches. However, it is not
clear which automatic discovery approach can discover the most meaningful
attributes. Furthermore, little research has been done on how to compare and
choose the best automatic attribute discovery methods. In this paper, we
propose a novel approach, based on the shared structure exhibited amongst
meaningful attributes, that enables us to compare between different automatic
attribute discovery approaches.We then validate our approach by comparing
various attribute discovery methods such as PiCoDeS on two attribute datasets.
The evaluation shows that our approach is able to select the automatic
discovery approach that discovers the most meaningful attributes. We then
employ the best discovery approach to generate keywords for videos recorded
from a surveillance system. This work shows it is possible to massively reduce
the amount of manual work in generating video keywords without limiting
ourselves to a particular video feature descriptor.
|
Semi-Markov Switching Vector Autoregressive Model-based Anomaly
Detection in Aviation Systems | In this work we consider the problem of anomaly detection in heterogeneous,
multivariate, variable-length time series datasets. Our focus is on the
aviation safety domain, where data objects are flights and time series are
sensor readings and pilot switches. In this context the goal is to detect
anomalous flight segments, due to mechanical, environmental, or human factors
in order to identifying operationally significant events and provide insights
into the flight operations and highlight otherwise unavailable potential safety
risks and precursors to accidents. For this purpose, we propose a framework
which represents each flight using a semi-Markov switching vector
autoregressive (SMS-VAR) model. Detection of anomalies is then based on
measuring dissimilarities between the model's prediction and data observation.
The framework is scalable, due to the inherent parallel nature of most
computations, and can be used to perform online anomaly detection. Extensive
experimental results on simulated and real datasets illustrate that the
framework can detect various types of anomalies along with the key parameters
involved.
|
Deep Learning in Finance | We explore the use of deep learning hierarchical models for problems in
financial prediction and classification. Financial prediction problems -- such
as those presented in designing and pricing securities, constructing
portfolios, and risk management -- often involve large data sets with complex
data interactions that currently are difficult or impossible to specify in a
full economic model. Applying deep learning methods to these problems can
produce more useful results than standard methods in finance. In particular,
deep learning can detect and exploit interactions in the data that are, at
least currently, invisible to any existing financial economic theory.
|
Interactive Storytelling over Document Collections | Storytelling algorithms aim to 'connect the dots' between disparate documents
by linking starting and ending documents through a series of intermediate
documents. Existing storytelling algorithms are based on notions of coherence
and connectivity, and thus the primary way by which users can steer the story
construction is via design of suitable similarity functions. We present an
alternative approach to storytelling wherein the user can interactively and
iteratively provide 'must use' constraints to preferentially support the
construction of some stories over others. The three innovations in our approach
are distance measures based on (inferred) topic distributions, the use of
constraints to define sets of linear inequalities over paths, and the
introduction of slack and surplus variables to condition the topic distribution
to preferentially emphasize desired terms over others. We describe experimental
results to illustrate the effectiveness of our interactive storytelling
approach over multiple text datasets.
|
2-Bit Random Projections, NonLinear Estimators, and Approximate Near
Neighbor Search | The method of random projections has become a standard tool for machine
learning, data mining, and search with massive data at Web scale. The effective
use of random projections requires efficient coding schemes for quantizing
(real-valued) projected data into integers. In this paper, we focus on a simple
2-bit coding scheme. In particular, we develop accurate nonlinear estimators of
data similarity based on the 2-bit strategy. This work will have important
practical applications. For example, in the task of near neighbor search, a
crucial step (often called re-ranking) is to compute or estimate data
similarities once a set of candidate data points have been identified by hash
table techniques. This re-ranking step can take advantage of the proposed
coding scheme and estimator.
As a related task, in this paper, we also study a simple uniform quantization
scheme for the purpose of building hash tables with projected data. Our
analysis shows that typically only a small number of bits are needed. For
example, when the target similarity level is high, 2 or 3 bits might be
sufficient. When the target similarity level is not so high, it is preferable
to use only 1 or 2 bits. Therefore, a 2-bit scheme appears to be overall a good
choice for the task of sublinear time approximate near neighbor search via hash
tables.
Combining these results, we conclude that 2-bit random projections should be
recommended for approximate near neighbor search and similarity estimation.
Extensive experimental results are provided.
|
Recovering Structured Probability Matrices | We consider the problem of accurately recovering a matrix B of size M by M ,
which represents a probability distribution over M2 outcomes, given access to
an observed matrix of "counts" generated by taking independent samples from the
distribution B. How can structural properties of the underlying matrix B be
leveraged to yield computationally efficient and information theoretically
optimal reconstruction algorithms? When can accurate reconstruction be
accomplished in the sparse data regime? This basic problem lies at the core of
a number of questions that are currently being considered by different
communities, including building recommendation systems and collaborative
filtering in the sparse data regime, community detection in sparse random
graphs, learning structured models such as topic models or hidden Markov
models, and the efforts from the natural language processing community to
compute "word embeddings".
Our results apply to the setting where B has a low rank structure. For this
setting, we propose an efficient algorithm that accurately recovers the
underlying M by M matrix using Theta(M) samples. This result easily translates
to Theta(M) sample algorithms for learning topic models and learning hidden
Markov Models. These linear sample complexities are optimal, up to constant
factors, in an extremely strong sense: even testing basic properties of the
underlying matrix (such as whether it has rank 1 or 2) requires Omega(M)
samples. We provide an even stronger lower bound where distinguishing whether a
sequence of observations were drawn from the uniform distribution over M
observations versus being generated by an HMM with two hidden states requires
Omega(M) observations. This precludes sublinear-sample hypothesis tests for
basic properties, such as identity or uniformity, as well as sublinear sample
estimators for quantities such as the entropy rate of HMMs.
|
Clustering subgaussian mixtures by semidefinite programming | We introduce a model-free relax-and-round algorithm for k-means clustering
based on a semidefinite relaxation due to Peng and Wei. The algorithm
interprets the SDP output as a denoised version of the original data and then
rounds this output to a hard clustering. We provide a generic method for
proving performance guarantees for this algorithm, and we analyze the algorithm
in the context of subgaussian mixture models. We also study the fundamental
limits of estimating Gaussian centers by k-means clustering in order to compare
our approximation guarantee to the theoretically optimal k-means clustering
solution.
|
Structured Learning of Binary Codes with Column Generation | Hashing methods aim to learn a set of hash functions which map the original
features to compact binary codes with similarity preserving in the Hamming
space. Hashing has proven a valuable tool for large-scale information
retrieval. We propose a column generation based binary code learning framework
for data-dependent hash function learning. Given a set of triplets that encode
the pairwise similarity comparison information, our column generation based
method learns hash functions that preserve the relative comparison relations
within the large-margin learning framework. Our method iteratively learns the
best hash functions during the column generation procedure. Existing hashing
methods optimize over simple objectives such as the reconstruction error or
graph Laplacian related loss functions, instead of the performance evaluation
criteria of interest---multivariate performance measures such as the AUC and
NDCG. Our column generation based method can be further generalized from the
triplet loss to a general structured learning based framework that allows one
to directly optimize multivariate performance measures. For optimizing general
ranking measures, the resulting optimization problem can involve exponentially
or infinitely many variables and constraints, which is more challenging than
standard structured output learning. We use a combination of column generation
and cutting-plane techniques to solve the optimization problem. To speed-up the
training we further explore stage-wise training and propose to use a simplified
NDCG loss for efficient inference. We demonstrate the generality of our method
by applying it to ranking prediction and image retrieval, and show that it
outperforms a few state-of-the-art hashing methods.
|
Recurrent Orthogonal Networks and Long-Memory Tasks | Although RNNs have been shown to be powerful tools for processing sequential
data, finding architectures or optimization strategies that allow them to model
very long term dependencies is still an active area of research. In this work,
we carefully analyze two synthetic datasets originally outlined in (Hochreiter
and Schmidhuber, 1997) which are used to evaluate the ability of RNNs to store
information over many time steps. We explicitly construct RNN solutions to
these problems, and using these constructions, illuminate both the problems
themselves and the way in which RNNs store different types of information in
their hidden states. These constructions furthermore explain the success of
recent methods that specify unitary initializations or constraints on the
transition matrices.
|
An Effective and Efficient Approach for Clusterability Evaluation | Clustering is an essential data mining tool that aims to discover inherent
cluster structure in data. As such, the study of clusterability, which
evaluates whether data possesses such structure, is an integral part of cluster
analysis. Yet, despite their central role in the theory and application of
clustering, current notions of clusterability fall short in two crucial aspects
that render them impractical; most are computationally infeasible and others
fail to classify the structure of real datasets.
In this paper, we propose a novel approach to clusterability evaluation that
is both computationally efficient and successfully captures the structure of
real data. Our method applies multimodality tests to the (one-dimensional) set
of pairwise distances based on the original, potentially high-dimensional data.
We present extensive analyses of our approach for both the Dip and Silverman
multimodality tests on real data as well as 17,000 simulations, demonstrating
the success of our approach as the first practical notion of clusterability.
|
Distributed Deep Learning Using Synchronous Stochastic Gradient Descent | We design and implement a distributed multinode synchronous SGD algorithm,
without altering hyper parameters, or compressing data, or altering algorithmic
behavior. We perform a detailed analysis of scaling, and identify optimal
design points for different networks. We demonstrate scaling of CNNs on 100s of
nodes, and present what we believe to be record training throughputs. A 512
minibatch VGG-A CNN training run is scaled 90X on 128 nodes. Also 256 minibatch
VGG-A and OverFeat-FAST networks are scaled 53X and 42X respectively on a 64
node cluster. We also demonstrate the generality of our approach via
best-in-class 6.5X scaling for a 7-layer DNN on 16 nodes. Thereafter we attempt
to democratize deep-learning by training on an Ethernet based AWS cluster and
show ~14X scaling on 16 nodes.
|
Variational inference for Monte Carlo objectives | Recent progress in deep latent variable models has largely been driven by the
development of flexible and scalable variational inference methods. Variational
training of this type involves maximizing a lower bound on the log-likelihood,
using samples from the variational posterior to compute the required gradients.
Recently, Burda et al. (2016) have derived a tighter lower bound using a
multi-sample importance sampling estimate of the likelihood and showed that
optimizing it yields models that use more of their capacity and achieve higher
likelihoods. This development showed the importance of such multi-sample
objectives and explained the success of several related approaches.
We extend the multi-sample approach to discrete latent variables and analyze
the difficulty encountered when estimating the gradients involved. We then
develop the first unbiased gradient estimator designed for importance-sampled
objectives and evaluate it at training generative and structured output
prediction models. The resulting estimator, which is based on low-variance
per-sample learning signals, is both simpler and more effective than the NVIL
estimator proposed for the single-sample variational objective, and is
competitive with the currently used biased estimators.
|
Convexification of Learning from Constraints | Regularized empirical risk minimization with constrained labels (in contrast
to fixed labels) is a remarkably general abstraction of learning. For common
loss and regularization functions, this optimization problem assumes the form
of a mixed integer program (MIP) whose objective function is non-convex. In
this form, the problem is resistant to standard optimization techniques. We
construct MIPs with the same solutions whose objective functions are convex.
Specifically, we characterize the tightest convex extension of the objective
function, given by the Legendre-Fenchel biconjugate. Computing values of this
tightest convex extension is NP-hard. However, by applying our characterization
to every function in an additive decomposition of the objective function, we
obtain a class of looser convex extensions that can be computed efficiently.
For some decompositions, common loss and regularization functions, we derive a
closed form.
|
Graph Regularized Low Rank Representation for Aerosol Optical Depth
Retrieval | In this paper, we propose a novel data-driven regression model for aerosol
optical depth (AOD) retrieval. First, we adopt a low rank representation (LRR)
model to learn a powerful representation of the spectral response. Then, graph
regularization is incorporated into the LRR model to capture the local
structure information and the nonlinear property of the remote-sensing data.
Since it is easy to acquire the rich satellite-retrieval results, we use them
as a baseline to construct the graph. Finally, the learned feature
representation is feeded into support vector machine (SVM) to retrieve AOD.
Experiments are conducted on two widely used data sets acquired by different
sensors, and the experimental results show that the proposed method can achieve
superior performance compared to the physical models and other state-of-the-art
empirical models.
|
Understanding Visual Concepts with Continuation Learning | We introduce a neural network architecture and a learning algorithm to
produce factorized symbolic representations. We propose to learn these concepts
by observing consecutive frames, letting all the components of the hidden
representation except a small discrete set (gating units) be predicted from the
previous frame, and let the factors of variation in the next frame be
represented entirely by these discrete gated units (corresponding to symbolic
representations). We demonstrate the efficacy of our approach on datasets of
faces undergoing 3D transformations and Atari 2600 games.
|
Higher-Order Low-Rank Regression | This paper proposes an efficient algorithm (HOLRR) to handle regression tasks
where the outputs have a tensor structure. We formulate the regression problem
as the minimization of a least square criterion under a multilinear rank
constraint, a difficult non convex problem. HOLRR computes efficiently an
approximate solution of this problem, with solid theoretical guarantees. A
kernel extension is also presented. Experiments on synthetic and real data show
that HOLRR outperforms multivariate and multilinear regression methods and is
considerably faster than existing tensor methods.
|
Principal Component Projection Without Principal Component Analysis | We show how to efficiently project a vector onto the top principal components
of a matrix, without explicitly computing these components. Specifically, we
introduce an iterative algorithm that provably computes the projection using
few calls to any black-box routine for ridge regression.
By avoiding explicit principal component analysis (PCA), our algorithm is the
first with no runtime dependence on the number of top principal components. We
show that it can be used to give a fast iterative method for the popular
principal component regression problem, giving the first major runtime
improvement over the naive method of combining PCA with regression.
To achieve our results, we first observe that ridge regression can be used to
obtain a "smooth projection" onto the top principal components. We then sharpen
this approximation to true projection using a low-degree polynomial
approximation to the matrix step function. Step function approximation is a
topic of long-term interest in scientific computing. We extend prior theory by
constructing polynomials with simple iterative structure and rigorously
analyzing their behavior under limited precision.
|
Clustering with a Reject Option: Interactive Clustering as Bayesian
Prior Elicitation | A good clustering can help a data analyst to explore and understand a data
set, but what constitutes a good clustering may depend on domain-specific and
application-specific criteria. These criteria can be difficult to formalize,
even when it is easy for an analyst to know a good clustering when she sees
one. We present a new approach to interactive clustering for data exploration,
called \ciif, based on a particularly simple feedback mechanism, in which an
analyst can choose to reject individual clusters and request new ones. The new
clusters should be different from previously rejected clusters while still
fitting the data well. We formalize this interaction in a novel Bayesian prior
elicitation framework. In each iteration, the prior is adapted to account for
all the previous feedback, and a new clustering is then produced from the
posterior distribution. To achieve the computational efficiency necessary for
an interactive setting, we propose an incremental optimization method over data
minibatches using Lagrangian relaxation. Experiments demonstrate that \ciif can
produce accurate and diverse clusterings.
|
Sparse Linear Regression via Generalized Orthogonal Least-Squares | Sparse linear regression, which entails finding a sparse solution to an
underdetermined system of linear equations, can formally be expressed as an
$l_0$-constrained least-squares problem. The Orthogonal Least-Squares (OLS)
algorithm sequentially selects the features (i.e., columns of the coefficient
matrix) to greedily find an approximate sparse solution. In this paper, a
generalization of Orthogonal Least-Squares which relies on a recursive relation
between the components of the optimal solution to select L features at each
step and solve the resulting overdetermined system of equations is proposed.
Simulation results demonstrate that the generalized OLS algorithm is
computationally efficient and achieves performance superior to that of existing
greedy algorithms broadly used in the literature.
|
Streaming PCA: Matching Matrix Bernstein and Near-Optimal Finite Sample
Guarantees for Oja's Algorithm | This work provides improved guarantees for streaming principle component
analysis (PCA). Given $A_1, \ldots, A_n\in \mathbb{R}^{d\times d}$ sampled
independently from distributions satisfying $\mathbb{E}[A_i] = \Sigma$ for
$\Sigma \succeq \mathbf{0}$, this work provides an $O(d)$-space linear-time
single-pass streaming algorithm for estimating the top eigenvector of $\Sigma$.
The algorithm nearly matches (and in certain cases improves upon) the accuracy
obtained by the standard batch method that computes top eigenvector of the
empirical covariance $\frac{1}{n} \sum_{i \in [n]} A_i$ as analyzed by the
matrix Bernstein inequality. Moreover, to achieve constant accuracy, our
algorithm improves upon the best previous known sample complexities of
streaming algorithms by either a multiplicative factor of $O(d)$ or
$1/\mathrm{gap}$ where $\mathrm{gap}$ is the relative distance between the top
two eigenvalues of $\Sigma$.
These results are achieved through a novel analysis of the classic Oja's
algorithm, one of the oldest and most popular algorithms for streaming PCA. In
particular, this work shows that simply picking a random initial point $w_0$
and applying the update rule $w_{i + 1} = w_i + \eta_i A_i w_i$ suffices to
accurately estimate the top eigenvector, with a suitable choice of $\eta_i$. We
believe our result sheds light on how to efficiently perform streaming PCA both
in theory and in practice and we hope that our analysis may serve as the basis
for analyzing many variants and extensions of streaming PCA.
|
Blind score normalization method for PLDA based speaker recognition | Probabilistic Linear Discriminant Analysis (PLDA) has become state-of-the-art
method for modeling $i$-vector space in speaker recognition task. However the
performance degradation is observed if enrollment data size differs from one
speaker to another. This paper presents a solution to such problem by
introducing new PLDA scoring normalization technique. Normalization parameters
are derived in a blind way, so that, unlike traditional \textit{ZT-norm}, no
extra development data is required. Moreover, proposed method has shown to be
optimal in terms of detection cost function. The experiments conducted on NIST
SRE 2014 database demonstrate an improved accuracy in a mixed enrollment number
condition.
|
Recovering the number of clusters in data sets with noise features using
feature rescaling factors | In this paper we introduce three methods for re-scaling data sets aiming at
improving the likelihood of clustering validity indexes to return the true
number of spherical Gaussian clusters with additional noise features. Our
method obtains feature re-scaling factors taking into account the structure of
a given data set and the intuitive idea that different features may have
different degrees of relevance at different clusters.
We experiment with the Silhouette (using squared Euclidean, Manhattan, and
the p$^{th}$ power of the Minkowski distance), Dunn's, Calinski-Harabasz and
Hartigan indexes on data sets with spherical Gaussian clusters with and without
noise features. We conclude that our methods indeed increase the chances of
estimating the true number of clusters in a data set.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.