title
stringlengths 7
246
| abstract
stringlengths 6
3.31k
|
|---|---|
Masking Strategies for Image Manifolds | We consider the problem of selecting an optimal mask for an image manifold,
i.e., choosing a subset of the pixels of the image that preserves the
manifold's geometric structure present in the original data. Such masking
implements a form of compressive sensing through emerging imaging sensor
platforms for which the power expense grows with the number of pixels acquired.
Our goal is for the manifold learned from masked images to resemble its full
image counterpart as closely as possible. More precisely, we show that one can
indeed accurately learn an image manifold without having to consider a large
majority of the image pixels. In doing so, we consider two masking methods that
preserve the local and global geometric structure of the manifold,
respectively. In each case, the process of finding the optimal masking pattern
can be cast as a binary integer program, which is computationally expensive but
can be approximated by a fast greedy algorithm. Numerical experiments show that
the relevant manifold structure is preserved through the data-dependent masking
process, even for modest mask sizes.
|
Finite-time Analysis for the Knowledge-Gradient Policy | We consider sequential decision problems in which we adaptively choose one of
finitely many alternatives and observe a stochastic reward. We offer a new
perspective of interpreting Bayesian ranking and selection problems as adaptive
stochastic multi-set maximization problems and derive the first finite-time
bound of the knowledge-gradient policy for adaptive submodular objective
functions. In addition, we introduce the concept of prior-optimality and
provide another insight into the performance of the knowledge gradient policy
based on the submodular assumption on the value of information. We demonstrate
submodularity for the two-alternative case and provide other conditions for
more general problems, bringing out the issue and importance of submodularity
in learning problems. Empirical experiments are conducted to further illustrate
the finite time behavior of the knowledge gradient policy.
|
Unsupervised Learning of Predictors from Unpaired Input-Output Samples | Unsupervised learning is the most challenging problem in machine learning and
especially in deep learning. Among many scenarios, we study an unsupervised
learning problem of high economic value --- learning to predict without costly
pairing of input data and corresponding labels. Part of the difficulty in this
problem is a lack of solid evaluation measures. In this paper, we take a
practical approach to grounding unsupervised learning by using the same success
criterion as for supervised learning in prediction tasks but we do not require
the presence of paired input-output training data. In particular, we propose an
objective function that aims to make the predicted outputs fit well the
structure of the output while preserving the correlation between the input and
the predicted output. We experiment with a synthetic structural prediction
problem and show that even with simple linear classifiers, the objective
function is already highly non-convex. We further demonstrate the nature of
this non-convex optimization problem as well as potential solutions. In
particular, we show that with regularization via a generative model, learning
with the proposed unsupervised objective function converges to an optimal
solution.
|
Progressive Neural Networks | Learning to solve complex sequences of tasks--while both leveraging transfer
and avoiding catastrophic forgetting--remains a key obstacle to achieving
human-level intelligence. The progressive networks approach represents a step
forward in this direction: they are immune to forgetting and can leverage prior
knowledge via lateral connections to previously learned features. We evaluate
this architecture extensively on a wide variety of reinforcement learning tasks
(Atari and 3D maze games), and show that it outperforms common baselines based
on pretraining and finetuning. Using a novel sensitivity measure, we
demonstrate that transfer occurs at both low-level sensory and high-level
control layers of the learned policy.
|
Strategic Attentive Writer for Learning Macro-Actions | We present a novel deep recurrent neural network architecture that learns to
build implicit plans in an end-to-end manner by purely interacting with an
environment in reinforcement learning setting. The network builds an internal
plan, which is continuously updated upon observation of the next input from the
environment. It can also partition this internal representation into contiguous
sub- sequences by learning for how long the plan can be committed to - i.e.
followed without re-planing. Combining these properties, the proposed model,
dubbed STRategic Attentive Writer (STRAW) can learn high-level, temporally
abstracted macro- actions of varying lengths that are solely learnt from data
without any prior information. These macro-actions enable both structured
exploration and economic computation. We experimentally demonstrate that STRAW
delivers strong improvements on several ATARI games by employing temporally
extended planning strategies (e.g. Ms. Pacman and Frostbite). It is at the same
time a general algorithm that can be applied on any sequence data. To that end,
we also show that when trained on text prediction task, STRAW naturally
predicts frequent n-grams (instead of macro-actions), demonstrating the
generality of the approach.
|
Bolt-on Differential Privacy for Scalable Stochastic Gradient
Descent-based Analytics | While significant progress has been made separately on analytics systems for
scalable stochastic gradient descent (SGD) and private SGD, none of the major
scalable analytics frameworks have incorporated differentially private SGD.
There are two inter-related issues for this disconnect between research and
practice: (1) low model accuracy due to added noise to guarantee privacy, and
(2) high development and runtime overhead of the private algorithms. This paper
takes a first step to remedy this disconnect and proposes a private SGD
algorithm to address \emph{both} issues in an integrated manner. In contrast to
the white-box approach adopted by previous work, we revisit and use the
classical technique of {\em output perturbation} to devise a novel "bolt-on"
approach to private SGD. While our approach trivially addresses (2), it makes
(1) even more challenging. We address this challenge by providing a novel
analysis of the $L_2$-sensitivity of SGD, which allows, under the same privacy
guarantees, better convergence of SGD when only a constant number of passes can
be made over the data. We integrate our algorithm, as well as other
state-of-the-art differentially private SGD, into Bismarck, a popular scalable
SGD-based analytics system on top of an RDBMS. Extensive experiments show that
our algorithm can be easily integrated, incurs virtually no overhead, scales
well, and most importantly, yields substantially better (up to 4X) test
accuracy than the state-of-the-art algorithms on many real datasets.
|
Multi-Modal Hybrid Deep Neural Network for Speech Enhancement | Deep Neural Networks (DNN) have been successful in en- hancing noisy speech
signals. Enhancement is achieved by learning a nonlinear mapping function from
the features of the corrupted speech signal to that of the reference clean
speech signal. The quality of predicted features can be improved by providing
additional side channel information that is robust to noise, such as visual
cues. In this paper we propose a novel deep learning model inspired by insights
from human audio visual perception. In the proposed unified hybrid
architecture, features from a Convolution Neural Network (CNN) that processes
the visual cues and features from a fully connected DNN that processes the
audio signal are integrated using a Bidirectional Long Short-Term Memory
(BiLSTM) network. The parameters of the hybrid model are jointly learned using
backpropagation. We compare the quality of enhanced speech from the hybrid
models with those from traditional DNN and BiLSTM models.
|
Safe Exploration in Finite Markov Decision Processes with Gaussian
Processes | In classical reinforcement learning, when exploring an environment, agents
accept arbitrary short term loss for long term gain. This is infeasible for
safety critical applications, such as robotics, where even a single unsafe
action may cause system failure. In this paper, we address the problem of
safely exploring finite Markov decision processes (MDP). We define safety in
terms of an, a priori unknown, safety constraint that depends on states and
actions. We aim to explore the MDP under this constraint, assuming that the
unknown function satisfies regularity conditions expressed via a Gaussian
process prior. We develop a novel algorithm for this task and prove that it is
able to completely explore the safely reachable part of the MDP without
violating the safety constraint. To achieve this, it cautiously explores safe
states and actions in order to gain statistical confidence about the safety of
unvisited state-action pairs from noisy observations collected while navigating
the environment. Moreover, the algorithm explicitly considers reachability when
exploring the MDP, ensuring that it does not get stuck in any state with no
safe way out. We demonstrate our method on digital terrain models for the task
of exploring an unknown map with a rover.
|
The Learning and Prediction of Application-level Traffic Data in
Cellular Networks | Traffic learning and prediction is at the heart of the evaluation of the
performance of telecommunications networks and attracts a lot of attention in
wired broadband networks. Now, benefiting from the big data in cellular
networks, it becomes possible to make the analyses one step further into the
application level. In this paper, we firstly collect a significant amount of
application-level traffic data from cellular network operators. Afterwards,
with the aid of the traffic "big data", we make a comprehensive study over the
modeling and prediction framework of cellular network traffic. Our results
solidly demonstrate that there universally exist some traffic statistical
modeling characteristics, including ALPHA-stable modeled property in the
temporal domain and the sparsity in the spatial domain. Meanwhile, the results
also demonstrate the distinctions originated from the uniqueness of different
service types of applications. Furthermore, we propose a new traffic prediction
framework to encompass and explore these aforementioned characteristics and
then develop a dictionary learning-based alternating direction method to solve
it. Besides, we validate the prediction accuracy improvement and the robustness
of the proposed framework through extensive simulation results.
|
A Powerful Generative Model Using Random Weights for the Deep Image
Representation | To what extent is the success of deep visualization due to the training?
Could we do deep visualization using untrained, random weight networks? To
address this issue, we explore new and powerful generative models for three
popular deep visualization tasks using untrained, random weight convolutional
neural networks. First we invert representations in feature spaces and
reconstruct images from white noise inputs. The reconstruction quality is
statistically higher than that of the same method applied on well trained
networks with the same architecture. Next we synthesize textures using scaled
correlations of representations in multiple layers and our results are almost
indistinguishable with the original natural texture and the synthesized
textures based on the trained network. Third, by recasting the content of an
image in the style of various artworks, we create artistic images with high
perceptual quality, highly competitive to the prior work of Gatys et al. on
pretrained networks. To our knowledge this is the first demonstration of image
representations using untrained deep neural networks. Our work provides a new
and fascinating tool to study the representation of deep network architecture
and sheds light on new understandings on deep visualization.
|
ASAGA: Asynchronous Parallel SAGA | We describe ASAGA, an asynchronous parallel version of the incremental
gradient algorithm SAGA that enjoys fast linear convergence rates. Through a
novel perspective, we revisit and clarify a subtle but important technical
issue present in a large fraction of the recent convergence rate proofs for
asynchronous parallel optimization algorithms, and propose a simplification of
the recently introduced "perturbed iterate" framework that resolves it. We
thereby prove that ASAGA can obtain a theoretical linear speedup on multi-core
systems even without sparsity assumptions. We present results of an
implementation on a 40-core architecture illustrating the practical speedup as
well as the hardware overhead.
|
Optimization Methods for Large-Scale Machine Learning | This paper provides a review and commentary on the past, present, and future
of numerical optimization algorithms in the context of machine learning
applications. Through case studies on text classification and the training of
deep neural networks, we discuss how optimization problems arise in machine
learning and what makes them challenging. A major theme of our study is that
large-scale machine learning represents a distinctive setting in which the
stochastic gradient (SG) method has traditionally played a central role while
conventional gradient-based nonlinear optimization techniques typically falter.
Based on this viewpoint, we present a comprehensive theory of a
straightforward, yet versatile SG algorithm, discuss its practical behavior,
and highlight opportunities for designing algorithms with improved performance.
This leads to a discussion about the next generation of optimization methods
for large-scale machine learning, including an investigation of two main
streams of research on techniques that diminish noise in the stochastic
directions and methods that make use of second-order derivative approximations.
|
Deep Learning for Music | Our goal is to be able to build a generative model from a deep neural network
architecture to try to create music that has both harmony and melody and is
passable as music composed by humans. Previous work in music generation has
mainly been focused on creating a single melody. More recent work on polyphonic
music modeling, centered around time series probability density estimation, has
met some partial success. In particular, there has been a lot of work based off
of Recurrent Neural Networks combined with Restricted Boltzmann Machines
(RNN-RBM) and other similar recurrent energy based models. Our approach,
however, is to perform end-to-end learning and generation with deep neural nets
alone.
|
Improving Variational Inference with Inverse Autoregressive Flow | The framework of normalizing flows provides a general strategy for flexible
variational inference of posteriors over latent variables. We propose a new
type of normalizing flow, inverse autoregressive flow (IAF), that, in contrast
to earlier published flows, scales well to high-dimensional latent spaces. The
proposed flow consists of a chain of invertible transformations, where each
transformation is based on an autoregressive neural network. In experiments, we
show that IAF significantly improves upon diagonal Gaussian approximate
posteriors. In addition, we demonstrate that a novel type of variational
autoencoder, coupled with IAF, is competitive with neural autoregressive models
in terms of attained log-likelihood on natural images, while allowing
significantly faster synthesis.
|
Combining multiscale features for classification of hyperspectral
images: a sequence based kernel approach | Nowadays, hyperspectral image classification widely copes with spatial
information to improve accuracy. One of the most popular way to integrate such
information is to extract hierarchical features from a multiscale segmentation.
In the classification context, the extracted features are commonly concatenated
into a long vector (also called stacked vector), on which is applied a
conventional vector-based machine learning technique (e.g. SVM with Gaussian
kernel). In this paper, we rather propose to use a sequence structured kernel:
the spectrum kernel. We show that the conventional stacked vector-based kernel
is actually a special case of this kernel. Experiments conducted on various
publicly available hyperspectral datasets illustrate the improvement of the
proposed kernel w.r.t. conventional ones using the same hierarchical spatial
features.
|
Logarithmic Time One-Against-Some | We create a new online reduction of multiclass classification to binary
classification for which training and prediction time scale logarithmically
with the number of classes. Compared to previous approaches, we obtain
substantially better statistical performance for two reasons: First, we prove a
tighter and more complete boosting theorem, and second we translate the results
more directly into an algorithm. We show that several simple techniques give
rise to an algorithm that can compete with one-against-all in both space and
predictive power while offering exponential improvements in speed when the
number of classes is large.
|
A Class of Parallel Doubly Stochastic Algorithms for Large-Scale
Learning | We consider learning problems over training sets in which both, the number of
training examples and the dimension of the feature vectors, are large. To solve
these problems we propose the random parallel stochastic algorithm (RAPSA). We
call the algorithm random parallel because it utilizes multiple parallel
processors to operate on a randomly chosen subset of blocks of the feature
vector. We call the algorithm stochastic because processors choose training
subsets uniformly at random. Algorithms that are parallel in either of these
dimensions exist, but RAPSA is the first attempt at a methodology that is
parallel in both the selection of blocks and the selection of elements of the
training set. In RAPSA, processors utilize the randomly chosen functions to
compute the stochastic gradient component associated with a randomly chosen
block. The technical contribution of this paper is to show that this minimally
coordinated algorithm converges to the optimal classifier when the training
objective is convex. Moreover, we present an accelerated version of RAPSA
(ARAPSA) that incorporates the objective function curvature information by
premultiplying the descent direction by a Hessian approximation matrix. We
further extend the results for asynchronous settings and show that if the
processors perform their updates without any coordination the algorithms are
still convergent to the optimal argument. RAPSA and its extensions are then
numerically evaluated on a linear estimation problem and a binary image
classification task using the MNIST handwritten digit dataset.
|
Automatic Pronunciation Generation by Utilizing a Semi-supervised Deep
Neural Networks | Phonemic or phonetic sub-word units are the most commonly used atomic
elements to represent speech signals in modern ASRs. However they are not the
optimal choice due to several reasons such as: large amount of effort required
to handcraft a pronunciation dictionary, pronunciation variations, human
mistakes and under-resourced dialects and languages. Here, we propose a
data-driven pronunciation estimation and acoustic modeling method which only
takes the orthographic transcription to jointly estimate a set of sub-word
units and a reliable dictionary. Experimental results show that the proposed
method which is based on semi-supervised training of a deep neural network
largely outperforms phoneme based continuous speech recognition on the TIMIT
dataset.
|
Improving Power Generation Efficiency using Deep Neural Networks | Recently there has been significant research on power generation,
distribution and transmission efficiency especially in the case of renewable
resources. The main objective is reduction of energy losses and this requires
improvements on data acquisition and analysis. In this paper we address these
concerns by using consumers' electrical smart meter readings to estimate
network loading and this information can then be used for better capacity
planning. We compare Deep Neural Network (DNN) methods with traditional methods
for load forecasting. Our results indicate that DNN methods outperform most
traditional methods. This comes at the cost of additional computational
complexity but this can be addressed with the use of cloud resources. We also
illustrate how these results can be used to better support dynamic pricing.
|
Learning Optimal Interventions | Our goal is to identify beneficial interventions from observational data. We
consider interventions that are narrowly focused (impacting few covariates) and
may be tailored to each individual or globally enacted over a population. For
applications where harmful intervention is drastically worse than proposing no
change, we propose a conservative definition of the optimal intervention.
Assuming the underlying relationship remains invariant under intervention, we
develop efficient algorithms to identify the optimal intervention policy from
limited data and provide theoretical guarantees for our approach in a Gaussian
Process setting. Although our methods assume covariates can be precisely
adjusted, they remain capable of improving outcomes in misspecified settings
where interventions incur unintentional downstream effects. Empirically, our
approach identifies good interventions in two practical applications: gene
perturbation and writing improvement.
|
Pruning Random Forests for Prediction on a Budget | We propose to prune a random forest (RF) for resource-constrained prediction.
We first construct a RF and then prune it to optimize expected feature cost &
accuracy. We pose pruning RFs as a novel 0-1 integer program with linear
constraints that encourages feature re-use. We establish total unimodularity of
the constraint set to prove that the corresponding LP relaxation solves the
original integer program. We then exploit connections to combinatorial
optimization and develop an efficient primal-dual algorithm, scalable to large
datasets. In contrast to our bottom-up approach, which benefits from good RF
initialization, conventional methods are top-down acquiring features based on
their utility value and is generally intractable, requiring heuristics.
Empirically, our pruning algorithm outperforms existing state-of-the-art
resource-constrained algorithms.
|
How many faces can be recognized? Performance extrapolation for
multi-class classification | The difficulty of multi-class classification generally increases with the
number of classes. Using data from a subset of the classes, can we predict how
well a classifier will scale with an increased number of classes? Under the
assumption that the classes are sampled exchangeably, and under the assumption
that the classifier is generative (e.g. QDA or Naive Bayes), we show that the
expected accuracy when the classifier is trained on $k$ classes is the $k-1$st
moment of a \emph{conditional accuracy distribution}, which can be estimated
from data. This provides the theoretical foundation for performance
extrapolation based on pseudolikelihood, unbiased estimation, and
high-dimensional asymptotics. We investigate the robustness of our methods to
non-generative classifiers in simulations and one optical character recognition
example.
|
Learning feed-forward one-shot learners | One-shot learning is usually tackled by using generative models or
discriminative embeddings. Discriminative methods based on deep learning, which
are very effective in other learning scenarios, are ill-suited for one-shot
learning as they need large amounts of training data. In this paper, we propose
a method to learn the parameters of a deep model in one shot. We construct the
learner as a second deep network, called a learnet, which predicts the
parameters of a pupil network from a single exemplar. In this manner we obtain
an efficient feed-forward one-shot learner, trained end-to-end by minimizing a
one-shot classification objective in a learning to learn formulation. In order
to make the construction feasible, we propose a number of factorizations of the
parameters of the pupil network. We demonstrate encouraging results by learning
characters from single exemplars in Omniglot, and by tracking visual objects
from a single initial exemplar in the Visual Object Tracking benchmark.
|
Generalized Direct Change Estimation in Ising Model Structure | We consider the problem of estimating change in the dependency structure
between two $p$-dimensional Ising models, based on respectively $n_1$ and $n_2$
samples drawn from the models. The change is assumed to be structured, e.g.,
sparse, block sparse, node-perturbed sparse, etc., such that it can be
characterized by a suitable (atomic) norm. We present and analyze a
norm-regularized estimator for directly estimating the change in structure,
without having to estimate the structures of the individual Ising models. The
estimator can work with any norm, and can be generalized to other graphical
models under mild assumptions. We show that only one set of samples, say $n_2$,
needs to satisfy the sample complexity requirement for the estimator to work,
and the estimation error decreases as $\frac{c}{\sqrt{\min(n_1,n_2)}}$, where
$c$ depends on the Gaussian width of the unit norm ball. For example, for
$\ell_1$ norm applied to $s$-sparse change, the change can be accurately
estimated with $\min(n_1,n_2)=O(s \log p)$ which is sharper than an existing
result $n_1= O(s^2 \log p)$ and $n_2 = O(n_1^2)$. Experimental results
illustrating the effectiveness of the proposed estimator are presented.
|
Unsupervised Risk Estimation Using Only Conditional Independence
Structure | We show how to estimate a model's test error from unlabeled data, on
distributions very different from the training distribution, while assuming
only that certain conditional independencies are preserved between train and
test. We do not need to assume that the optimal predictor is the same between
train and test, or that the true distribution lies in any parametric family. We
can also efficiently differentiate the error estimate to perform unsupervised
discriminative learning. Our technical tool is the method of moments, which
allows us to exploit conditional independencies in the absence of a
fully-specified model. Our framework encompasses a large family of losses
including the log and exponential loss, and extends to structured output
settings such as hidden Markov models.
|
Learning Infinite-Layer Networks: Without the Kernel Trick | Infinite--Layer Networks (ILN) have recently been proposed as an architecture
that mimics neural networks while enjoying some of the advantages of kernel
methods. ILN are networks that integrate over infinitely many nodes within a
single hidden layer. It has been demonstrated by several authors that the
problem of learning ILN can be reduced to the kernel trick, implying that
whenever a certain integral can be computed analytically they are efficiently
learnable.
In this work we give an online algorithm for ILN, which avoids the kernel
trick assumption. More generally and of independent interest, we show that
kernel methods in general can be exploited even when the kernel cannot be
efficiently computed but can only be estimated via sampling.
We provide a regret analysis for our algorithm, showing that it matches the
sample complexity of methods which have access to kernel values. Thus, our
method is the first to demonstrate that the kernel trick is not necessary as
such, and random features suffice to obtain comparable performance.
|
Increasing the Interpretability of Recurrent Neural Networks Using
Hidden Markov Models | As deep neural networks continue to revolutionize various application
domains, there is increasing interest in making these powerful models more
understandable and interpretable, and narrowing down the causes of good and bad
predictions. We focus on recurrent neural networks (RNNs), state of the art
models in speech recognition and translation. Our approach to increasing
interpretability is by combining an RNN with a hidden Markov model (HMM), a
simpler and more transparent model. We explore various combinations of RNNs and
HMMs: an HMM trained on LSTM states; a hybrid model where an HMM is trained
first, then a small LSTM is given HMM state distributions and trained to fill
in gaps in the HMM's performance; and a jointly trained hybrid model. We find
that the LSTM and HMM learn complementary information about the features in the
text.
|
ACDC: $\alpha$-Carving Decision Chain for Risk Stratification | In many healthcare settings, intuitive decision rules for risk stratification
can help effective hospital resource allocation. This paper introduces a novel
variant of decision tree algorithms that produces a chain of decisions, not a
general tree. Our algorithm, $\alpha$-Carving Decision Chain (ACDC),
sequentially carves out "pure" subsets of the majority class examples. The
resulting chain of decision rules yields a pure subset of the minority class
examples. Our approach is particularly effective in exploring large and
class-imbalanced health datasets. Moreover, ACDC provides an interactive
interpretation in conjunction with visual performance metrics such as Receiver
Operating Characteristics curve and Lift chart.
|
Conditional Image Generation with PixelCNN Decoders | This work explores conditional image generation with a new image density
model based on the PixelCNN architecture. The model can be conditioned on any
vector, including descriptive labels or tags, or latent embeddings created by
other networks. When conditioned on class labels from the ImageNet database,
the model is able to generate diverse, realistic scenes representing distinct
animals, objects, landscapes and structures. When conditioned on an embedding
produced by a convolutional network given a single image of an unseen face, it
generates a variety of new portraits of the same person with different facial
expressions, poses and lighting conditions. We also show that conditional
PixelCNN can serve as a powerful decoder in an image autoencoder. Additionally,
the gated convolutional layers in the proposed model improve the log-likelihood
of PixelCNN to match the state-of-the-art performance of PixelRNN on ImageNet,
with greatly reduced computational cost.
|
On the Expressive Power of Deep Neural Networks | We propose a new approach to the problem of neural network expressivity,
which seeks to characterize how structural properties of a neural network
family affect the functions it is able to compute. Our approach is based on an
interrelated set of measures of expressivity, unified by the novel notion of
trajectory length, which measures how the output of a network changes as the
input sweeps along a one-dimensional path. Our findings can be summarized as
follows:
(1) The complexity of the computed function grows exponentially with depth.
(2) All weights are not equal: trained networks are more sensitive to their
lower (initial) layer weights.
(3) Regularizing on trajectory length (trajectory regularization) is a
simpler alternative to batch normalization, with the same performance.
|
Exponential expressivity in deep neural networks through transient chaos | We combine Riemannian geometry with the mean field theory of high dimensional
chaos to study the nature of signal propagation in generic, deep neural
networks with random weights. Our results reveal an order-to-chaos expressivity
phase transition, with networks in the chaotic phase computing nonlinear
functions whose global curvature grows exponentially with depth but not width.
We prove this generic class of deep random functions cannot be efficiently
computed by any shallow network, going beyond prior work restricted to the
analysis of single functions. Moreover, we formalize and quantitatively
demonstrate the long conjectured idea that deep networks can disentangle highly
curved manifolds in input space into flat manifolds in hidden space. Our
theoretical analysis of the expressive power of deep networks broadly applies
to arbitrary nonlinearities, and provides a quantitative underpinning for
previously abstract notions about the geometry of deep functions.
|
Avoiding Imposters and Delinquents: Adversarial Crowdsourcing and Peer
Prediction | We consider a crowdsourcing model in which $n$ workers are asked to rate the
quality of $n$ items previously generated by other workers. An unknown set of
$\alpha n$ workers generate reliable ratings, while the remaining workers may
behave arbitrarily and possibly adversarially. The manager of the experiment
can also manually evaluate the quality of a small number of items, and wishes
to curate together almost all of the high-quality items with at most an
$\epsilon$ fraction of low-quality items. Perhaps surprisingly, we show that
this is possible with an amount of work required of the manager, and each
worker, that does not scale with $n$: the dataset can be curated with
$\tilde{O}\Big(\frac{1}{\beta\alpha^3\epsilon^4}\Big)$ ratings per worker, and
$\tilde{O}\Big(\frac{1}{\beta\epsilon^2}\Big)$ ratings by the manager, where
$\beta$ is the fraction of high-quality items. Our results extend to the more
general setting of peer prediction, including peer grading in online
classrooms.
|
Sampling Method for Fast Training of Support Vector Data Description | Support Vector Data Description (SVDD) is a popular outlier detection
technique which constructs a flexible description of the input data. SVDD
computation time is high for large training datasets which limits its use in
big-data process-monitoring applications. We propose a new iterative
sampling-based method for SVDD training. The method incrementally learns the
training data description at each iteration by computing SVDD on an independent
random sample selected with replacement from the training data set. The
experimental results indicate that the proposed method is extremely fast and
provides a good data description .
|
Model-Agnostic Interpretability of Machine Learning | Understanding why machine learning models behave the way they do empowers
both system designers and end-users in many ways: in model selection, feature
engineering, in order to trust and act upon the predictions, and in more
intuitive user interfaces. Thus, interpretability has become a vital concern in
machine learning, and work in the area of interpretable models has found
renewed interest. In some applications, such models are as accurate as
non-interpretable ones, and thus are preferred for their transparency. Even
when they are not accurate, they may still be preferred when interpretability
is of paramount importance. However, restricting machine learning to
interpretable models is often a severe limitation. In this paper we argue for
explaining machine learning predictions using model-agnostic approaches. By
treating the machine learning models as black-box functions, these approaches
provide crucial flexibility in the choice of models, explanations, and
representations, improving debugging, comparison, and interfaces for a variety
of users and models. We also outline the main challenges for such methods, and
review a recently-introduced model-agnostic explanation approach (LIME) that
addresses these challenges.
|
Proceedings First International Workshop on Hammers for Type Theories | This volume of EPTCS contains the proceedings of the First Workshop on
Hammers for Type Theories (HaTT 2016), held on 1 July 2016 as part of the
International Joint Conference on Automated Reasoning (IJCAR 2016) in Coimbra,
Portugal. The proceedings contain four regular papers, as well as abstracts of
the two invited talks by Pierre Corbineau (Verimag, France) and Aleksy Schubert
(University of Warsaw, Poland).
|
Stance Detection with Bidirectional Conditional Encoding | Stance detection is the task of classifying the attitude expressed in a text
towards a target such as Hillary Clinton to be "positive", negative" or
"neutral". Previous work has assumed that either the target is mentioned in the
text or that training data for every target is given. This paper considers the
more challenging version of this task, where targets are not always mentioned
and no training data is available for the test targets. We experiment with
conditional LSTM encoding, which builds a representation of the tweet that is
dependent on the target, and demonstrate that it outperforms encoding the tweet
and the target independently. Performance is improved further when the
conditional model is augmented with bidirectional encoding. We evaluate our
approach on the SemEval 2016 Task 6 Twitter Stance Detection corpus achieving
performance second best only to a system trained on semi-automatically labelled
tweets for the test target. When such weak supervision is added, our approach
achieves state-of-the-art results.
|
SMS Spam Filtering using Probabilistic Topic Modelling and Stacked
Denoising Autoencoder | In This paper we present a novel approach to spam filtering and demonstrate
its applicability with respect to SMS messages. Our approach requires minimum
features engineering and a small set of la- belled data samples. Features are
extracted using topic modelling based on latent Dirichlet allocation, and then
a comprehensive data model is created using a Stacked Denoising Autoencoder
(SDA). Topic modelling summarises the data providing ease of use and high
interpretability by visualising the topics using word clouds. Given that the
SMS messages can be regarded as either spam (unwanted) or ham (wanted), the SDA
is able to model the messages and accurately discriminate between the two
classes without the need for a pre-labelled training set. The results are
compared against the state-of-the-art spam detection algorithms with our
proposed approach achieving over 97% accuracy which compares favourably to the
best reported algorithms presented in the literature.
|
Learning Interpretable Musical Compositional Rules and Traces | Throughout music history, theorists have identified and documented
interpretable rules that capture the decisions of composers. This paper asks,
"Can a machine behave like a music theorist?" It presents MUS-ROVER, a
self-learning system for automatically discovering rules from symbolic music.
MUS-ROVER performs feature learning via $n$-gram models to extract
compositional rules --- statistical patterns over the resulting features. We
evaluate MUS-ROVER on Bach's (SATB) chorales, demonstrating that it can recover
known rules, as well as identify new, characteristic patterns for further
study. We discuss how the extracted rules can be used in both machine and human
composition.
|
Early Visual Concept Learning with Unsupervised Deep Learning | Automated discovery of early visual concepts from raw image data is a major
open challenge in AI research. Addressing this problem, we propose an
unsupervised approach for learning disentangled representations of the
underlying factors of variation. We draw inspiration from neuroscience, and
show how this can be achieved in an unsupervised generative model by applying
the same learning pressures as have been suggested to act in the ventral visual
stream in the brain. By enforcing redundancy reduction, encouraging statistical
independence, and exposure to data with transform continuities analogous to
those to which human infants are exposed, we obtain a variational autoencoder
(VAE) framework capable of learning disentangled factors. Our approach makes
few assumptions and works well across a wide variety of datasets. Furthermore,
our solution has useful emergent properties, such as zero-shot inference and an
intuitive understanding of "objectness".
|
Ground Truth Bias in External Cluster Validity Indices | It has been noticed that some external CVIs exhibit a preferential bias
towards a larger or smaller number of clusters which is monotonic (directly or
inversely) in the number of clusters in candidate partitions. This type of bias
is caused by the functional form of the CVI model. For example, the popular
Rand index (RI) exhibits a monotone increasing (NCinc) bias, while the Jaccard
Index (JI) index suffers from a monotone decreasing (NCdec) bias. This type of
bias has been previously recognized in the literature. In this work, we
identify a new type of bias arising from the distribution of the ground truth
(reference) partition against which candidate partitions are compared. We call
this new type of bias ground truth (GT) bias. This type of bias occurs if a
change in the reference partition causes a change in the bias status (e.g.,
NCinc, NCdec) of a CVI. For example, NCinc bias in the RI can be changed to
NCdec bias by skewing the distribution of clusters in the ground truth
partition. It is important for users to be aware of this new type of biased
behaviour, since it may affect the interpretations of CVI results. The
objective of this article is to study the empirical and theoretical
implications of GT bias. To the best of our knowledge, this is the first
extensive study of such a property for external cluster validity indices.
|
Guaranteed Non-convex Optimization: Submodular Maximization over
Continuous Domains | Submodular continuous functions are a category of (generally)
non-convex/non-concave functions with a wide spectrum of applications. We
characterize these functions and demonstrate that they can be maximized
efficiently with approximation guarantees. Specifically, i) We introduce the
weak DR property that gives a unified characterization of submodularity for all
set, integer-lattice and continuous functions; ii) for maximizing monotone
DR-submodular continuous functions under general down-closed convex
constraints, we propose a Frank-Wolfe variant with $(1-1/e)$ approximation
guarantee, and sub-linear convergence rate; iii) for maximizing general
non-monotone submodular continuous functions subject to box constraints, we
propose a DoubleGreedy algorithm with $1/3$ approximation guarantee. Submodular
continuous functions naturally find applications in various real-world
settings, including influence and revenue maximization with continuous
assignments, sensor energy management, multi-resolution data summarization,
facility location, etc. Experimental results show that the proposed algorithms
efficiently generate superior solutions compared to baseline algorithms.
|
Balancing New Against Old Information: The Role of Surprise in Learning | Surprise describes a range of phenomena from unexpected events to behavioral
responses. We propose a measure of surprise and use it for surprise-driven
learning. Our surprise measure takes into account data likelihood as well as
the degree of commitment to a belief via the entropy of the belief
distribution. We find that surprise-minimizing learning dynamically adjusts the
balance between new and old information without the need of knowledge about the
temporal statistics of the environment. We apply our framework to a dynamic
decision-making task and a maze exploration task. Our surprise minimizing
framework is suitable for learning in complex environments, even if the
environment undergoes gradual or sudden changes and could eventually provide a
framework to study the behavior of humans and animals encountering surprising
events.
|
Linear Classification of data with Support Vector Machines and
Generalized Support Vector Machines | In this paper, we study the support vector machine and introduced the notion
of generalized support vector machine for classification of data. We show that
the problem of generalized support vector machine is equivalent to the problem
of generalized variational inequality and establish various results for the
existence of solutions. Moreover, we provide various examples to support our
results.
|
Using Visual Analytics to Interpret Predictive Machine Learning Models | It is commonly believed that increasing the interpretability of a machine
learning model may decrease its predictive power. However, inspecting
input-output relationships of those models using visual analytics, while
treating them as black-box, can help to understand the reasoning behind
outcomes without sacrificing predictive quality. We identify a space of
possible solutions and provide two examples of where such techniques have been
successfully used in practice.
|
ZNNi - Maximizing the Inference Throughput of 3D Convolutional Networks
on Multi-Core CPUs and GPUs | Sliding window convolutional networks (ConvNets) have become a popular
approach to computer vision problems such as image segmentation, and object
detection and localization. Here we consider the problem of inference, the
application of a previously trained ConvNet, with emphasis on 3D images. Our
goal is to maximize throughput, defined as average number of output voxels
computed per unit time. Other things being equal, processing a larger image
tends to increase throughput, because fractionally less computation is wasted
on the borders of the image. It follows that an apparently slower algorithm may
end up having higher throughput if it can process a larger image within the
constraint of the available RAM. We introduce novel CPU and GPU primitives for
convolutional and pooling layers, which are designed to minimize memory
overhead. The primitives include convolution based on highly efficient pruned
FFTs. Our theoretical analyses and empirical tests reveal a number of
interesting findings. For some ConvNet architectures, cuDNN is outperformed by
our FFT-based GPU primitives, and these in turn can be outperformed by our CPU
primitives. The CPU manages to achieve higher throughput because of its fast
access to more RAM. A novel primitive in which the GPU accesses host RAM can
significantly increase GPU throughput. Finally, a CPU-GPU algorithm achieves
the greatest throughput of all, 10x or more than other publicly available
implementations of sliding window 3D ConvNets. All of our code has been made
available as open source project.
|
Structured Stochastic Linear Bandits | The stochastic linear bandit problem proceeds in rounds where at each round
the algorithm selects a vector from a decision set after which it receives a
noisy linear loss parameterized by an unknown vector. The goal in such a
problem is to minimize the (pseudo) regret which is the difference between the
total expected loss of the algorithm and the total expected loss of the best
fixed vector in hindsight. In this paper, we consider settings where the
unknown parameter has structure, e.g., sparse, group sparse, low-rank, which
can be captured by a norm, e.g., $L_1$, $L_{(1,2)}$, nuclear norm. We focus on
constructing confidence ellipsoids which contain the unknown parameter across
all rounds with high-probability. We show the radius of such ellipsoids depend
on the Gaussian width of sets associated with the norm capturing the structure.
Such characterization leads to tighter confidence ellipsoids and, therefore,
sharper regret bounds compared to bounds in the existing literature which are
based on the ambient dimensionality.
|
An Efficient Large-scale Semi-supervised Multi-label Classifier Capable
of Handling Missing labels | Multi-label classification has received considerable interest in recent
years. Multi-label classifiers have to address many problems including:
handling large-scale datasets with many instances and a large set of labels,
compensating missing label assignments in the training set, considering
correlations between labels, as well as exploiting unlabeled data to improve
prediction performance. To tackle datasets with a large set of labels,
embedding-based methods have been proposed which seek to represent the label
assignments in a low-dimensional space. Many state-of-the-art embedding-based
methods use a linear dimensionality reduction to represent the label
assignments in a low-dimensional space. However, by doing so, these methods
actually neglect the tail labels - labels that are infrequently assigned to
instances. We propose an embedding-based method that non-linearly embeds the
label vectors using an stochastic approach, thereby predicting the tail labels
more accurately. Moreover, the proposed method have excellent mechanisms for
handling missing labels, dealing with large-scale datasets, as well as
exploiting unlabeled data. With the best of our knowledge, our proposed method
is the first multi-label classifier that simultaneously addresses all of the
mentioned challenges. Experiments on real-world datasets show that our method
outperforms stateof-the-art multi-label classifiers by a large margin, in terms
of prediction performance, as well as training time.
|
A Comparative Analysis of classification data mining techniques :
Deriving key factors useful for predicting students performance | Students opting for Engineering as their discipline is increasing rapidly.
But due to various factors and inappropriate primary education in India,
failure rates are high. Students are unable to excel in core engineering
because of complex and mathematical subjects. Hence, they fail in such
subjects. With the help of data mining techniques, we can predict the
performance of students in terms of grades and failure in subjects. This paper
performs a comparative analysis of various classification techniques, such as
Na\"ive Bayes, LibSVM, J48, Random Forest, and JRip and tries to choose best
among these. Based on the results obtained, we found that Na\"ive Bayes is the
most accurate method in terms of students failure prediction and JRip is most
accurate in terms of students grade prediction. We also found that JRip
marginally differs from Na\"ive Bayes in terms of accuracy for students failure
prediction and gives us a set of rules from which we derive the key factors
influencing students performance. Finally, we suggest various ways to mitigate
these factors. This study is limited to Indian Education system scenarios.
However, the factors found can be helpful in other scenarios as well.
|
Interpretable Two-level Boolean Rule Learning for Classification | As a contribution to interpretable machine learning research, we develop a
novel optimization framework for learning accurate and sparse two-level Boolean
rules. We consider rules in both conjunctive normal form (AND-of-ORs) and
disjunctive normal form (OR-of-ANDs). A principled objective function is
proposed to trade classification accuracy and interpretability, where we use
Hamming loss to characterize accuracy and sparsity to characterize
interpretability. We propose efficient procedures to optimize these objectives
based on linear programming (LP) relaxation, block coordinate descent, and
alternating minimization. Experiments show that our new algorithms provide very
good tradeoffs between accuracy and interpretability.
|
Building an Interpretable Recommender via Loss-Preserving Transformation | We propose a method for building an interpretable recommender system for
personalizing online content and promotions. Historical data available for the
system consists of customer features, provided content (promotions), and user
responses. Unlike in a standard multi-class classification setting,
misclassification costs depend on both recommended actions and customers. Our
method transforms such a data set to a new set which can be used with standard
interpretable multi-class classification algorithms. The transformation has the
desirable property that minimizing the standard misclassification penalty in
this new space is equivalent to minimizing the custom cost function.
|
Statistical Parametric Speech Synthesis Using Bottleneck Representation
From Sequence Auto-encoder | In this paper, we describe a statistical parametric speech synthesis approach
with unit-level acoustic representation. In conventional deep neural network
based speech synthesis, the input text features are repeated for the entire
duration of phoneme for mapping text and speech parameters. This mapping is
learnt at the frame-level which is the de-facto acoustic representation.
However much of this computational requirement can be drastically reduced if
every unit can be represented with a fixed-dimensional representation. Using
recurrent neural network based auto-encoder, we show that it is indeed possible
to map units of varying duration to a single vector. We then use this acoustic
representation at unit-level to synthesize speech using deep neural network
based statistical parametric speech synthesis technique. Results show that the
proposed approach is able to synthesize at the same quality as the conventional
frame based approach at a highly reduced computational cost.
|
Guaranteed bounds on the Kullback-Leibler divergence of univariate
mixtures using piecewise log-sum-exp inequalities | Information-theoretic measures such as the entropy, cross-entropy and the
Kullback-Leibler divergence between two mixture models is a core primitive in
many signal processing tasks. Since the Kullback-Leibler divergence of mixtures
provably does not admit a closed-form formula, it is in practice either
estimated using costly Monte-Carlo stochastic integration, approximated, or
bounded using various techniques. We present a fast and generic method that
builds algorithmically closed-form lower and upper bounds on the entropy, the
cross-entropy and the Kullback-Leibler divergence of mixtures. We illustrate
the versatile method by reporting on our experiments for approximating the
Kullback-Leibler divergence between univariate exponential mixtures, Gaussian
mixtures, Rayleigh mixtures, and Gamma mixtures.
|
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 they see
one. We present a new approach to interactive clustering for data exploration
called TINDER, based on a particularly simple feedback mechanism, in which an
analyst can reject a given clustering and request a new one, which is chosen to
be different from the previous clustering while fitting the data well. We
formalize this interaction in a Bayesian framework as a method for prior
elicitation, in which each different clustering is produced by a prior
distribution that is modified to discourage previously rejected clusterings. We
show that TINDER successfully produces a diverse set of clusterings, each of
equivalent quality, that are much more diverse than would be obtained by
randomized restarts.
|
Tutorial on Variational Autoencoders | In just three years, Variational Autoencoders (VAEs) have emerged as one of
the most popular approaches to unsupervised learning of complicated
distributions. VAEs are appealing because they are built on top of standard
function approximators (neural networks), and can be trained with stochastic
gradient descent. VAEs have already shown promise in generating many kinds of
complicated data, including handwritten digits, faces, house numbers, CIFAR
images, physical models of scenes, segmentation, and predicting the future from
static images. This tutorial introduces the intuitions behind VAEs, explains
the mathematics behind them, and describes some empirical behavior. No prior
knowledge of variational Bayesian methods is assumed.
|
Slack and Margin Rescaling as Convex Extensions of Supermodular
Functions | Slack and margin rescaling are variants of the structured output SVM, which
is frequently applied to problems in computer vision such as image
segmentation, object localization, and learning parts based object models. They
define convex surrogates to task specific loss functions, which, when
specialized to non-additive loss functions for multi-label problems, yield
extensions to increasing set functions. We demonstrate in this paper that we
may use these concepts to define polynomial time convex extensions of arbitrary
supermodular functions, providing an analysis framework for the tightness of
these surrogates. This analysis framework shows that, while neither margin nor
slack rescaling dominate the other, known bounds on supermodular functions can
be used to derive extensions that dominate both of these, indicating possible
directions for defining novel structured output prediction surrogates. In
addition to the analysis of structured prediction loss functions, these results
imply an approach to supermodular minimization in which margin rescaling is
combined with non-polynomial time convex extensions to compute a sequence of LP
relaxations reminiscent of a cutting plane method. This approach is applied to
the problem of selecting representative exemplars from a set of images,
validating our theoretical contributions.
|
Graph based manifold regularized deep neural networks for automatic
speech recognition | Deep neural networks (DNNs) have been successfully applied to a wide variety
of acoustic modeling tasks in recent years. These include the applications of
DNNs either in a discriminative feature extraction or in a hybrid acoustic
modeling scenario. Despite the rapid progress in this area, a number of
challenges remain in training DNNs. This paper presents an effective way of
training DNNs using a manifold learning based regularization framework. In this
framework, the parameters of the network are optimized to preserve underlying
manifold based relationships between speech feature vectors while minimizing a
measure of loss between network outputs and targets. This is achieved by
incorporating manifold based locality constraints in the objective criterion of
DNNs. Empirical evidence is provided to demonstrate that training a network
with manifold constraints preserves structural compactness in the hidden layers
of the network. Manifold regularization is applied to train bottleneck DNNs for
feature extraction in hidden Markov model (HMM) based speech recognition. The
experiments in this work are conducted on the Aurora-2 spoken digits and the
Aurora-4 read news large vocabulary continuous speech recognition tasks. The
performance is measured in terms of word error rate (WER) on these tasks. It is
shown that the manifold regularized DNNs result in up to 37% reduction in WER
relative to standard DNNs.
|
Adapting ELM to Time Series Classification: A Novel Diversified Top-k
Shapelets Extraction Method | ELM (Extreme Learning Machine) is a single hidden layer feed-forward network,
where the weights between input and hidden layer are initialized randomly. ELM
is efficient due to its utilization of the analytical approach to compute
weights between hidden and output layer. However, ELM still fails to output the
semantic classification outcome. To address such limitation, in this paper, we
propose a diversified top-k shapelets transform framework, where the shapelets
are the subsequences i.e., the best representative and interpretative features
of each class. As we identified, the most challenge problems are how to extract
the best k shapelets in original candidate sets and how to automatically
determine the k value. Specifically, we first define the similar shapelets and
diversified top-k shapelets to construct diversity shapelets graph. Then, a
novel diversity graph based top-k shapelets extraction algorithm named as
\textbf{DivTopkshapelets}\ is proposed to search top-k diversified shapelets.
Finally, we propose a shapelets transformed ELM algorithm named as
\textbf{DivShapELM} to automatically determine the k value, which is further
utilized for time series classification. The experimental results over public
data sets demonstrate that the proposed approach significantly outperforms
traditional ELM algorithm in terms of effectiveness and efficiency.
|
Continuum directions for supervised dimension reduction | Dimension reduction of multivariate data supervised by auxiliary information
is considered. A series of basis for dimension reduction is obtained as
minimizers of a novel criterion. The proposed method is akin to continuum
regression, and the resulting basis is called continuum directions. With a
presence of binary supervision data, these directions continuously bridge the
principal component, mean difference and linear discriminant directions, thus
ranging from unsupervised to fully supervised dimension reduction.
High-dimensional asymptotic studies of continuum directions for binary
supervision reveal several interesting facts. The conditions under which the
sample continuum directions are inconsistent, but their classification
performance is good, are specified. While the proposed method can be directly
used for binary and multi-category classification, its generalizations to
incorporate any form of auxiliary data are also presented. The proposed method
enjoys fast computation, and the performance is better or on par with more
computer-intensive alternatives.
|
A New Training Method for Feedforward Neural Networks Based on Geometric
Contraction Property of Activation Functions | We propose a new training method for a feedforward neural network having the
activation functions with the geometric contraction property. The method
consists of constructing a new functional that is less nonlinear in comparison
with the classical functional by removing the nonlinearity of the activation
function from the output layer. We validate this new method by a series of
experiments that show an improved learning speed and better classification
error.
|
The LAMBADA dataset: Word prediction requiring a broad discourse context | We introduce LAMBADA, a dataset to evaluate the capabilities of computational
models for text understanding by means of a word prediction task. LAMBADA is a
collection of narrative passages sharing the characteristic that human subjects
are able to guess their last word if they are exposed to the whole passage, but
not if they only see the last sentence preceding the target word. To succeed on
LAMBADA, computational models cannot simply rely on local context, but must be
able to keep track of information in the broader discourse. We show that
LAMBADA exemplifies a wide range of linguistic phenomena, and that none of
several state-of-the-art language models reaches accuracy above 1% on this
novel benchmark. We thus propose LAMBADA as a challenging test set, meant to
encourage the development of new models capable of genuine understanding of
broad context in natural language text.
|
Mining Local Process Models | In this paper we describe a method to discover frequent behavioral patterns
in event logs. We express these patterns as \emph{local process models}. Local
process model mining can be positioned in-between process discovery and episode
/ sequential pattern mining. The technique presented in this paper is able to
learn behavioral patterns involving sequential composition, concurrency, choice
and loop, like in process mining. However, we do not look at start-to-end
models, which distinguishes our approach from process discovery and creates a
link to episode / sequential pattern mining. We propose an incremental
procedure for building local process models capturing frequent patterns based
on so-called process trees. We propose five quality dimensions and
corresponding metrics for local process models, given an event log. We show
monotonicity properties for some quality dimensions, enabling a speedup of
local process model discovery through pruning. We demonstrate through a real
life case study that mining local patterns allows us to get insights in
processes where regular start-to-end process discovery techniques are only able
to learn unstructured, flower-like, models.
|
Relative Natural Gradient for Learning Large Complex Models | Fisher information and natural gradient provided deep insights and powerful
tools to artificial neural networks. However related analysis becomes more and
more difficult as the learner's structure turns large and complex. This paper
makes a preliminary step towards a new direction. We extract a local component
of a large neuron system, and defines its relative Fisher information metric
that describes accurately this small component, and is invariant to the other
parts of the system. This concept is important because the geometry structure
is much simplified and it can be easily applied to guide the learning of neural
networks. We provide an analysis on a list of commonly used components, and
demonstrate how to use this concept to further improve optimization.
|
Quantifying and Reducing Stereotypes in Word Embeddings | Machine learning algorithms are optimized to model statistical properties of
the training data. If the input data reflects stereotypes and biases of the
broader society, then the output of the learning algorithm also captures these
stereotypes. In this paper, we initiate the study of gender stereotypes in {\em
word embedding}, a popular framework to represent text data. As their use
becomes increasingly common, applications can inadvertently amplify unwanted
stereotypes. We show across multiple datasets that the embeddings contain
significant gender stereotypes, especially with regard to professions. We
created a novel gender analogy task and combined it with crowdsourcing to
systematically quantify the gender bias in a given embedding. We developed an
efficient algorithm that reduces gender stereotype using just a handful of
training examples while preserving the useful geometric properties of the
embedding. We evaluated our algorithm on several metrics. While we focus on
male/female stereotypes, our framework may be applicable to other types of
embedding biases.
|
Bootstrapping with Models: Confidence Intervals for Off-Policy
Evaluation | For an autonomous agent, executing a poor policy may be costly or even
dangerous. For such agents, it is desirable to determine confidence interval
lower bounds on the performance of any given policy without executing said
policy. Current methods for exact high confidence off-policy evaluation that
use importance sampling require a substantial amount of data to achieve a tight
lower bound. Existing model-based methods only address the problem in discrete
state spaces. Since exact bounds are intractable for many domains we trade off
strict guarantees of safety for more data-efficient approximate bounds. In this
context, we propose two bootstrapping off-policy evaluation methods which use
learned MDP transition models in order to estimate lower confidence bounds on
policy performance with limited data in both continuous and discrete state
spaces. Since direct use of a model may introduce bias, we derive a theoretical
upper bound on model bias for when the model transition function is estimated
with i.i.d. trajectories. This bound broadens our understanding of the
conditions under which model-based methods have high bias. Finally, we
empirically evaluate our proposed methods and analyze the settings in which
different bootstrapping off-policy confidence interval methods succeed and
fail.
|
DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low
Bitwidth Gradients | We propose DoReFa-Net, a method to train convolutional neural networks that
have low bitwidth weights and activations using low bitwidth parameter
gradients. In particular, during backward pass, parameter gradients are
stochastically quantized to low bitwidth numbers before being propagated to
convolutional layers. As convolutions during forward/backward passes can now
operate on low bitwidth weights and activations/gradients respectively,
DoReFa-Net can use bit convolution kernels to accelerate both training and
inference. Moreover, as bit convolutions can be efficiently implemented on CPU,
FPGA, ASIC and GPU, DoReFa-Net opens the way to accelerate training of low
bitwidth neural network on these hardware. Our experiments on SVHN and ImageNet
datasets prove that DoReFa-Net can achieve comparable prediction accuracy as
32-bit counterparts. For example, a DoReFa-Net derived from AlexNet that has
1-bit weights, 2-bit activations, can be trained from scratch using 6-bit
gradients to get 46.1\% top-1 accuracy on ImageNet validation set. The
DoReFa-Net AlexNet model is released publicly.
|
CNNLab: a Novel Parallel Framework for Neural Networks using GPU and
FPGA-a Practical Study with Trade-off Analysis | Designing and implementing efficient, provably correct parallel neural
network processing is challenging. Existing high-level parallel abstractions
like MapReduce are insufficiently expressive while low-level tools like MPI and
Pthreads leave ML experts repeatedly solving the same design challenges.
However, the diversity and large-scale data size have posed a significant
challenge to construct a flexible and high-performance implementation of deep
learning neural networks. To improve the performance and maintain the
scalability, we present CNNLab, a novel deep learning framework using GPU and
FPGA-based accelerators. CNNLab provides a uniform programming model to users
so that the hardware implementation and the scheduling are invisible to the
programmers. At runtime, CNNLab leverages the trade-offs between GPU and FPGA
before offloading the tasks to the accelerators. Experimental results on the
state-of-the-art Nvidia K40 GPU and Altera DE5 FPGA board demonstrate that the
CNNLab can provide a universal framework with efficient support for diverse
applications without increasing the burden of the programmers. Moreover, we
analyze the detailed quantitative performance, throughput, power, energy, and
performance density for both approaches. Experimental results leverage the
trade-offs between GPU and FPGA and provide useful practical experiences for
the deep learning research community.
|
Online and Differentially-Private Tensor Decomposition | In this paper, we resolve many of the key algorithmic questions regarding
robustness, memory efficiency, and differential privacy of tensor
decomposition. We propose simple variants of the tensor power method which
enjoy these strong properties. We present the first guarantees for online
tensor power method which has a linear memory requirement. Moreover, we present
a noise calibrated tensor power method with efficient privacy guarantees. At
the heart of all these guarantees lies a careful perturbation analysis derived
in this paper which improves up on the existing results significantly.
|
Learning in Games: Robustness of Fast Convergence | We show that learning algorithms satisfying a $\textit{low approximate
regret}$ property experience fast convergence to approximate optimality in a
large class of repeated games. Our property, which simply requires that each
learner has small regret compared to a $(1+\epsilon)$-multiplicative
approximation to the best action in hindsight, is ubiquitous among learning
algorithms; it is satisfied even by the vanilla Hedge forecaster. Our results
improve upon recent work of Syrgkanis et al. [SALS15] in a number of ways. We
require only that players observe payoffs under other players' realized
actions, as opposed to expected payoffs. We further show that convergence
occurs with high probability, and show convergence under bandit feedback.
Finally, we improve upon the speed of convergence by a factor of $n$, the
number of players. Both the scope of settings and the class of algorithms for
which our analysis provides fast convergence are considerably broader than in
previous work.
Our framework applies to dynamic population games via a low approximate
regret property for shifting experts. Here we strengthen the results of
Lykouris et al. [LST16] in two ways: We allow players to select learning
algorithms from a larger class, which includes a minor variant of the basic
Hedge algorithm, and we increase the maximum churn in players for which
approximate optimality is achieved.
In the bandit setting we present a new algorithm which provides a "small
loss"-type bound with improved dependence on the number of actions in utility
settings, and is both simple and efficient. This result may be of independent
interest.
|
An Empirical Comparison of Sampling Quality Metrics: A Case Study for
Bayesian Nonnegative Matrix Factorization | In this work, we empirically explore the question: how can we assess the
quality of samples from some target distribution? We assume that the samples
are provided by some valid Monte Carlo procedure, so we are guaranteed that the
collection of samples will asymptotically approximate the true distribution.
Most current evaluation approaches focus on two questions: (1) Has the chain
mixed, that is, is it sampling from the distribution? and (2) How independent
are the samples (as MCMC procedures produce correlated samples)? Focusing on
the case of Bayesian nonnegative matrix factorization, we empirically evaluate
standard metrics of sampler quality as well as propose new metrics to capture
aspects that these measures fail to expose. The aspect of sampling that is of
particular interest to us is the ability (or inability) of sampling methods to
move between multiple optima in NMF problems. As a proxy, we propose and study
a number of metrics that might quantify the diversity of a set of NMF
factorizations obtained by a sampler through quantifying the coverage of the
posterior distribution. We compare the performance of a number of standard
sampling methods for NMF in terms of these new metrics.
|
Visualizing textual models with in-text and word-as-pixel highlighting | We explore two techniques which use color to make sense of statistical text
models. One method uses in-text annotations to illustrate a model's view of
particular tokens in particular documents. Another uses a high-level,
"words-as-pixels" graphic to display an entire corpus. Together, these methods
offer both zoomed-in and zoomed-out perspectives into a model's understanding
of text. We show how these interconnected methods help diagnose a classifier's
poor performance on Twitter slang, and make sense of a topic model on
historical political texts.
|
Complex Embeddings for Simple Link Prediction | In statistical relational learning, the link prediction problem is key to
automatically understand the structure of large knowledge bases. As in previous
studies, we propose to solve this problem through latent factorization.
However, here we make use of complex valued embeddings. The composition of
complex embeddings can handle a large variety of binary relations, among them
symmetric and antisymmetric relations. Compared to state-of-the-art models such
as Neural Tensor Network and Holographic Embeddings, our approach based on
complex embeddings is arguably simpler, as it only uses the Hermitian dot
product, the complex counterpart of the standard dot product between real
vectors. Our approach is scalable to large datasets as it remains linear in
both space and time, while consistently outperforming alternative approaches on
standard link prediction benchmarks.
|
A Probabilistic Generative Grammar for Semantic Parsing | Domain-general semantic parsing is a long-standing goal in natural language
processing, where the semantic parser is capable of robustly parsing sentences
from domains outside of which it was trained. Current approaches largely rely
on additional supervision from new domains in order to generalize to those
domains. We present a generative model of natural language utterances and
logical forms and demonstrate its application to semantic parsing. Our approach
relies on domain-independent supervision to generalize to new domains. We
derive and implement efficient algorithms for training, parsing, and sentence
generation. The work relies on a novel application of hierarchical Dirichlet
processes (HDPs) for structured prediction, which we also present in this
manuscript.
This manuscript is an excerpt of chapter 4 from the Ph.D. thesis of Saparov
(2022), where the model plays a central role in a larger natural language
understanding system.
This manuscript provides a new simplified and more complete presentation of
the work first introduced in Saparov, Saraswat, and Mitchell (2017). The
description and proofs of correctness of the training algorithm, parsing
algorithm, and sentence generation algorithm are much simplified in this new
presentation. We also describe the novel application of hierarchical Dirichlet
processes for structured prediction. In addition, we extend the earlier work
with a new model of word morphology, which utilizes the comprehensive
morphological data from Wiktionary.
|
FSMJ: Feature Selection with Maximum Jensen-Shannon Divergence for Text
Categorization | In this paper, we present a new wrapper feature selection approach based on
Jensen-Shannon (JS) divergence, termed feature selection with maximum
JS-divergence (FSMJ), for text categorization. Unlike most existing feature
selection approaches, the proposed FSMJ approach is based on real-valued
features which provide more information for discrimination than binary-valued
features used in conventional approaches. We show that the FSMJ is a greedy
approach and the JS-divergence monotonically increases when more features are
selected. We conduct several experiments on real-life data sets, compared with
the state-of-the-art feature selection approaches for text categorization. The
superior performance of the proposed FSMJ approach demonstrates its
effectiveness and further indicates its wide potential applications on data
mining.
|
Unanimous Prediction for 100% Precision with Application to Learning
Semantic Mappings | Can we train a system that, on any new input, either says "don't know" or
makes a prediction that is guaranteed to be correct? We answer the question in
the affirmative provided our model family is well-specified. Specifically, we
introduce the unanimity principle: only predict when all models consistent with
the training data predict the same output. We operationalize this principle for
semantic parsing, the task of mapping utterances to logical forms. We develop a
simple, efficient method that reasons over the infinite set of all consistent
models by only checking two of the models. We prove that our method obtains
100% precision even with a modest amount of training data from a possibly
adversarial distribution. Empirically, we demonstrate the effectiveness of our
approach on the standard GeoQuery dataset.
|
Contextual Weisfeiler-Lehman Graph Kernel For Malware Detection | In this paper, we propose a novel graph kernel specifically to address a
challenging problem in the field of cyber-security, namely, malware detection.
Previous research has revealed the following: (1) Graph representations of
programs are ideally suited for malware detection as they are robust against
several attacks, (2) Besides capturing topological neighbourhoods (i.e.,
structural information) from these graphs it is important to capture the
context under which the neighbourhoods are reachable to accurately detect
malicious neighbourhoods.
We observe that state-of-the-art graph kernels, such as Weisfeiler-Lehman
kernel (WLK) capture the structural information well but fail to capture
contextual information. To address this, we develop the Contextual
Weisfeiler-Lehman kernel (CWLK) which is capable of capturing both these types
of information. We show that for the malware detection problem, CWLK is more
expressive and hence more accurate than WLK while maintaining comparable
efficiency. Through our large-scale experiments with more than 50,000
real-world Android apps, we demonstrate that CWLK outperforms two
state-of-the-art graph kernels (including WLK) and three malware detection
techniques by more than 5.27% and 4.87% F-measure, respectively, while
maintaining high efficiency. This high accuracy and efficiency make CWLK
suitable for large-scale real-world malware detection.
|
Kernel-based Generative Learning in Distortion Feature Space | This paper presents a novel kernel-based generative classifier which is
defined in a distortion subspace using polynomial series expansion, named
Kernel-Distortion (KD) classifier. An iterative kernel selection algorithm is
developed to steadily improve classification performance by repeatedly removing
and adding kernels. The experimental results on character recognition
application not only show that the proposed generative classifier performs
better than many existing classifiers, but also illustrate that it has
different recognition capability compared to the state-of-the-art
discriminative classifier - deep belief network. The recognition diversity
indicates that a hybrid combination of the proposed generative classifier and
the discriminative classifier could further improve the classification
performance. Two hybrid combination methods, cascading and stacking, have been
implemented to verify the diversity and the improvement of the proposed
classifier.
|
A Novel Framework to Expedite Systematic Reviews by Automatically
Building Information Extraction Training Corpora | A systematic review identifies and collates various clinical studies and
compares data elements and results in order to provide an evidence based answer
for a particular clinical question. The process is manual and involves lot of
time. A tool to automate this process is lacking. The aim of this work is to
develop a framework using natural language processing and machine learning to
build information extraction algorithms to identify data elements in a new
primary publication, without having to go through the expensive task of manual
annotation to build gold standards for each data element type. The system is
developed in two stages. Initially, it uses information contained in existing
systematic reviews to identify the sentences from the PDF files of the included
references that contain specific data elements of interest using a modified
Jaccard similarity measure. These sentences have been treated as labeled data.A
Support Vector Machine (SVM) classifier is trained on this labeled data to
extract data elements of interests from a new article. We conducted experiments
on Cochrane Database systematic reviews related to congestive heart failure
using inclusion criteria as an example data element. The empirical results show
that the proposed system automatically identifies sentences containing the data
element of interest with a high recall (93.75%) and reasonable precision
(27.05% - which means the reviewers have to read only 3.7 sentences on
average). The empirical results suggest that the tool is retrieving valuable
information from the reference articles, even when it is time-consuming to
identify them manually. Thus we hope that the tool will be useful for automatic
data extraction from biomedical research publications. The future scope of this
work is to generalize this information framework for all types of systematic
reviews.
|
An artificial neural network to find correlation patterns in an
arbitrary number of variables | Methods to find correlation among variables are of interest to many
disciplines, including statistics, machine learning, (big) data mining and
neurosciences. Parameters that measure correlation between two variables are of
limited utility when used with multiple variables. In this work, I propose a
simple criterion to measure correlation among an arbitrary number of variables,
based on a data set. The central idea is to i) design a function of the
variables that can take different forms depending on a set of parameters, ii)
calculate the difference between a statistics associated to the function
computed on the data set and the same statistics computed on a randomised
version of the data set, called "scrambled" data set, and iii) optimise the
parameters to maximise this difference. Many such functions can be organised in
layers, which can in turn be stacked one on top of the other, forming a neural
network. The function parameters are searched with an enhanced genetic
algortihm called POET and the resulting method is tested on a cancer gene data
set. The method may have potential implications for some issues that affect the
field of neural networks, such as overfitting, the need to process huge amounts
of data for training and the presence of "adversarial examples".
|
Concrete Problems in AI Safety | Rapid progress in machine learning and artificial intelligence (AI) has
brought increasing attention to the potential impacts of AI technologies on
society. In this paper we discuss one such potential impact: the problem of
accidents in machine learning systems, defined as unintended and harmful
behavior that may emerge from poor design of real-world AI systems. We present
a list of five practical research problems related to accident risk,
categorized according to whether the problem originates from having the wrong
objective function ("avoiding side effects" and "avoiding reward hacking"), an
objective function that is too expensive to evaluate frequently ("scalable
supervision"), or undesirable behavior during the learning process ("safe
exploration" and "distributional shift"). We review previous work in these
areas as well as suggesting research directions with a focus on relevance to
cutting-edge AI systems. Finally, we consider the high-level question of how to
think most productively about the safety of forward-looking applications of AI.
|
Augmenting Supervised Neural Networks with Unsupervised Objectives for
Large-scale Image Classification | Unsupervised learning and supervised learning are key research topics in deep
learning. However, as high-capacity supervised neural networks trained with a
large amount of labels have achieved remarkable success in many computer vision
tasks, the availability of large-scale labeled images reduced the significance
of unsupervised learning. Inspired by the recent trend toward revisiting the
importance of unsupervised learning, we investigate joint supervised and
unsupervised learning in a large-scale setting by augmenting existing neural
networks with decoding pathways for reconstruction. First, we demonstrate that
the intermediate activations of pretrained large-scale classification networks
preserve almost all the information of input images except a portion of local
spatial details. Then, by end-to-end training of the entire augmented
architecture with the reconstructive objective, we show improvement of the
network performance for supervised tasks. We evaluate several variants of
autoencoders, including the recently proposed "what-where" autoencoder that
uses the encoder pooling switches, to study the importance of the architecture
design. Taking the 16-layer VGGNet trained under the ImageNet ILSVRC 2012
protocol as a strong baseline for image classification, our methods improve the
validation-set accuracy by a noticeable margin.
|
ML-based tactile sensor calibration: A universal approach | We study the responses of two tactile sensors, the fingertip sensor from the
iCub and the BioTac under different external stimuli. The question of interest
is to which degree both sensors i) allow the estimation of force exerted on the
sensor and ii) enable the recognition of differing degrees of curvature. Making
use of a force controlled linear motor affecting the tactile sensors we acquire
several high-quality data sets allowing the study of both sensors under exactly
the same conditions. We also examined the structure of the representation of
tactile stimuli in the recorded tactile sensor data using t-SNE embeddings. The
experiments show that both the iCub and the BioTac excel in different settings.
|
Question Relevance in VQA: Identifying Non-Visual And False-Premise
Questions | Visual Question Answering (VQA) is the task of answering natural-language
questions about images. We introduce the novel problem of determining the
relevance of questions to images in VQA. Current VQA models do not reason about
whether a question is even related to the given image (e.g. What is the capital
of Argentina?) or if it requires information from external resources to answer
correctly. This can break the continuity of a dialogue in human-machine
interaction. Our approaches for determining relevance are composed of two
stages. Given an image and a question, (1) we first determine whether the
question is visual or not, (2) if visual, we determine whether the question is
relevant to the given image or not. Our approaches, based on LSTM-RNNs, VQA
model uncertainty, and caption-question similarity, are able to outperform
strong baselines on both relevance tasks. We also present human studies showing
that VQA models augmented with such question relevance reasoning are perceived
as more intelligent, reasonable, and human-like.
|
On Multiplicative Integration with Recurrent Neural Networks | We introduce a general and simple structural design called Multiplicative
Integration (MI) to improve recurrent neural networks (RNNs). MI changes the
way in which information from difference sources flows and is integrated in the
computational building block of an RNN, while introducing almost no extra
parameters. The new structure can be easily embedded into many popular RNN
models, including LSTMs and GRUs. We empirically analyze its learning behaviour
and conduct evaluations on several tasks using different RNN models. Our
experimental results demonstrate that Multiplicative Integration can provide a
substantial performance boost over many of the existing RNN models.
|
Tracking Time-Vertex Propagation using Dynamic Graph Wavelets | Graph Signal Processing generalizes classical signal processing to signal or
data indexed by the vertices of a weighted graph. So far, the research efforts
have been focused on static graph signals. However numerous applications
involve graph signals evolving in time, such as spreading or propagation of
waves on a network. The analysis of this type of data requires a new set of
methods that fully takes into account the time and graph dimensions. We propose
a novel class of wavelet frames named Dynamic Graph Wavelets, whose time-vertex
evolution follows a dynamic process. We demonstrate that this set of functions
can be combined with sparsity based approaches such as compressive sensing to
reveal information on the dynamic processes occurring on a graph. Experiments
on real seismological data show the efficiency of the technique, allowing to
estimate the epicenter of earthquake events recorded by a seismic network.
|
A Stackelberg Game Perspective on the Conflict Between Machine Learning
and Data Obfuscation | Data is the new oil; this refrain is repeated extensively in the age of
internet tracking, machine learning, and data analytics. Social network
analysis, cookie-based advertising, and government surveillance are all
evidence of the use of data for commercial and national interests. Public
pressure, however, is mounting for the protection of privacy. Frameworks such
as differential privacy offer machine learning algorithms methods to guarantee
limits to information disclosure, but they are seldom implemented. Recently,
however, developers have made significant efforts to undermine tracking through
obfuscation tools that hide user characteristics in a sea of noise. These
services highlight an emerging clash between tracking and data obfuscation. In
this paper, we conceptualize this conflict through a dynamic game between users
and a machine learning algorithm that uses empirical risk minimization. First,
a machine learner declares a privacy protection level, and then users respond
by choosing their own perturbation amounts. We study the interaction between
the users and the learner using a Stackelberg game. The utility functions
quantify accuracy using expected loss and privacy in terms of the bounds of
differential privacy. In equilibrium, we find selfish users tend to cause
significant utility loss to trackers by perturbing heavily, in a phenomenon
reminiscent of public good games. Trackers, however, can improve the balance by
proactively perturbing the data themselves. While other work in this area has
studied privacy markets and mechanism design for truthful reporting of user
information, we take a different viewpoint by considering both user and learner
perturbation.
|
Scalable Semi-supervised Learning with Graph-based Kernel Machine | Acquiring labels are often costly, whereas unlabeled data are usually easy to
obtain in modern machine learning applications. Semi-supervised learning
provides a principled machine learning framework to address such situations,
and has been applied successfully in many real-word applications and
industries. Nonetheless, most of existing semi-supervised learning methods
encounter two serious limitations when applied to modern and large-scale
datasets: computational burden and memory usage demand. To this end, we present
in this paper the Graph-based semi-supervised Kernel Machine (GKM), a method
that leverages the generalization ability of kernel-based method with the
geometrical and distributive information formulated through a spectral graph
induced from data for semi-supervised learning purpose. Our proposed GKM can be
solved directly in the primal form using the Stochastic Gradient Descent method
with the ideal convergence rate $O(\frac{1}{T})$. Besides, our formulation is
suitable for a wide spectrum of important loss functions in the literature of
machine learning (e.g., Hinge, smooth Hinge, Logistic, L1, and
{\epsilon}-insensitive) and smoothness functions (i.e., $l_p(t) = |t|^p$ with
$p\ge1$). We further show that the well-known Laplacian Support Vector Machine
is a special case of our formulation. We validate our proposed method on
several benchmark datasets to demonstrate that GKM is appropriate for the
large-scale datasets since it is optimal in memory usage and yields superior
classification accuracy whilst simultaneously achieving a significant
computation speed-up in comparison with the state-of-the-art baselines.
|
Link Prediction via Matrix Completion | Inspired by practical importance of social networks, economic networks,
biological networks and so on, studies on large and complex networks have
attracted a surge of attentions in the recent years. Link prediction is a
fundamental issue to understand the mechanisms by which new links are added to
the networks. We introduce the method of robust principal component analysis
(robust PCA) into link prediction, and estimate the missing entries of the
adjacency matrix. On one hand, our algorithm is based on the sparsity and low
rank property of the matrix, on the other hand, it also performs very well when
the network is dense. This is because a relatively dense real network is also
sparse in comparison to the complete graph. According to extensive experiments
on real networks from disparate fields, when the target network is connected
and sufficiently dense, whatever it is weighted or unweighted, our method is
demonstrated to be very effective and with prediction accuracy being
considerably improved comparing with many state-of-the-art algorithms.
|
A Curriculum Learning Method for Improved Noise Robustness in Automatic
Speech Recognition | The performance of automatic speech recognition systems under noisy
environments still leaves room for improvement. Speech enhancement or feature
enhancement techniques for increasing noise robustness of these systems usually
add components to the recognition system that need careful optimization. In
this work, we propose the use of a relatively simple curriculum training
strategy called accordion annealing (ACCAN). It uses a multi-stage training
schedule where samples at signal-to-noise ratio (SNR) values as low as 0dB are
first added and samples at increasing higher SNR values are gradually added up
to an SNR value of 50dB. We also use a method called per-epoch noise mixing
(PEM) that generates noisy training samples online during training and thus
enables dynamically changing the SNR of our training data. Both the ACCAN and
the PEM methods are evaluated on a end-to-end speech recognition pipeline on
the Wall Street Journal corpus. ACCAN decreases the average word error rate
(WER) on the 20dB to -10dB SNR range by up to 31.4% when compared to a
conventional multi-condition training method.
|
A Comprehensive Study of Deep Bidirectional LSTM RNNs for Acoustic
Modeling in Speech Recognition | We present a comprehensive study of deep bidirectional long short-term memory
(LSTM) recurrent neural network (RNN) based acoustic models for automatic
speech recognition (ASR). We study the effect of size and depth and train
models of up to 8 layers. We investigate the training aspect and study
different variants of optimization methods, batching, truncated
backpropagation, different regularization techniques such as dropout and $L_2$
regularization, and different gradient clipping variants.
The major part of the experimental analysis was performed on the Quaero
corpus. Additional experiments also were performed on the Switchboard corpus.
Our best LSTM model has a relative improvement in word error rate of over 14\%
compared to our best feed-forward neural network (FFNN) baseline on the Quaero
task. On this task, we get our best result with an 8 layer bidirectional LSTM
and we show that a pretraining scheme with layer-wise construction helps for
deep LSTMs.
Finally we compare the training calculation time of many of the presented
experiments in relation with recognition performance.
All the experiments were done with RETURNN, the RWTH extensible training
framework for universal recurrent neural networks in combination with RASR, the
RWTH ASR toolkit.
|
A segmental framework for fully-unsupervised large-vocabulary speech
recognition | Zero-resource speech technology is a growing research area that aims to
develop methods for speech processing in the absence of transcriptions,
lexicons, or language modelling text. Early term discovery systems focused on
identifying isolated recurring patterns in a corpus, while more recent
full-coverage systems attempt to completely segment and cluster the audio into
word-like units---effectively performing unsupervised speech recognition. This
article presents the first attempt we are aware of to apply such a system to
large-vocabulary multi-speaker data. Our system uses a Bayesian modelling
framework with segmental word representations: each word segment is represented
as a fixed-dimensional acoustic embedding obtained by mapping the sequence of
feature frames to a single embedding vector. We compare our system on English
and Xitsonga datasets to state-of-the-art baselines, using a variety of
measures including word error rate (obtained by mapping the unsupervised output
to ground truth transcriptions). Very high word error rates are reported---in
the order of 70--80% for speaker-dependent and 80--95% for speaker-independent
systems---highlighting the difficulty of this task. Nevertheless, in terms of
cluster quality and word segmentation metrics, we show that by imposing a
consistent top-down segmentation while also using bottom-up knowledge from
detected syllable boundaries, both single-speaker and multi-speaker versions of
our system outperform a purely bottom-up single-speaker syllable-based
approach. We also show that the discovered clusters can be made less speaker-
and gender-specific by using an unsupervised autoencoder-like feature extractor
to learn better frame-level features (prior to embedding). Our system's
discovered clusters are still less pure than those of unsupervised term
discovery systems, but provide far greater coverage.
|
Towards stationary time-vertex signal processing | Graph-based methods for signal processing have shown promise for the analysis
of data exhibiting irregular structure, such as those found in social,
transportation, and sensor networks. Yet, though these systems are often
dynamic, state-of-the-art methods for signal processing on graphs ignore the
dimension of time, treating successive graph signals independently or taking a
global average. To address this shortcoming, this paper considers the
statistical analysis of time-varying graph signals. We introduce a novel
definition of joint (time-vertex) stationarity, which generalizes the classical
definition of time stationarity and the more recent definition appropriate for
graphs. Joint stationarity gives rise to a scalable Wiener optimization
framework for joint denoising, semi-supervised learning, or more generally
inversing a linear operator, that is provably optimal. Experimental results on
real weather data demonstrate that taking into account graph and time
dimensions jointly can yield significant accuracy improvements in the
reconstruction effort.
|
Ancestral Causal Inference | Constraint-based causal discovery from limited data is a notoriously
difficult challenge due to the many borderline independence test decisions.
Several approaches to improve the reliability of the predictions by exploiting
redundancy in the independence information have been proposed recently. Though
promising, existing approaches can still be greatly improved in terms of
accuracy and scalability. We present a novel method that reduces the
combinatorial explosion of the search space by using a more coarse-grained
representation of causal information, drastically reducing computation time.
Additionally, we propose a method to score causal predictions based on their
confidence. Crucially, our implementation also allows one to easily combine
observational and interventional data and to incorporate various types of
available background knowledge. We prove soundness and asymptotic consistency
of our method and demonstrate that it can outperform the state-of-the-art on
synthetic data, achieving a speedup of several orders of magnitude. We
illustrate its practical feasibility by applying it on a challenging protein
data set.
|
Toward Interpretable Topic Discovery via Anchored Correlation
Explanation | Many predictive tasks, such as diagnosing a patient based on their medical
chart, are ultimately defined by the decisions of human experts. Unfortunately,
encoding experts' knowledge is often time consuming and expensive. We propose a
simple way to use fuzzy and informal knowledge from experts to guide discovery
of interpretable latent topics in text. The underlying intuition of our
approach is that latent factors should be informative about both correlations
in the data and a set of relevance variables specified by an expert.
Mathematically, this approach is a combination of the information bottleneck
and Total Correlation Explanation (CorEx). We give a preliminary evaluation of
Anchored CorEx, showing that it produces more coherent and interpretable topics
on two distinct corpora.
|
Finite Sample Prediction and Recovery Bounds for Ordinal Embedding | The goal of ordinal embedding is to represent items as points in a
low-dimensional Euclidean space given a set of constraints in the form of
distance comparisons like "item $i$ is closer to item $j$ than item $k$".
Ordinal constraints like this often come from human judgments. To account for
errors and variation in judgments, we consider the noisy situation in which the
given constraints are independently corrupted by reversing the correct
constraint with some probability. This paper makes several new contributions to
this problem. First, we derive prediction error bounds for ordinal embedding
with noise by exploiting the fact that the rank of a distance matrix of points
in $\mathbb{R}^d$ is at most $d+2$. These bounds characterize how well a
learned embedding predicts new comparative judgments. Second, we investigate
the special case of a known noise model and study the Maximum Likelihood
estimator. Third, knowledge of the noise model enables us to relate prediction
errors to embedding accuracy. This relationship is highly non-trivial since we
show that the linear map corresponding to distance comparisons is
non-invertible, but there exists a nonlinear map that is invertible. Fourth,
two new algorithms for ordinal embedding are proposed and evaluated in
experiments.
|
Manifold Approximation by Moving Least-Squares Projection (MMLS) | In order to avoid the curse of dimensionality, frequently encountered in Big
Data analysis, there was a vast development in the field of linear and
nonlinear dimension reduction techniques in recent years. These techniques
(sometimes referred to as manifold learning) assume that the scattered input
data is lying on a lower dimensional manifold, thus the high dimensionality
problem can be overcome by learning the lower dimensionality behavior. However,
in real life applications, data is often very noisy. In this work, we propose a
method to approximate $\mathcal{M}$ a $d$-dimensional $C^{m+1}$ smooth
submanifold of $\mathbb{R}^n$ ($d \ll n$) based upon noisy scattered data
points (i.e., a data cloud). We assume that the data points are located "near"
the lower dimensional manifold and suggest a non-linear moving least-squares
projection on an approximating $d$-dimensional manifold. Under some mild
assumptions, the resulting approximant is shown to be infinitely smooth and of
high approximation order (i.e., $O(h^{m+1})$, where $h$ is the fill distance
and $m$ is the degree of the local polynomial approximation). The method
presented here assumes no analytic knowledge of the approximated manifold and
the approximation algorithm is linear in the large dimension $n$. Furthermore,
the approximating manifold can serve as a framework to perform operations
directly on the high dimensional data in a computationally efficient manner.
This way, the preparatory step of dimension reduction, which induces
distortions to the data, can be avoided altogether.
|
Visualizing Dynamics: from t-SNE to SEMI-MDPs | Deep Reinforcement Learning (DRL) is a trending field of research, showing
great promise in many challenging problems such as playing Atari, solving Go
and controlling robots. While DRL agents perform well in practice we are still
missing the tools to analayze their performance and visualize the temporal
abstractions that they learn. In this paper, we present a novel method that
automatically discovers an internal Semi Markov Decision Process (SMDP) model
in the Deep Q Network's (DQN) learned representation. We suggest a novel
visualization method that represents the SMDP model by a directed graph and
visualize it above a t-SNE map. We show how can we interpret the agent's policy
and give evidence for the hierarchical state aggregation that DQNs are learning
automatically. Our algorithm is fully automatic, does not require any domain
specific knowledge and is evaluated by a novel likelihood based evaluation
criteria.
|
Explainable Restricted Boltzmann Machines for Collaborative Filtering | Most accurate recommender systems are black-box models, hiding the reasoning
behind their recommendations. Yet explanations have been shown to increase the
user's trust in the system in addition to providing other benefits such as
scrutability, meaning the ability to verify the validity of recommendations.
This gap between accuracy and transparency or explainability has generated an
interest in automated explanation generation methods. Restricted Boltzmann
Machines (RBM) are accurate models for CF that also lack interpretability. In
this paper, we focus on RBM based collaborative filtering recommendations, and
further assume the absence of any additional data source, such as item content
or user attributes. We thus propose a new Explainable RBM technique that
computes the top-n recommendation list from items that are explainable.
Experimental results show that our method is effective in generating accurate
and explainable recommendations.
|
An Approach to Stable Gradient Descent Adaptation of Higher-Order Neural
Units | Stability evaluation of a weight-update system of higher-order neural units
(HONUs) with polynomial aggregation of neural inputs (also known as classes of
polynomial neural networks) for adaptation of both feedforward and recurrent
HONUs by a gradient descent method is introduced. An essential core of the
approach is based on spectral radius of a weight-update system, and it allows
stability monitoring and its maintenance at every adaptation step individually.
Assuring stability of the weight-update system (at every single adaptation
step) naturally results in adaptation stability of the whole neural
architecture that adapts to target data. As an aside, the used approach
highlights the fact that the weight optimization of HONU is a linear problem,
so the proposed approach can be generally extended to any neural architecture
that is linear in its adaptable parameters.
|
Adaptive and Scalable Android Malware Detection through Online Learning | It is well-known that malware constantly evolves so as to evade detection and
this causes the entire malware population to be non-stationary. Contrary to
this fact, prior works on machine learning based Android malware detection have
assumed that the distribution of the observed malware characteristics (i.e.,
features) do not change over time. In this work, we address the problem of
malware population drift and propose a novel online machine learning based
framework, named DroidOL to handle it and effectively detect malware. In order
to perform accurate detection, security-sensitive behaviors are captured from
apps in the form of inter-procedural control-flow sub-graph features using a
state-of-the-art graph kernel. In order to perform scalable detection and to
adapt to the drift and evolution in malware population, an online
passive-aggressive classifier is used.
In a large-scale comparative analysis with more than 87,000 apps, DroidOL
achieves 84.29% accuracy outperforming two state-of-the-art malware techniques
by more than 20% in their typical batch learning setting and more than 3% when
they are continuously re-trained. Our experimental findings strongly indicate
that online learning based approaches are highly suitable for real-world
malware detection.
|
Interpretable Machine Learning Models for the Digital Clock Drawing Test | The Clock Drawing Test (CDT) is a rapid, inexpensive, and popular
neuropsychological screening tool for cognitive conditions. The Digital Clock
Drawing Test (dCDT) uses novel software to analyze data from a digitizing
ballpoint pen that reports its position with considerable spatial and temporal
precision, making possible the analysis of both the drawing process and final
product. We developed methodology to analyze pen stroke data from these
drawings, and computed a large collection of features which were then analyzed
with a variety of machine learning techniques. The resulting scoring systems
were designed to be more accurate than the systems currently used by
clinicians, but just as interpretable and easy to use. The systems also allow
us to quantify the tradeoff between accuracy and interpretability. We created
automated versions of the CDT scoring systems currently used by clinicians,
allowing us to benchmark our models, which indicated that our machine learning
models substantially outperformed the existing scoring systems.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.