title
stringlengths 7
246
| abstract
stringlengths 6
3.31k
|
|---|---|
Multiscale Residual Mixture of PCA: Dynamic Dictionaries for Optimal
Basis Learning | In this paper we are interested in the problem of learning an over-complete
basis and a methodology such that the reconstruction or inverse problem does
not need optimization. We analyze the optimality of the presented approaches,
their link to popular already known techniques s.a. Artificial Neural
Networks,k-means or Oja's learning rule. Finally, we will see that one approach
to reach the optimal dictionary is a factorial and hierarchical approach. The
derived approach lead to a formulation of a Deep Oja Network. We present
results on different tasks and present the resulting very efficient learning
algorithm which brings a new vision on the training of deep nets. Finally, the
theoretical work shows that deep frameworks are one way to efficiently have
over-complete (combinatorially large) dictionary yet allowing easy
reconstruction. We thus present the Deep Residual Oja Network (DRON). We
demonstrate that a recursive deep approach working on the residuals allow
exponential decrease of the error w.r.t. the depth.
|
Linear Time Complexity Deep Fourier Scattering Network and Extension to
Nonlinear Invariants | In this paper we propose a scalable version of a state-of-the-art
deterministic time-invariant feature extraction approach based on consecutive
changes of basis and nonlinearities, namely, the scattering network. The first
focus of the paper is to extend the scattering network to allow the use of
higher order nonlinearities as well as extracting nonlinear and Fourier based
statistics leading to the required invariants of any inherently structured
input. In order to reach fast convolutions and to leverage the intrinsic
structure of wavelets, we derive our complete model in the Fourier domain. In
addition of providing fast computations, we are now able to exploit sparse
matrices due to extremely high sparsity well localized in the Fourier domain.
As a result, we are able to reach a true linear time complexity with inputs in
the Fourier domain allowing fast and energy efficient solutions to machine
learning tasks. Validation of the features and computational results will be
presented through the use of these invariant coefficients to perform
classification on audio recordings of bird songs captured in multiple different
soundscapes. In the end, the applicability of the presented solutions to deep
artificial neural networks is discussed.
|
On-line Building Energy Optimization using Deep Reinforcement Learning | Unprecedented high volumes of data are becoming available with the growth of
the advanced metering infrastructure. These are expected to benefit planning
and operation of the future power system, and to help the customers transition
from a passive to an active role. In this paper, we explore for the first time
in the smart grid context the benefits of using Deep Reinforcement Learning, a
hybrid type of methods that combines Reinforcement Learning with Deep Learning,
to perform on-line optimization of schedules for building energy management
systems. The learning procedure was explored using two methods, Deep Q-learning
and Deep Policy Gradient, both of them being extended to perform multiple
actions simultaneously. The proposed approach was validated on the large-scale
Pecan Street Inc. database. This highly-dimensional database includes
information about photovoltaic power generation, electric vehicles as well as
buildings appliances. Moreover, these on-line energy scheduling strategies
could be used to provide real-time feedback to consumers to encourage more
efficient use of electricity.
|
Recovering Latent Signals from a Mixture of Measurements using a
Gaussian Process Prior | In sensing applications, sensors cannot always measure the latent quantity of
interest at the required resolution, sometimes they can only acquire a blurred
version of it due the sensor's transfer function. To recover latent signals
when only noisy mixed measurements of the signal are available, we propose the
Gaussian process mixture of measurements (GPMM), which models the latent signal
as a Gaussian process (GP) and allows us to perform Bayesian inference on such
signal conditional to a set of noisy mixture of measurements. We describe how
to train GPMM, that is, to find the hyperparameters of the GP and the mixing
weights, and how to perform inference on the latent signal under GPMM;
additionally, we identify the solution to the underdetermined linear system
resulting from a sensing application as a particular case of GPMM. The proposed
model is validated in the recovery of three signals: a smooth synthetic signal,
a real-world heart-rate time series and a step function, where GPMM
outperformed the standard GP in terms of estimation error, uncertainty
representation and recovery of the spectral content of the latent signal.
|
Equivalence between LINE and Matrix Factorization | LINE [1], as an efficient network embedding method, has shown its
effectiveness in dealing with large-scale undirected, directed, and/or weighted
networks. Particularly, it proposes to preserve both the local structure
(represented by First-order Proximity) and global structure (represented by
Second-order Proximity) of the network. In this study, we prove that LINE with
these two proximities (LINE(1st) and LINE(2nd)) are actually factoring two
different matrices separately. Specifically, LINE(1st) is factoring a matrix M
(1), whose entries are the doubled Pointwise Mutual Information (PMI) of vertex
pairs in undirected networks, shifted by a constant. LINE(2nd) is factoring a
matrix M (2), whose entries are the PMI of vertex and context pairs in directed
networks, shifted by a constant. We hope this finding would provide a basis for
further extensions and generalizations of LINE.
|
Generalization Bounds of SGLD for Non-convex Learning: Two Theoretical
Viewpoints | Algorithm-dependent generalization error bounds are central to statistical
learning theory. A learning algorithm may use a large hypothesis space, but the
limited number of iterations controls its model capacity and generalization
error. The impacts of stochastic gradient methods on generalization error for
non-convex learning problems not only have important theoretical consequences,
but are also critical to generalization errors of deep learning.
In this paper, we study the generalization errors of Stochastic Gradient
Langevin Dynamics (SGLD) with non-convex objectives. Two theories are proposed
with non-asymptotic discrete-time analysis, using Stability and PAC-Bayesian
results respectively. The stability-based theory obtains a bound of
$O\left(\frac{1}{n}L\sqrt{\beta T_k}\right)$, where $L$ is uniform Lipschitz
parameter, $\beta$ is inverse temperature, and $T_k$ is aggregated step sizes.
For PAC-Bayesian theory, though the bound has a slower $O(1/\sqrt{n})$ rate,
the contribution of each step is shown with an exponentially decaying factor by
imposing $\ell^2$ regularization, and the uniform Lipschitz constant is also
replaced by actual norms of gradients along trajectory. Our bounds have no
implicit dependence on dimensions, norms or other capacity measures of
parameter, which elegantly characterizes the phenomenon of "Fast Training
Guarantees Generalization" in non-convex settings. This is the first
algorithm-dependent result with reasonable dependence on aggregated step sizes
for non-convex learning, and has important implications to statistical learning
aspects of stochastic gradient methods in complicated models such as deep
learning.
|
Generic Black-Box End-to-End Attack Against State of the Art API Call
Based Malware Classifiers | In this paper, we present a black-box attack against API call based machine
learning malware classifiers, focusing on generating adversarial sequences
combining API calls and static features (e.g., printable strings) that will be
misclassified by the classifier without affecting the malware functionality. We
show that this attack is effective against many classifiers due to the
transferability principle between RNN variants, feed forward DNNs, and
traditional machine learning classifiers such as SVM. We also implement GADGET,
a software framework to convert any malware binary to a binary undetected by
malware classifiers, using the proposed attack, without access to the malware
source code.
|
Probably approximate Bayesian computation: nonasymptotic convergence of
ABC under misspecification | Approximate Bayesian computation (ABC) is a widely used inference method in
Bayesian statistics to bypass the point-wise computation of the likelihood. In
this paper we develop theoretical bounds for the distance between the
statistics used in ABC. We show that some versions of ABC are inherently robust
to misspecification. The bounds are given in the form of oracle inequalities
for a finite sample size. The dependence on the dimension of the parameter
space and the number of statistics is made explicit. The results are shown to
be amenable to oracle inequalities in parameter space. We apply our theoretical
results to given prior distributions and data generating processes, including a
non-parametric regression model. In a second part of the paper, we propose a
sequential Monte Carlo (SMC) to sample from the pseudo-posterior, improving
upon the state of the art samplers.
|
Dynamic Layer Normalization for Adaptive Neural Acoustic Modeling in
Speech Recognition | Layer normalization is a recently introduced technique for normalizing the
activities of neurons in deep neural networks to improve the training speed and
stability. In this paper, we introduce a new layer normalization technique
called Dynamic Layer Normalization (DLN) for adaptive neural acoustic modeling
in speech recognition. By dynamically generating the scaling and shifting
parameters in layer normalization, DLN adapts neural acoustic models to the
acoustic variability arising from various factors such as speakers, channel
noises, and environments. Unlike other adaptive acoustic models, our proposed
approach does not require additional adaptation data or speaker information
such as i-vectors. Moreover, the model size is fixed as it dynamically
generates adaptation parameters. We apply our proposed DLN to deep
bidirectional LSTM acoustic models and evaluate them on two benchmark datasets
for large vocabulary ASR experiments: WSJ and TED-LIUM release 2. The
experimental results show that our DLN improves neural acoustic models in terms
of transcription accuracy by dynamically adapting to various speakers and
environments.
|
Naive Bayes Classification for Subset Selection | This article focuses on the question of learning how to automatically select
a subset of items among a bigger set. We introduce a methodology for the
inference of ensembles of discrete values, based on the Naive Bayes assumption.
Our motivation stems from practical use cases where one wishes to predict an
unordered set of (possibly interdependent) values from a set of observed
features. This problem can be considered in the context of Multi-label
Classification (MLC) where such values are seen as labels associated to
continuous or discrete features. We introduce the \nbx algorithm, an extension
of Naive Bayes classification into the multi-label domain, discuss its
properties and evaluate our approach on real-world problems.
|
Self-paced Convolutional Neural Network for Computer Aided Detection in
Medical Imaging Analysis | Tissue characterization has long been an important component of Computer
Aided Diagnosis (CAD) systems for automatic lesion detection and further
clinical planning. Motivated by the superior performance of deep learning
methods on various computer vision problems, there has been increasing work
applying deep learning to medical image analysis. However, the development of a
robust and reliable deep learning model for computer-aided diagnosis is still
highly challenging due to the combination of the high heterogeneity in the
medical images and the relative lack of training samples. Specifically,
annotation and labeling of the medical images is much more expensive and
time-consuming than other applications and often involves manual labor from
multiple domain experts. In this work, we propose a multi-stage, self-paced
learning framework utilizing a convolutional neural network (CNN) to classify
Computed Tomography (CT) image patches. The key contribution of this approach
is that we augment the size of training samples by refining the unlabeled
instances with a self-paced learning CNN. By implementing the framework on high
performance computing servers including the NVIDIA DGX1 machine, we obtained
the experimental result, showing that the self-pace boosted network
consistently outperformed the original network even with very scarce manual
labels. The performance gain indicates that applications with limited training
samples such as medical image analysis can benefit from using the proposed
framework.
|
Learning model-based planning from scratch | Conventional wisdom holds that model-based planning is a powerful approach to
sequential decision-making. It is often very challenging in practice, however,
because while a model can be used to evaluate a plan, it does not prescribe how
to construct a plan. Here we introduce the "Imagination-based Planner", the
first model-based, sequential decision-making agent that can learn to
construct, evaluate, and execute plans. Before any action, it can perform a
variable number of imagination steps, which involve proposing an imagined
action and evaluating it with its model-based imagination. All imagined actions
and outcomes are aggregated, iteratively, into a "plan context" which
conditions future real and imagined actions. The agent can even decide how to
imagine: testing out alternative imagined actions, chaining sequences of
actions together, or building a more complex "imagination tree" by navigating
flexibly among the previously imagined states using a learned policy. And our
agent can learn to plan economically, jointly optimizing for external rewards
and computational costs associated with using its imagination. We show that our
architecture can learn to solve a challenging continuous control problem, and
also learn elaborate planning strategies in a discrete maze-solving task. Our
work opens a new direction toward learning the components of a model-based
planning system and how to use them.
|
Deformable Part-based Fully Convolutional Network for Object Detection | Existing region-based object detectors are limited to regions with fixed box
geometry to represent objects, even if those are highly non-rectangular. In
this paper we introduce DP-FCN, a deep model for object detection which
explicitly adapts to shapes of objects with deformable parts. Without
additional annotations, it learns to focus on discriminative elements and to
align them, and simultaneously brings more invariance for classification and
geometric information to refine localization. DP-FCN is composed of three main
modules: a Fully Convolutional Network to efficiently maintain spatial
resolution, a deformable part-based RoI pooling layer to optimize positions of
parts and build invariance, and a deformation-aware localization module
explicitly exploiting displacements of parts to improve accuracy of bounding
box regression. We experimentally validate our model and show significant
gains. DP-FCN achieves state-of-the-art performances of 83.1% and 80.9% on
PASCAL VOC 2007 and 2012 with VOC data only.
|
Can GAN Learn Topological Features of a Graph? | This paper is first-line research expanding GANs into graph topology
analysis. By leveraging the hierarchical connectivity structure of a graph, we
have demonstrated that generative adversarial networks (GANs) can successfully
capture topological features of any arbitrary graph, and rank edge sets by
different stages according to their contribution to topology reconstruction.
Moreover, in addition to acting as an indicator of graph reconstruction, we
find that these stages can also preserve important topological features in a
graph.
|
Imagination-Augmented Agents for Deep Reinforcement Learning | We introduce Imagination-Augmented Agents (I2As), a novel architecture for
deep reinforcement learning combining model-free and model-based aspects. In
contrast to most existing model-based reinforcement learning and planning
methods, which prescribe how a model should be used to arrive at a policy, I2As
learn to interpret predictions from a learned environment model to construct
implicit plans in arbitrary ways, by using the predictions as additional
context in deep policy networks. I2As show improved data efficiency,
performance, and robustness to model misspecification compared to several
baselines.
|
Analysis of $p$-Laplacian Regularization in Semi-Supervised Learning | We investigate a family of regression problems in a semi-supervised setting.
The task is to assign real-valued labels to a set of $n$ sample points,
provided a small training subset of $N$ labeled points. A goal of
semi-supervised learning is to take advantage of the (geometric) structure
provided by the large number of unlabeled data when assigning labels. We
consider random geometric graphs, with connection radius $\epsilon(n)$, to
represent the geometry of the data set. Functionals which model the task reward
the regularity of the estimator function and impose or reward the agreement
with the training data. Here we consider the discrete $p$-Laplacian
regularization.
We investigate asymptotic behavior when the number of unlabeled points
increases, while the number of training points remains fixed. We uncover a
delicate interplay between the regularizing nature of the functionals
considered and the nonlocality inherent to the graph constructions. We
rigorously obtain almost optimal ranges on the scaling of $\epsilon(n)$ for the
asymptotic consistency to hold. We prove that the minimizers of the discrete
functionals in random setting converge uniformly to the desired continuum
limit. Furthermore we discover that for the standard model used there is a
restrictive upper bound on how quickly $\epsilon(n)$ must converge to zero as
$n \to \infty$. We introduce a new model which is as simple as the original
model, but overcomes this restriction.
|
Worst-case vs Average-case Design for Estimation from Fixed Pairwise
Comparisons | Pairwise comparison data arises in many domains, including tournament
rankings, web search, and preference elicitation. Given noisy comparisons of a
fixed subset of pairs of items, we study the problem of estimating the
underlying comparison probabilities under the assumption of strong stochastic
transitivity (SST). We also consider the noisy sorting subclass of the SST
model. We show that when the assignment of items to the topology is arbitrary,
these permutation-based models, unlike their parametric counterparts, do not
admit consistent estimation for most comparison topologies used in practice. We
then demonstrate that consistent estimation is possible when the assignment of
items to the topology is randomized, thus establishing a dichotomy between
worst-case and average-case designs. We propose two estimators in the
average-case setting and analyze their risk, showing that it depends on the
comparison topology only through the degree sequence of the topology. The rates
achieved by these estimators are shown to be optimal for a large class of
graphs. Our results are corroborated by simulations on multiple comparison
topologies.
|
The Role of Conversation Context for Sarcasm Detection in Online
Interactions | Computational models for sarcasm detection have often relied on the content
of utterances in isolation. However, speaker's sarcastic intent is not always
obvious without additional context. Focusing on social media discussions, we
investigate two issues: (1) does modeling of conversation context help in
sarcasm detection and (2) can we understand what part of conversation context
triggered the sarcastic reply. To address the first issue, we investigate
several types of Long Short-Term Memory (LSTM) networks that can model both the
conversation context and the sarcastic response. We show that the conditional
LSTM network (Rocktaschel et al., 2015) and LSTM networks with sentence level
attention on context and response outperform the LSTM model that reads only the
response. To address the second issue, we present a qualitative analysis of
attention weights produced by the LSTM models with attention and discuss the
results compared with human performance on the task.
|
From Bach to the Beatles: The simulation of human tonal expectation
using ecologically-trained predictive models | Tonal structure is in part conveyed by statistical regularities between
musical events, and research has shown that computational models reflect tonal
structure in music by capturing these regularities in schematic constructs like
pitch histograms. Of the few studies that model the acquisition of perceptual
learning from musical data, most have employed self-organizing models that
learn a topology of static descriptions of musical contexts. Also, the stimuli
used to train these models are often symbolic rather than acoustically faithful
representations of musical material. In this work we investigate whether
sequential predictive models of musical memory (specifically, recurrent neural
networks), trained on audio from commercial CD recordings, induce tonal
knowledge in a similar manner to listeners (as shown in behavioral studies in
music perception). Our experiments indicate that various types of recurrent
neural networks produce musical expectations that clearly convey tonal
structure. Furthermore, the results imply that although implicit knowledge of
tonal structure is a necessary condition for accurate musical expectation, the
most accurate predictive models also use other cues beyond the tonal structure
of the musical context.
|
Learning Approximate Neural Estimators for Wireless Channel State
Information | Estimation is a critical component of synchronization in wireless and signal
processing systems. There is a rich body of work on estimator derivation,
optimization, and statistical characterization from analytic system models
which are used pervasively today. We explore an alternative approach to
building estimators which relies principally on approximate regression using
large datasets and large computationally efficient artificial neural network
models capable of learning non-linear function mappings which provide compact
and accurate estimates. For single carrier PSK modulation, we explore the
accuracy and computational complexity of such estimators compared with the
current gold-standard analytically derived alternatives. We compare performance
in various wireless operating conditions and consider the trade offs between
the two different classes of systems. Our results show the learned estimators
can provide improvements in areas such as short-time estimation and estimation
under non-trivial real world channel conditions such as fading or other
non-linear hardware or propagation effects.
|
Non-Asymptotic Uniform Rates of Consistency for k-NN Regression | We derive high-probability finite-sample uniform rates of consistency for
$k$-NN regression that are optimal up to logarithmic factors under mild
assumptions. We moreover show that $k$-NN regression adapts to an unknown lower
intrinsic dimension automatically. We then apply the $k$-NN regression rates to
establish new results about estimating the level sets and global maxima of a
function from noisy observations.
|
Deformable Registration through Learning of Context-Specific Metric
Aggregation | We propose a novel weakly supervised discriminative algorithm for learning
context specific registration metrics as a linear combination of conventional
similarity measures. Conventional metrics have been extensively used over the
past two decades and therefore both their strengths and limitations are known.
The challenge is to find the optimal relative weighting (or parameters) of
different metrics forming the similarity measure of the registration algorithm.
Hand-tuning these parameters would result in sub optimal solutions and quickly
become infeasible as the number of metrics increases. Furthermore, such
hand-crafted combination can only happen at global scale (entire volume) and
therefore will not be able to account for the different tissue properties. We
propose a learning algorithm for estimating these parameters locally,
conditioned to the data semantic classes. The objective function of our
formulation is a special case of non-convex function, difference of convex
function, which we optimize using the concave convex procedure. As a proof of
concept, we show the impact of our approach on three challenging datasets for
different anatomical structures and modalities.
|
Unsupervised Domain Adaptation for Robust Speech Recognition via
Variational Autoencoder-Based Data Augmentation | Domain mismatch between training and testing can lead to significant
degradation in performance in many machine learning scenarios. Unfortunately,
this is not a rare situation for automatic speech recognition deployments in
real-world applications. Research on robust speech recognition can be regarded
as trying to overcome this domain mismatch issue. In this paper, we address the
unsupervised domain adaptation problem for robust speech recognition, where
both source and target domain speech are presented, but word transcripts are
only available for the source domain speech. We present novel
augmentation-based methods that transform speech in a way that does not change
the transcripts. Specifically, we first train a variational autoencoder on both
source and target domain data (without supervision) to learn a latent
representation of speech. We then transform nuisance attributes of speech that
are irrelevant to recognition by modifying the latent representations, in order
to augment labeled training data with additional data whose distribution is
more similar to the target domain. The proposed method is evaluated on the
CHiME-4 dataset and reduces the absolute word error rate (WER) by as much as
35% compared to the non-adapted baseline.
|
Proximal Policy Optimization Algorithms | We propose a new family of policy gradient methods for reinforcement
learning, which alternate between sampling data through interaction with the
environment, and optimizing a "surrogate" objective function using stochastic
gradient ascent. Whereas standard policy gradient methods perform one gradient
update per data sample, we propose a novel objective function that enables
multiple epochs of minibatch updates. The new methods, which we call proximal
policy optimization (PPO), have some of the benefits of trust region policy
optimization (TRPO), but they are much simpler to implement, more general, and
have better sample complexity (empirically). Our experiments test PPO on a
collection of benchmark tasks, including simulated robotic locomotion and Atari
game playing, and we show that PPO outperforms other online policy gradient
methods, and overall strikes a favorable balance between sample complexity,
simplicity, and wall-time.
|
Pragmatic-Pedagogic Value Alignment | As intelligent systems gain autonomy and capability, it becomes vital to
ensure that their objectives match those of their human users; this is known as
the value-alignment problem. In robotics, value alignment is key to the design
of collaborative robots that can integrate into human workflows, successfully
inferring and adapting to their users' objectives as they go. We argue that a
meaningful solution to value alignment must combine multi-agent decision theory
with rich mathematical models of human cognition, enabling robots to tap into
people's natural collaborative capabilities. We present a solution to the
cooperative inverse reinforcement learning (CIRL) dynamic game based on
well-established cognitive models of decision making and theory of mind. The
solution captures a key reciprocity relation: the human will not plan her
actions in isolation, but rather reason pedagogically about how the robot might
learn from them; the robot, in turn, can anticipate this and interpret the
human's actions pragmatically. To our knowledge, this work constitutes the
first formal analysis of value alignment grounded in empirically validated
cognitive models.
|
Domain Adaptation by Using Causal Inference to Predict Invariant
Conditional Distributions | An important goal common to domain adaptation and causal inference is to make
accurate predictions when the distributions for the source (or training)
domain(s) and target (or test) domain(s) differ. In many cases, these different
distributions can be modeled as different contexts of a single underlying
system, in which each distribution corresponds to a different perturbation of
the system, or in causal terms, an intervention. We focus on a class of such
causal domain adaptation problems, where data for one or more source domains
are given, and the task is to predict the distribution of a certain target
variable from measurements of other variables in one or more target domains. We
propose an approach for solving these problems that exploits causal inference
and does not rely on prior knowledge of the causal graph, the type of
interventions or the intervention targets. We demonstrate our approach by
evaluating a possible implementation on simulated and real world data.
|
Breaking the Nonsmooth Barrier: A Scalable Parallel Method for Composite
Optimization | Due to their simplicity and excellent performance, parallel asynchronous
variants of stochastic gradient descent have become popular methods to solve a
wide range of large-scale optimization problems on multi-core architectures.
Yet, despite their practical success, support for nonsmooth objectives is still
lacking, making them unsuitable for many problems of interest in machine
learning, such as the Lasso, group Lasso or empirical risk minimization with
convex constraints.
In this work, we propose and analyze ProxASAGA, a fully asynchronous sparse
method inspired by SAGA, a variance reduced incremental gradient algorithm. The
proposed method is easy to implement and significantly outperforms the state of
the art on several nonsmooth, large-scale problems. We prove that our method
achieves a theoretical linear speedup with respect to the sequential version
under assumptions on the sparsity of gradients and block-separability of the
proximal term. Empirical benchmarks on a multi-core architecture illustrate
practical speedups of up to 12x on a 20-core machine.
|
Deep Layer Aggregation | Visual recognition requires rich representations that span levels from low to
high, scales from small to large, and resolutions from fine to coarse. Even
with the depth of features in a convolutional network, a layer in isolation is
not enough: compounding and aggregating these representations improves
inference of what and where. Architectural efforts are exploring many
dimensions for network backbones, designing deeper or wider architectures, but
how to best aggregate layers and blocks across a network deserves further
attention. Although skip connections have been incorporated to combine layers,
these connections have been "shallow" themselves, and only fuse by simple,
one-step operations. We augment standard architectures with deeper aggregation
to better fuse information across layers. Our deep layer aggregation structures
iteratively and hierarchically merge the feature hierarchy to make networks
with better accuracy and fewer parameters. Experiments across architectures and
tasks show that deep layer aggregation improves recognition and resolution
compared to existing branching and merging schemes. The code is at
https://github.com/ucbdrive/dla.
|
A Nonlinear Kernel Support Matrix Machine for Matrix Learning | In many problems of supervised tensor learning (STL), real world data such as
face images or MRI scans are naturally represented as matrices, which are also
called as second order tensors. Most existing classifiers based on tensor
representation, such as support tensor machine (STM) need to solve iteratively
which occupy much time and may suffer from local minima. In this paper, we
present a kernel support matrix machine (KSMM) to perform supervised learning
when data are represented as matrices. KSMM is a general framework for the
construction of matrix-based hyperplane to exploit structural information. We
analyze a unifying optimization problem for which we propose an asymptotically
convergent algorithm. Theoretical analysis for the generalization bounds is
derived based on Rademacher complexity with respect to a probability
distribution. We demonstrate the merits of the proposed method by exhaustive
experiments on both simulation study and a number of real-word datasets from a
variety of application domains.
|
Language Transfer of Audio Word2Vec: Learning Audio Segment
Representations without Target Language Data | Audio Word2Vec offers vector representations of fixed dimensionality for
variable-length audio segments using Sequence-to-sequence Autoencoder (SA).
These vector representations are shown to describe the sequential phonetic
structures of the audio segments to a good degree, with real world applications
such as query-by-example Spoken Term Detection (STD). This paper examines the
capability of language transfer of Audio Word2Vec. We train SA from one
language (source language) and use it to extract the vector representation of
the audio segments of another language (target language). We found that SA can
still catch phonetic structure from the audio segments of the target language
if the source and target languages are similar. In query-by-example STD, we
obtain the vector representations from the SA learned from a large amount of
source language data, and found them surpass the representations from naive
encoder and SA directly learned from a small amount of target language data.
The result shows that it is possible to learn Audio Word2Vec model from
high-resource languages and use it on low-resource languages. This further
expands the usability of Audio Word2Vec.
|
Single-Channel Multi-talker Speech Recognition with Permutation
Invariant Training | Although great progresses have been made in automatic speech recognition
(ASR), significant performance degradation is still observed when recognizing
multi-talker mixed speech. In this paper, we propose and evaluate several
architectures to address this problem under the assumption that only a single
channel of mixed signal is available. Our technique extends permutation
invariant training (PIT) by introducing the front-end feature separation module
with the minimum mean square error (MSE) criterion and the back-end recognition
module with the minimum cross entropy (CE) criterion. More specifically, during
training we compute the average MSE or CE over the whole utterance for each
possible utterance-level output-target assignment, pick the one with the
minimum MSE or CE, and optimize for that assignment. This strategy elegantly
solves the label permutation problem observed in the deep learning based
multi-talker mixed speech separation and recognition systems. The proposed
architectures are evaluated and compared on an artificially mixed AMI dataset
with both two- and three-talker mixed speech. The experimental results indicate
that our proposed architectures can cut the word error rate (WER) by 45.0% and
25.0% relatively against the state-of-the-art single-talker speech recognition
system across all speakers when their energies are comparable, for two- and
three-talker mixed speech, respectively. To our knowledge, this is the first
work on the multi-talker mixed speech recognition on the challenging
speaker-independent spontaneous large vocabulary continuous speech task.
|
Discretization-free Knowledge Gradient Methods for Bayesian Optimization | This paper studies Bayesian ranking and selection (R&S) problems with
correlated prior beliefs and continuous domains, i.e. Bayesian optimization
(BO). Knowledge gradient methods [Frazier et al., 2008, 2009] have been widely
studied for discrete R&S problems, which sample the one-step Bayes-optimal
point. When used over continuous domains, previous work on the knowledge
gradient [Scott et al., 2011, Wu and Frazier, 2016, Wu et al., 2017] often rely
on a discretized finite approximation. However, the discretization introduces
error and scales poorly as the dimension of domain grows. In this paper, we
develop a fast discretization-free knowledge gradient method for Bayesian
optimization. Our method is not restricted to the fully sequential setting, but
useful in all settings where knowledge gradient can be used over continuous
domains. We show how our method can be generalized to handle (i) batch of
points suggestion (parallel knowledge gradient); (ii) the setting where
derivative information is available in the optimization process
(derivative-enabled knowledge gradient). In numerical experiments, we
demonstrate that the discretization-free knowledge gradient method finds global
optima significantly faster than previous Bayesian optimization algorithms on
both synthetic test functions and real-world applications, especially when
function evaluations are noisy; and derivative-enabled knowledge gradient can
further improve the performances, even outperforming the gradient-based
optimizer such as BFGS when derivative information is available.
|
High-risk learning: acquiring new word vectors from tiny data | Distributional semantics models are known to struggle with small data. It is
generally accepted that in order to learn 'a good vector' for a word, a model
must have sufficient examples of its usage. This contradicts the fact that
humans can guess the meaning of a word from a few occurrences only. In this
paper, we show that a neural language model such as Word2Vec only necessitates
minor modifications to its standard architecture to learn new terms from tiny
data, using background knowledge from a previously learnt semantic space. We
test our model on word definitions and on a nonce task involving 2-6 sentences'
worth of context, showing a large increase in performance over state-of-the-art
models on the definitional task.
|
VoiceLoop: Voice Fitting and Synthesis via a Phonological Loop | We present a new neural text to speech (TTS) method that is able to transform
text to speech in voices that are sampled in the wild. Unlike other systems,
our solution is able to deal with unconstrained voice samples and without
requiring aligned phonemes or linguistic features. The network architecture is
simpler than those in the existing literature and is based on a novel shifting
buffer working memory. The same buffer is used for estimating the attention,
computing the output audio, and for updating the buffer itself. The input
sentence is encoded using a context-free lookup table that contains one entry
per character or phoneme. The speakers are similarly represented by a short
vector that can also be fitted to new identities, even with only a few samples.
Variability in the generated speech is achieved by priming the buffer prior to
generating the audio. Experimental results on several datasets demonstrate
convincing capabilities, making TTS accessible to a wider range of
applications. In order to promote reproducibility, we release our source code
and models.
|
Decoupled classifiers for fair and efficient machine learning | When it is ethical and legal to use a sensitive attribute (such as gender or
race) in machine learning systems, the question remains how to do so. We show
that the naive application of machine learning algorithms using sensitive
features leads to an inherent tradeoff in accuracy between groups. We provide a
simple and efficient decoupling technique, that can be added on top of any
black-box machine learning algorithm, to learn different classifiers for
different groups. Transfer learning is used to mitigate the problem of having
too little data on any one group.
The method can apply to a range of fairness criteria. In particular, we
require the application designer to specify as joint loss function that makes
explicit the trade-off between fairness and accuracy. Our reduction is shown to
efficiently find the minimum loss as long as the objective has a certain
natural monotonicity property which may be of independent interest in the study
of fairness in algorithms.
|
Global Convergence of Langevin Dynamics Based Algorithms for Nonconvex
Optimization | We present a unified framework to analyze the global convergence of Langevin
dynamics based algorithms for nonconvex finite-sum optimization with $n$
component functions. At the core of our analysis is a direct analysis of the
ergodicity of the numerical approximations to Langevin dynamics, which leads to
faster convergence rates. Specifically, we show that gradient Langevin dynamics
(GLD) and stochastic gradient Langevin dynamics (SGLD) converge to the almost
minimizer within $\tilde O\big(nd/(\lambda\epsilon) \big)$ and $\tilde
O\big(d^7/(\lambda^5\epsilon^5) \big)$ stochastic gradient evaluations
respectively, where $d$ is the problem dimension, and $\lambda$ is the spectral
gap of the Markov chain generated by GLD. Both results improve upon the best
known gradient complexity results (Raginsky et al., 2017). Furthermore, for the
first time we prove the global convergence guarantee for variance reduced
stochastic gradient Langevin dynamics (SVRG-LD) to the almost minimizer within
$\tilde O\big(\sqrt{n}d^5/(\lambda^4\epsilon^{5/2})\big)$ stochastic gradient
evaluations, which outperforms the gradient complexities of GLD and SGLD in a
wide regime. Our theoretical analyses shed some light on using Langevin
dynamics based algorithms for nonconvex optimization with provable guarantees.
|
Acting Thoughts: Towards a Mobile Robotic Service Assistant for Users
with Limited Communication Skills | As autonomous service robots become more affordable and thus available also
for the general public, there is a growing need for user friendly interfaces to
control the robotic system. Currently available control modalities typically
expect users to be able to express their desire through either touch, speech or
gesture commands. While this requirement is fulfilled for the majority of
users, paralyzed users may not be able to use such systems. In this paper, we
present a novel framework, that allows these users to interact with a robotic
service assistant in a closed-loop fashion, using only thoughts. The
brain-computer interface (BCI) system is composed of several interacting
components, i.e., non-invasive neuronal signal recording and decoding,
high-level task planning, motion and manipulation planning as well as
environment perception. In various experiments, we demonstrate its
applicability and robustness in real world scenarios, considering
fetch-and-carry tasks and tasks involving human-robot interaction. As our
results demonstrate, our system is capable of adapting to frequent changes in
the environment and reliably completing given tasks within a reasonable amount
of time. Combined with high-level planning and autonomous robotic systems,
interesting new perspectives open up for non-invasive BCI-based human-robot
interactions.
|
RAIL: Risk-Averse Imitation Learning | Imitation learning algorithms learn viable policies by imitating an expert's
behavior when reward signals are not available. Generative Adversarial
Imitation Learning (GAIL) is a state-of-the-art algorithm for learning policies
when the expert's behavior is available as a fixed set of trajectories. We
evaluate in terms of the expert's cost function and observe that the
distribution of trajectory-costs is often more heavy-tailed for GAIL-agents
than the expert at a number of benchmark continuous-control tasks. Thus,
high-cost trajectories, corresponding to tail-end events of catastrophic
failure, are more likely to be encountered by the GAIL-agents than the expert.
This makes the reliability of GAIL-agents questionable when it comes to
deployment in risk-sensitive applications like robotic surgery and autonomous
driving. In this work, we aim to minimize the occurrence of tail-end events by
minimizing tail risk within the GAIL framework. We quantify tail risk by the
Conditional-Value-at-Risk (CVaR) of trajectories and develop the Risk-Averse
Imitation Learning (RAIL) algorithm. We observe that the policies learned with
RAIL show lower tail-end risk than those of vanilla GAIL. Thus the proposed
RAIL algorithm appears as a potent alternative to GAIL for improved reliability
in risk-sensitive applications.
|
Resting state fMRI functional connectivity-based classification using a
convolutional neural network architecture | Machine learning techniques have become increasingly popular in the field of
resting state fMRI (functional magnetic resonance imaging) network based
classification. However, the application of convolutional networks has been
proposed only very recently and has remained largely unexplored. In this paper
we describe a convolutional neural network architecture for functional
connectome classification called connectome-convolutional neural network
(CCNN). Our results on simulated datasets and a publicly available dataset for
amnestic mild cognitive impairment classification demonstrate that our CCNN
model can efficiently distinguish between subject groups. We also show that the
connectome-convolutional network is capable to combine information from diverse
functional connectivity metrics and that models using a combination of
different connectivity descriptors are able to outperform classifiers using
only one metric. From this flexibility follows that our proposed CCNN model can
be easily adapted to a wide range of connectome based classification or
regression tasks, by varying which connectivity descriptor combinations are
used to train the network.
|
Efficient Defenses Against Adversarial Attacks | Following the recent adoption of deep neural networks (DNN) accross a wide
range of applications, adversarial attacks against these models have proven to
be an indisputable threat. Adversarial samples are crafted with a deliberate
intention of undermining a system. In the case of DNNs, the lack of better
understanding of their working has prevented the development of efficient
defenses. In this paper, we propose a new defense method based on practical
observations which is easy to integrate into models and performs better than
state-of-the-art defenses. Our proposed solution is meant to reinforce the
structure of a DNN, making its prediction more stable and less likely to be
fooled by adversarial samples. We conduct an extensive experimental study
proving the efficiency of our method against multiple attacks, comparing it to
numerous defenses, both in white-box and black-box setups. Additionally, the
implementation of our method brings almost no overhead to the training
procedure, while maintaining the prediction performance of the original model
on clean samples.
|
Machine Teaching: A New Paradigm for Building Machine Learning Systems | The current processes for building machine learning systems require
practitioners with deep knowledge of machine learning. This significantly
limits the number of machine learning systems that can be created and has led
to a mismatch between the demand for machine learning systems and the ability
for organizations to build them. We believe that in order to meet this growing
demand for machine learning systems we must significantly increase the number
of individuals that can teach machines. We postulate that we can achieve this
goal by making the process of teaching machines easy, fast and above all,
universally accessible.
While machine learning focuses on creating new algorithms and improving the
accuracy of "learners", the machine teaching discipline focuses on the efficacy
of the "teachers". Machine teaching as a discipline is a paradigm shift that
follows and extends principles of software engineering and programming
languages. We put a strong emphasis on the teacher and the teacher's
interaction with data, as well as crucial components such as techniques and
design principles of interaction and visualization.
In this paper, we present our position regarding the discipline of machine
teaching and articulate fundamental machine teaching principles. We also
describe how, by decoupling knowledge about machine learning algorithms from
the process of teaching, we can accelerate innovation and empower millions of
new uses for machine learning models.
|
An Infinite Hidden Markov Model With Similarity-Biased Transitions | We describe a generalization of the Hierarchical Dirichlet Process Hidden
Markov Model (HDP-HMM) which is able to encode prior information that state
transitions are more likely between "nearby" states. This is accomplished by
defining a similarity function on the state space and scaling transition
probabilities by pair-wise similarities, thereby inducing correlations among
the transition distributions. We present an augmented data representation of
the model as a Markov Jump Process in which: (1) some jump attempts fail, and
(2) the probability of success is proportional to the similarity between the
source and destination states. This augmentation restores conditional conjugacy
and admits a simple Gibbs sampler. We evaluate the model and inference method
on a speaker diarization task and a "harmonic parsing" task using four-part
chorale data, as well as on several synthetic datasets, achieving favorable
comparisons to existing models.
|
A Nonlinear Dimensionality Reduction Framework Using Smooth Geodesics | Existing dimensionality reduction methods are adept at revealing hidden
underlying manifolds arising from high-dimensional data and thereby producing a
low-dimensional representation. However, the smoothness of the manifolds
produced by classic techniques over sparse and noisy data is not guaranteed. In
fact, the embedding generated using such data may distort the geometry of the
manifold and thereby produce an unfaithful embedding. Herein, we propose a
framework for nonlinear dimensionality reduction that generates a manifold in
terms of smooth geodesics that is designed to treat problems in which manifold
measurements are either sparse or corrupted by noise. Our method generates a
network structure for given high-dimensional data using a nearest neighbors
search and then produces piecewise linear shortest paths that are defined as
geodesics. Then, we fit points in each geodesic by a smoothing spline to
emphasize the smoothness. The robustness of this approach for sparse and noisy
datasets is demonstrated by the implementation of the method on synthetic and
real-world datasets.
|
An Error-Oriented Approach to Word Embedding Pre-Training | We propose a novel word embedding pre-training approach that exploits writing
errors in learners' scripts. We compare our method to previous models that tune
the embeddings based on script scores and the discrimination between correct
and corrupt word contexts in addition to the generic commonly-used embeddings
pre-trained on large corpora. The comparison is achieved by using the
aforementioned models to bootstrap a neural network that learns to predict a
holistic score for scripts. Furthermore, we investigate augmenting our model
with error corrections and monitor the impact on performance. Our results show
that our error-oriented approach outperforms other comparable ones which is
further demonstrated when training on more data. Additionally, extending the
model with corrections provides further performance gains when data sparsity is
an issue.
|
A Distributional Perspective on Reinforcement Learning | In this paper we argue for the fundamental importance of the value
distribution: the distribution of the random return received by a reinforcement
learning agent. This is in contrast to the common approach to reinforcement
learning which models the expectation of this return, or value. Although there
is an established body of literature studying the value distribution, thus far
it has always been used for a specific purpose such as implementing risk-aware
behaviour. We begin with theoretical results in both the policy evaluation and
control settings, exposing a significant distributional instability in the
latter. We then use the distributional perspective to design a new algorithm
which applies Bellman's equation to the learning of approximate value
distributions. We evaluate our algorithm using the suite of games from the
Arcade Learning Environment. We obtain both state-of-the-art results and
anecdotal evidence demonstrating the importance of the value distribution in
approximate reinforcement learning. Finally, we combine theoretical and
empirical evidence to highlight the ways in which the value distribution
impacts learning in the approximate setting.
|
A New Family of Near-metrics for Universal Similarity | We propose a family of near-metrics based on local graph diffusion to capture
similarity for a wide class of data sets. These quasi-metametrics, as their
names suggest, dispense with one or two standard axioms of metric spaces,
specifically distinguishability and symmetry, so that similarity between data
points of arbitrary type and form could be measured broadly and effectively.
The proposed near-metric family includes the forward k-step diffusion and its
reverse, typically on the graph consisting of data objects and their features.
By construction, this family of near-metrics is particularly appropriate for
categorical data, continuous data, and vector representations of images and
text extracted via deep learning approaches. We conduct extensive experiments
to evaluate the performance of this family of similarity measures and compare
and contrast with traditional measures of similarity used for each specific
application and with the ground truth when available. We show that for
structured data including categorical and continuous data, the near-metrics
corresponding to normalized forward k-step diffusion (k small) work as one of
the best performing similarity measures; for vector representations of text and
images including those extracted from deep learning, the near-metrics derived
from normalized and reverse k-step graph diffusion (k very small) exhibit
outstanding ability to distinguish data points from different classes.
|
Dictionary Learning and Sparse Coding-based Denoising for
High-Resolution Task Functional Connectivity MRI Analysis | We propose a novel denoising framework for task functional Magnetic Resonance
Imaging (tfMRI) data to delineate the high-resolution spatial pattern of the
brain functional connectivity via dictionary learning and sparse coding (DLSC).
In order to address the limitations of the unsupervised DLSC-based fMRI
studies, we utilize the prior knowledge of task paradigm in the learning step
to train a data-driven dictionary and to model the sparse representation. We
apply the proposed DLSC-based method to Human Connectome Project (HCP) motor
tfMRI dataset. Studies on the functional connectivity of cerebrocerebellar
circuits in somatomotor networks show that the DLSC-based denoising framework
can significantly improve the prominent connectivity patterns, in comparison to
the temporal non-local means (tNLM)-based denoising method as well as the case
without denoising, which is consistent and neuroscientifically meaningful
within motor area. The promising results show that the proposed method can
provide an important foundation for the high-resolution functional connectivity
analysis, and provide a better approach for fMRI preprocessing.
|
Ideological Sublations: Resolution of Dialectic in Population-based
Optimization | A population-based optimization algorithm was designed, inspired by two main
thinking modes in philosophy, both based on dialectic concept and
thesis-antithesis paradigm. They impose two different kinds of dialectics.
Idealistic and materialistic antitheses are formulated as optimization models.
Based on the models, the population is coordinated for dialectical
interactions. At the population-based context, the formulated optimization
models are reduced to a simple detection problem for each thinker (particle).
According to the assigned thinking mode to each thinker and her/his
measurements of corresponding dialectic with other candidate particles, they
deterministically decide to interact with a thinker in maximum dialectic with
their theses. The position of a thinker at maximum dialectic is known as an
available antithesis among the existing solutions. The dialectical interactions
at each ideological community are distinguished by meaningful distributions of
step-sizes for each thinking mode. In fact, the thinking modes are regarded as
exploration and exploitation elements of the proposed algorithm. The result is
a delicate balance without any requirement for adjustment of step-size
coefficients. Main parameter of the proposed algorithm is the number of
particles appointed to each thinking modes, or equivalently for each kind of
motions. An additional integer parameter is defined to boost the stability of
the final algorithm in some particular problems. The proposed algorithm is
evaluated by a testbed of 12 single-objective continuous benchmark functions.
Moreover, its performance and speed were highlighted in sparse reconstruction
and antenna selection problems, at the context of compressed sensing and
massive MIMO, respectively. The results indicate fast and efficient performance
in comparison with well-known evolutionary algorithms and dedicated
state-of-the-art algorithms.
|
Machine Learning for Structured Clinical Data | Research is a tertiary priority in the EHR, where the priorities are patient
care and billing. Because of this, the data is not standardized or formatted in
a manner easily adapted to machine learning approaches. Data may be missing for
a large variety of reasons ranging from individual input styles to differences
in clinical decision making, for example, which lab tests to issue. Few
patients are annotated at a research quality, limiting sample size and
presenting a moving gold standard. Patient progression over time is key to
understanding many diseases but many machine learning algorithms require a
snapshot, at a single time point, to create a usable vector form. Furthermore,
algorithms that produce black box results do not provide the interpretability
required for clinical adoption. This chapter discusses these challenges and
others in applying machine learning techniques to the structured EHR (i.e.
Patient Demographics, Family History, Medication Information, Vital Signs,
Laboratory Tests, Genetic Testing). It does not cover feature extraction from
additional sources such as imaging data or free text patient notes but the
approaches discussed can include features extracted from these sources.
|
Learning Transferable Architectures for Scalable Image Recognition | Developing neural network image classification models often requires
significant architecture engineering. In this paper, we study a method to learn
the model architectures directly on the dataset of interest. As this approach
is expensive when the dataset is large, we propose to search for an
architectural building block on a small dataset and then transfer the block to
a larger dataset. The key contribution of this work is the design of a new
search space (the "NASNet search space") which enables transferability. In our
experiments, we search for the best convolutional layer (or "cell") on the
CIFAR-10 dataset and then apply this cell to the ImageNet dataset by stacking
together more copies of this cell, each with their own parameters to design a
convolutional architecture, named "NASNet architecture". We also introduce a
new regularization technique called ScheduledDropPath that significantly
improves generalization in the NASNet models. On CIFAR-10 itself, NASNet
achieves 2.4% error rate, which is state-of-the-art. On ImageNet, NASNet
achieves, among the published works, state-of-the-art accuracy of 82.7% top-1
and 96.2% top-5 on ImageNet. Our model is 1.2% better in top-1 accuracy than
the best human-invented architectures while having 9 billion fewer FLOPS - a
reduction of 28% in computational demand from the previous state-of-the-art
model. When evaluated at different levels of computational cost, accuracies of
NASNets exceed those of the state-of-the-art human-designed models. For
instance, a small version of NASNet also achieves 74% top-1 accuracy, which is
3.1% better than equivalently-sized, state-of-the-art models for mobile
platforms. Finally, the learned features by NASNet used with the Faster-RCNN
framework surpass state-of-the-art by 4.0% achieving 43.1% mAP on the COCO
dataset.
|
Adversarial Variational Optimization of Non-Differentiable Simulators | Complex computer simulators are increasingly used across fields of science as
generative models tying parameters of an underlying theory to experimental
observations. Inference in this setup is often difficult, as simulators rarely
admit a tractable density or likelihood function. We introduce Adversarial
Variational Optimization (AVO), a likelihood-free inference algorithm for
fitting a non-differentiable generative model incorporating ideas from
generative adversarial networks, variational optimization and empirical Bayes.
We adapt the training procedure of generative adversarial networks by replacing
the differentiable generative network with a domain-specific simulator. We
solve the resulting non-differentiable minimax problem by minimizing
variational upper bounds of the two adversarial objectives. Effectively, the
procedure results in learning a proposal distribution over simulator
parameters, such that the JS divergence between the marginal distribution of
the synthetic data and the empirical distribution of observed data is
minimized. We evaluate and compare the method with simulators producing both
discrete and continuous data.
|
Sketched Subspace Clustering | The immense amount of daily generated and communicated data presents unique
challenges in their processing. Clustering, the grouping of data without the
presence of ground-truth labels, is an important tool for drawing inferences
from data. Subspace clustering (SC) is a relatively recent method that is able
to successfully classify nonlinearly separable data in a multitude of settings.
In spite of their high clustering accuracy, SC methods incur prohibitively high
computational complexity when processing large volumes of high-dimensional
data. Inspired by random sketching approaches for dimensionality reduction, the
present paper introduces a randomized scheme for SC, termed Sketch-SC, tailored
for large volumes of high-dimensional data. Sketch-SC accelerates the
computationally heavy parts of state-of-the-art SC approaches by compressing
the data matrix across both dimensions using random projections, thus enabling
fast and accurate large-scale SC. Performance analysis as well as extensive
numerical tests on real data corroborate the potential of Sketch-SC and its
competitive performance relative to state-of-the-art scalable SC approaches.
|
Language modeling with Neural trans-dimensional random fields | Trans-dimensional random field language models (TRF LMs) have recently been
introduced, where sentences are modeled as a collection of random fields. The
TRF approach has been shown to have the advantages of being computationally
more efficient in inference than LSTM LMs with close performance and being able
to flexibly integrating rich features. In this paper we propose neural TRFs,
beyond of the previous discrete TRFs that only use linear potentials with
discrete features. The idea is to use nonlinear potentials with continuous
features, implemented by neural networks (NNs), in the TRF framework. Neural
TRFs combine the advantages of both NNs and TRFs. The benefits of word
embedding, nonlinear feature learning and larger context modeling are inherited
from the use of NNs. At the same time, the strength of efficient inference by
avoiding expensive softmax is preserved. A number of technical contributions,
including employing deep convolutional neural networks (CNNs) to define the
potentials and incorporating the joint stochastic approximation (JSA) strategy
in the training algorithm, are developed in this work, which enable us to
successfully train neural TRF LMs. Various LMs are evaluated in terms of speech
recognition WERs by rescoring the 1000-best lists of WSJ'92 test data. The
results show that neural TRF LMs not only improve over discrete TRF LMs, but
also perform slightly better than LSTM LMs with only one fifth of parameters
and 16x faster inference efficiency.
|
Pairing an arbitrary regressor with an artificial neural network
estimating aleatoric uncertainty | We suggest a general approach to quantification of different forms of
aleatoric uncertainty in regression tasks performed by artificial neural
networks. It is based on the simultaneous training of two neural networks with
a joint loss function and a specific hyperparameter $\lambda>0$ that allows for
automatically detecting noisy and clean regions in the input space and
controlling their {\em relative contribution} to the loss and its gradients.
After the model has been trained, one of the networks performs predictions and
the other quantifies the uncertainty of these predictions by estimating the
locally averaged loss of the first one. Unlike in many classical uncertainty
quantification methods, we do not assume any a priori knowledge of the ground
truth probability distribution, neither do we, in general, maximize the
likelihood of a chosen parametric family of distributions. We analyze the
learning process and the influence of clean and noisy regions of the input
space on the loss surface, depending on $\lambda$. In particular, we show that
small values of $\lambda$ increase the relative contribution of clean regions
to the loss and its gradients. This explains why choosing small $\lambda$
allows for better predictions compared with neural networks without uncertainty
counterparts and those based on classical likelihood maximization. Finally, we
demonstrate that one can naturally form ensembles of pairs of our networks and
thus capture both aleatoric and epistemic uncertainty and avoid overfitting.
|
Joint DOA Estimation and Array Calibration Using Multiple Parametric
Dictionary Learning | This letter proposes a multiple parametric dictionary learning algorithm for
direction of arrival (DOA) estimation in presence of array gain-phase error and
mutual coupling. It jointly solves both the DOA estimation and array
imperfection problems to yield a robust DOA estimation in presence of array
imperfection errors and off-grid. In the proposed method, a multiple parametric
dictionary learning-based algorithm with an steepest-descent iteration is used
for learning the parametric perturbation matrices and the steering matrix
simultaneously. It also exploits the multiple snapshots information to enhance
the performance of DOA estimation. Simulation results show the efficiency of
the proposed algorithm when both off-grid problem and array imperfection exist.
|
Adversarial Examples for Evaluating Reading Comprehension Systems | Standard accuracy metrics indicate that reading comprehension systems are
making rapid progress, but the extent to which these systems truly understand
language remains unclear. To reward systems with real language understanding
abilities, we propose an adversarial evaluation scheme for the Stanford
Question Answering Dataset (SQuAD). Our method tests whether systems can answer
questions about paragraphs that contain adversarially inserted sentences, which
are automatically generated to distract computer systems without changing the
correct answer or misleading humans. In this adversarial setting, the accuracy
of sixteen published models drops from an average of $75\%$ F1 score to $36\%$;
when the adversary is allowed to add ungrammatical sequences of words, average
accuracy on four models decreases further to $7\%$. We hope our insights will
motivate the development of new models that understand language more precisely.
|
Prediction-Constrained Training for Semi-Supervised Mixture and Topic
Models | Supervisory signals have the potential to make low-dimensional data
representations, like those learned by mixture and topic models, more
interpretable and useful. We propose a framework for training latent variable
models that explicitly balances two goals: recovery of faithful generative
explanations of high-dimensional data, and accurate prediction of associated
semantic labels. Existing approaches fail to achieve these goals due to an
incomplete treatment of a fundamental asymmetry: the intended application is
always predicting labels from data, not data from labels. Our
prediction-constrained objective for training generative models coherently
integrates loss-based supervisory signals while enabling effective
semi-supervised learning from partially labeled data. We derive learning
algorithms for semi-supervised mixture and topic models using stochastic
gradient descent with automatic differentiation. We demonstrate improved
prediction quality compared to several previous supervised topic models,
achieving predictions competitive with high-dimensional logistic regression on
text sentiment analysis and electronic health records tasks while
simultaneously learning interpretable topics.
|
An Online Learning Approach to Buying and Selling Demand Response | We adopt the perspective of an aggregator, which seeks to coordinate its
purchase of demand reductions from a fixed group of residential electricity
customers, with its sale of the aggregate demand reduction in a two-settlement
wholesale energy market. The aggregator procures reductions in demand by
offering its customers a uniform price for reductions in consumption relative
to their predetermined baselines. Prior to its realization of the aggregate
demand reduction, the aggregator must also determine how much energy to sell
into the two-settlement energy market. In the day-ahead market, the aggregator
commits to a forward contract, which calls for the delivery of energy in the
real-time market. The underlying aggregate demand curve, which relates the
aggregate demand reduction to the aggregator's offered price, is assumed to be
affine and subject to unobservable, random shocks. Assuming that both the
parameters of the demand curve and the distribution of the random shocks are
initially unknown to the aggregator, we investigate the extent to which the
aggregator might dynamically adapt its offered prices and forward contracts to
maximize its expected profit over a time window of $T$ days. Specifically, we
design a dynamic pricing and contract offering policy that resolves the
aggregator's need to learn the unknown demand model with its desire to maximize
its cumulative expected profit over time. In particular, the proposed pricing
policy is proven to incur a regret over $T$ days that is no greater than
$O(\log(T)\sqrt{T})$.
|
Wavelet Convolutional Neural Networks for Texture Classification | Texture classification is an important and challenging problem in many image
processing applications. While convolutional neural networks (CNNs) achieved
significant successes for image classification, texture classification remains
a difficult problem since textures usually do not contain enough information
regarding the shape of object. In image processing, texture classification has
been traditionally studied well with spectral analyses which exploit repeated
structures in many textures. Since CNNs process images as-is in the spatial
domain whereas spectral analyses process images in the frequency domain, these
models have different characteristics in terms of performance. We propose a
novel CNN architecture, wavelet CNNs, which integrates a spectral analysis into
CNNs. Our insight is that the pooling layer and the convolution layer can be
viewed as a limited form of a spectral analysis. Based on this insight, we
generalize both layers to perform a spectral analysis with wavelet transform.
Wavelet CNNs allow us to utilize spectral information which is lost in
conventional CNNs but useful in texture classification. The experiments
demonstrate that our model achieves better accuracy in texture classification
than existing models. We also show that our model has significantly fewer
parameters than CNNs, making our model easier to train with less memory.
|
Learning for Multi-robot Cooperation in Partially Observable Stochastic
Environments with Macro-actions | This paper presents a data-driven approach for multi-robot coordination in
partially-observable domains based on Decentralized Partially Observable Markov
Decision Processes (Dec-POMDPs) and macro-actions (MAs). Dec-POMDPs provide a
general framework for cooperative sequential decision making under uncertainty
and MAs allow temporally extended and asynchronous action execution. To date,
most methods assume the underlying Dec-POMDP model is known a priori or a full
simulator is available during planning time. Previous methods which aim to
address these issues suffer from local optimality and sensitivity to initial
conditions. Additionally, few hardware demonstrations involving a large team of
heterogeneous robots and with long planning horizons exist. This work addresses
these gaps by proposing an iterative sampling based Expectation-Maximization
algorithm (iSEM) to learn polices using only trajectory data containing
observations, MAs, and rewards. Our experiments show the algorithm is able to
achieve better solution quality than the state-of-the-art learning-based
methods. We implement two variants of multi-robot Search and Rescue (SAR)
domains (with and without obstacles) on hardware to demonstrate the learned
policies can effectively control a team of distributed robots to cooperate in a
partially observable stochastic environment.
|
Reinforcement Learning for Bandit Neural Machine Translation with
Simulated Human Feedback | Machine translation is a natural candidate problem for reinforcement learning
from human feedback: users provide quick, dirty ratings on candidate
translations to guide a system to improve. Yet, current neural machine
translation training focuses on expensive human-generated reference
translations. We describe a reinforcement learning algorithm that improves
neural machine translation systems from simulated human feedback. Our algorithm
combines the advantage actor-critic algorithm (Mnih et al., 2016) with the
attention-based neural encoder-decoder architecture (Luong et al., 2015). This
algorithm (a) is well-designed for problems with a large action space and
delayed rewards, (b) effectively optimizes traditional corpus-level machine
translation metrics, and (c) is robust to skewed, high-variance, granular
feedback modeled after actual human behaviors.
|
Big Data Regression Using Tree Based Segmentation | Scaling regression to large datasets is a common problem in many application
areas. We propose a two step approach to scaling regression to large datasets.
Using a regression tree (CART) to segment the large dataset constitutes the
first step of this approach. The second step of this approach is to develop a
suitable regression model for each segment. Since segment sizes are not very
large, we have the ability to apply sophisticated regression techniques if
required. A nice feature of this two step approach is that it can yield models
that have good explanatory power as well as good predictive performance.
Ensemble methods like Gradient Boosted Trees can offer excellent predictive
performance but may not provide interpretable models. In the experiments
reported in this study, we found that the predictive performance of the
proposed approach matched the predictive performance of Gradient Boosted Trees.
|
Combinatorial Multi-armed Bandit with Probabilistically Triggered Arms:
A Case with Bounded Regret | In this paper, we study the combinatorial multi-armed bandit problem (CMAB)
with probabilistically triggered arms (PTAs). Under the assumption that the arm
triggering probabilities (ATPs) are positive for all arms, we prove that a
class of upper confidence bound (UCB) policies, named Combinatorial UCB with
exploration rate $\kappa$ (CUCB-$\kappa$), and Combinatorial Thompson Sampling
(CTS), which estimates the expected states of the arms via Thompson sampling,
achieve bounded regret. In addition, we prove that CUCB-$0$ and CTS incur
$O(\sqrt{T})$ gap-independent regret. These results improve the results in
previous works, which show $O(\log T)$ gap-dependent and $O(\sqrt{T\log T})$
gap-independent regrets, respectively, under no assumptions on the ATPs. Then,
we numerically evaluate the performance of CUCB-$\kappa$ and CTS in a
real-world movie recommendation problem, where the actions correspond to
recommending a set of movies, the arms correspond to the edges between the
movies and the users, and the goal is to maximize the total number of users
that are attracted by at least one movie. Our numerical results complement our
theoretical findings on bounded regret. Apart from this problem, our results
also directly apply to the online influence maximization (OIM) problem studied
in numerous prior works.
|
Character-level Intra Attention Network for Natural Language Inference | Natural language inference (NLI) is a central problem in language
understanding. End-to-end artificial neural networks have reached
state-of-the-art performance in NLI field recently.
In this paper, we propose Character-level Intra Attention Network (CIAN) for
the NLI task. In our model, we use the character-level convolutional network to
replace the standard word embedding layer, and we use the intra attention to
capture the intra-sentence semantics. The proposed CIAN model provides improved
results based on a newly published MNLI corpus.
|
Likelihood Estimation for Generative Adversarial Networks | We present a simple method for assessing the quality of generated images in
Generative Adversarial Networks (GANs). The method can be applied in any kind
of GAN without interfering with the learning procedure or affecting the
learning objective. The central idea is to define a likelihood function that
correlates with the quality of the generated images. In particular, we derive a
Gaussian likelihood function from the distribution of the embeddings (hidden
activations) of the real images in the discriminator, and based on this, define
two simple measures of how likely it is that the embeddings of generated images
are from the distribution of the embeddings of the real images. This yields a
simple measure of fitness for generated images, for all varieties of GANs.
Empirical results on CIFAR-10 demonstrate a strong correlation between the
proposed measures and the perceived quality of the generated images.
|
Exploring Outliers in Crowdsourced Ranking for QoE | Outlier detection is a crucial part of robust evaluation for crowdsourceable
assessment of Quality of Experience (QoE) and has attracted much attention in
recent years. In this paper, we propose some simple and fast algorithms for
outlier detection and robust QoE evaluation based on the nonconvex optimization
principle. Several iterative procedures are designed with or without knowing
the number of outliers in samples. Theoretical analysis is given to show that
such procedures can reach statistically good estimates under mild conditions.
Finally, experimental results with simulated and real-world crowdsourcing
datasets show that the proposed algorithms could produce similar performance to
Huber-LASSO approach in robust ranking, yet with nearly 8 or 90 times speed-up,
without or with a prior knowledge on the sparsity size of outliers,
respectively. Therefore the proposed methodology provides us a set of helpful
tools for robust QoE evaluation with crowdsourcing data.
|
Automatic breast cancer grading in lymph nodes using a deep neural
network | The progression of breast cancer can be quantified in lymph node whole-slide
images (WSIs). We describe a novel method for effectively performing
classification of whole-slide images and patient level breast cancer grading.
Our method utilises a deep neural network. The method performs classification
on small patches and uses model averaging for boosting. In the first step,
region of interest patches are determined and cropped automatically by color
thresholding and then classified by the deep neural network. The classification
results are used to determine a slide level class and for further aggregation
to predict a patient level grade. Fast processing speed of our method enables
high throughput image analysis.
|
Interpreting Classifiers through Attribute Interactions in Datasets | In this work we present the novel ASTRID method for investigating which
attribute interactions classifiers exploit when making predictions. Attribute
interactions in classification tasks mean that two or more attributes together
provide stronger evidence for a particular class label. Knowledge of such
interactions makes models more interpretable by revealing associations between
attributes. This has applications, e.g., in pharmacovigilance to identify
interactions between drugs or in bioinformatics to investigate associations
between single nucleotide polymorphisms. We also show how the found attribute
partitioning is related to a factorisation of the data generating distribution
and empirically demonstrate the utility of the proposed method.
|
Stock Prediction: a method based on extraction of news features and
recurrent neural networks | This paper proposed a method for stock prediction. In terms of feature
extraction, we extract the features of stock-related news besides stock prices.
We first select some seed words based on experience which are the symbols of
good news and bad news. Then we propose an optimization method and calculate
the positive polar of all words. After that, we construct the features of news
based on the positive polar of their words. In consideration of sequential
stock prices and continuous news effects, we propose a recurrent neural network
model to help predict stock prices. Compared to SVM classifier with price
features, we find our proposed method has an over 5% improvement on stock
prediction accuracy in experiments.
|
Share your Model instead of your Data: Privacy Preserving Mimic Learning
for Ranking | Deep neural networks have become a primary tool for solving problems in many
fields. They are also used for addressing information retrieval problems and
show strong performance in several tasks. Training these models requires large,
representative datasets and for most IR tasks, such data contains sensitive
information from users. Privacy and confidentiality concerns prevent many data
owners from sharing the data, thus today the research community can only
benefit from research on large-scale datasets in a limited manner. In this
paper, we discuss privacy preserving mimic learning, i.e., using predictions
from a privacy preserving trained model instead of labels from the original
sensitive training data as a supervision signal. We present the results of
preliminary experiments in which we apply the idea of mimic learning and
privacy preserving mimic learning for the task of document re-ranking as one of
the core IR tasks. This research is a step toward laying the ground for
enabling researchers from data-rich environments to share knowledge learned
from actual users' data, which should facilitate research collaborations.
|
A Deep Learning Approach to Digitally Stain Optical Coherence Tomography
Images of the Optic Nerve Head | Purpose: To develop a deep learning approach to digitally-stain optical
coherence tomography (OCT) images of the optic nerve head (ONH).
Methods: A horizontal B-scan was acquired through the center of the ONH using
OCT (Spectralis) for 1 eye of each of 100 subjects (40 normal & 60 glaucoma).
All images were enhanced using adaptive compensation. A custom deep learning
network was then designed and trained with the compensated images to digitally
stain (i.e. highlight) 6 tissue layers of the ONH. The accuracy of our
algorithm was assessed (against manual segmentations) using the Dice
coefficient, sensitivity, and specificity. We further studied how compensation
and the number of training images affected the performance of our algorithm.
Results: For images it had not yet assessed, our algorithm was able to
digitally stain the retinal nerve fiber layer + prelamina, the retinal pigment
epithelium, all other retinal layers, the choroid, and the peripapillary sclera
and lamina cribrosa. For all tissues, the mean dice coefficient was $0.84 \pm
0.03$, the mean sensitivity $0.92 \pm 0.03$, and the mean specificity $0.99 \pm
0.00$. Our algorithm performed significantly better when compensated images
were used for training. Increasing the number of images (from 10 to 40) to
train our algorithm did not significantly improve performance, except for the
RPE.
Conclusion. Our deep learning algorithm can simultaneously stain neural and
connective tissues in ONH images. Our approach offers a framework to
automatically measure multiple key structural parameters of the ONH that may be
critical to improve glaucoma management.
|
Comparison of Decision Tree Based Classification Strategies to Detect
External Chemical Stimuli from Raw and Filtered Plant Electrical Response | Plants monitor their surrounding environment and control their physiological
functions by producing an electrical response. We recorded electrical signals
from different plants by exposing them to Sodium Chloride (NaCl), Ozone (O3)
and Sulfuric Acid (H2SO4) under laboratory conditions. After applying
pre-processing techniques such as filtering and drift removal, we extracted few
statistical features from the acquired plant electrical signals. Using these
features, combined with different classification algorithms, we used a decision
tree based multi-class classification strategy to identify the three different
external chemical stimuli. We here present our exploration to obtain the
optimum set of ranked feature and classifier combination that can separate a
particular chemical stimulus from the incoming stream of plant electrical
signals. The paper also reports an exhaustive comparison of similar feature
based classification using the filtered and the raw plant signals, containing
the high frequency stochastic part and also the low frequency trends present in
it, as two different cases for feature extraction. The work, presented in this
paper opens up new possibilities for using plant electrical signals to monitor
and detect other environmental stimuli apart from NaCl, O3 and H2SO4 in future.
|
Engineering fast multilevel support vector machines | The computational complexity of solving nonlinear support vector machine
(SVM) is prohibitive on large-scale data. In particular, this issue becomes
very sensitive when the data represents additional difficulties such as highly
imbalanced class sizes. Typically, nonlinear kernels produce significantly
higher classification quality to linear kernels but introduce extra kernel and
model parameters which requires computationally expensive fitting. This
increases the quality but also reduces the performance dramatically. We
introduce a generalized fast multilevel framework for regular and weighted SVM
and discuss several versions of its algorithmic components that lead to a good
trade-off between quality and time. Our framework is implemented using PETSc
which allows an easy integration with scientific computing tasks. The
experimental results demonstrate significant speed up compared to the
state-of-the-art nonlinear SVM libraries.
Reproducibility: our source code, documentation and parameters are available
at https:// github.com/esadr/mlsvm.
|
Per-instance Differential Privacy | We consider a refinement of differential privacy --- per instance
differential privacy (pDP), which captures the privacy of a specific individual
with respect to a fixed data set. We show that this is a strict generalization
of the standard DP and inherits all its desirable properties, e.g.,
composition, invariance to side information and closedness to postprocessing,
except that they all hold for every instance separately. When the data is drawn
from a distribution, we show that per-instance DP implies generalization.
Moreover, we provide explicit calculations of the per-instance DP for the
output perturbation on a class of smooth learning problems. The result reveals
an interesting and intuitive fact that an individual has stronger privacy if
he/she has small "leverage score" with respect to the data set and if he/she
can be predicted more accurately using the leave-one-out data set. Our
simulation shows several orders-of-magnitude more favorable privacy and utility
trade-off when we consider the privacy of only the users in the data set. In a
case study on differentially private linear regression, provide a novel
analysis of the One-Posterior-Sample (OPS) estimator and show that when the
data set is well-conditioned it provides $(\epsilon,\delta)$-pDP for any target
individuals and matches the exact lower bound up to a
$1+\tilde{O}(n^{-1}\epsilon^{-2})$ multiplicative factor. We also demonstrate
how we can use a "pDP to DP conversion" step to design AdaOPS which uses
adaptive regularization to achieve the same results with
$(\epsilon,\delta)$-DP.
|
Stochastic Gradient Descent for Relational Logistic Regression via
Partial Network Crawls | Research in statistical relational learning has produced a number of methods
for learning relational models from large-scale network data. While these
methods have been successfully applied in various domains, they have been
developed under the unrealistic assumption of full data access. In practice,
however, the data are often collected by crawling the network, due to
proprietary access, limited resources, and privacy concerns. Recently, we
showed that the parameter estimates for relational Bayes classifiers computed
from network samples collected by existing network crawlers can be quite
inaccurate, and developed a crawl-aware estimation method for such models
(Yang, Ribeiro, and Neville, 2017). In this work, we extend the methodology to
learning relational logistic regression models via stochastic gradient descent
from partial network crawls, and show that the proposed method yields accurate
parameter estimates and confidence intervals.
|
Bellman Gradient Iteration for Inverse Reinforcement Learning | This paper develops an inverse reinforcement learning algorithm aimed at
recovering a reward function from the observed actions of an agent. We
introduce a strategy to flexibly handle different types of actions with two
approximations of the Bellman Optimality Equation, and a Bellman Gradient
Iteration method to compute the gradient of the Q-value with respect to the
reward function. These methods allow us to build a differentiable relation
between the Q-value and the reward function and learn an approximately optimal
reward function with gradient methods. We test the proposed method in two
simulated environments by evaluating the accuracy of different approximations
and comparing the proposed method with existing solutions. The results show
that even with a linear reward function, the proposed method has a comparable
accuracy with the state-of-the-art method adopting a non-linear reward
function, and the proposed method is more flexible because it is defined on
observed actions instead of trajectories.
|
Exact Identification of a Quantum Change Point | The detection of change points is a pivotal task in statistical analysis. In
the quantum realm, it is a new primitive where one aims at identifying the
point where a source that supposedly prepares a sequence of particles in
identical quantum states starts preparing a mutated one. We obtain the optimal
procedure to identify the change point with certainty---naturally at the price
of having a certain probability of getting an inconclusive answer. We obtain
the analytical form of the optimal probability of successful identification for
any length of the particle sequence. We show that the conditional success
probabilities of identifying each possible change point show an unexpected
oscillatory behaviour. We also discuss local (online) protocols and compare
them with the optimal procedure.
|
Desensitized RDCA Subspaces for Compressive Privacy in Machine Learning | The quest for better data analysis and artificial intelligence has lead to
more and more data being collected and stored. As a consequence, more data are
exposed to malicious entities. This paper examines the problem of privacy in
machine learning for classification. We utilize the Ridge Discriminant
Component Analysis (RDCA) to desensitize data with respect to a privacy label.
Based on five experiments, we show that desensitization by RDCA can effectively
protect privacy (i.e. low accuracy on the privacy label) with small loss in
utility. On HAR and CMU Faces datasets, the use of desensitized data results in
random guess level accuracies for privacy at a cost of 5.14% and 0.04%, on
average, drop in the utility accuracies. For Semeion Handwritten Digit dataset,
accuracies of the privacy-sensitive digits are almost zero, while the
accuracies for the utility-relevant digits drop by 7.53% on average. This
presents a promising solution to the problem of privacy in machine learning for
classification.
|
Comparing Aggregators for Relational Probabilistic Models | Relational probabilistic models have the challenge of aggregation, where one
variable depends on a population of other variables. Consider the problem of
predicting gender from movie ratings; this is challenging because the number of
movies per user and users per movie can vary greatly. Surprisingly, aggregation
is not well understood. In this paper, we show that existing relational models
(implicitly or explicitly) either use simple numerical aggregators that lose
great amounts of information, or correspond to naive Bayes, logistic
regression, or noisy-OR that suffer from overconfidence. We propose new simple
aggregators and simple modifications of existing models that empirically
outperform the existing ones. The intuition we provide on different (existing
or new) models and their shortcomings plus our empirical findings promise to
form the foundation for future representations.
|
Concept Drift Detection and Adaptation with Hierarchical Hypothesis
Testing | A fundamental issue for statistical classification models in a streaming
environment is that the joint distribution between predictor and response
variables changes over time (a phenomenon also known as concept drifts), such
that their classification performance deteriorates dramatically. In this paper,
we first present a hierarchical hypothesis testing (HHT) framework that can
detect and also adapt to various concept drift types (e.g., recurrent or
irregular, gradual or abrupt), even in the presence of imbalanced data labels.
A novel concept drift detector, namely Hierarchical Linear Four Rates (HLFR),
is implemented under the HHT framework thereafter. By substituting a
widely-acknowledged retraining scheme with an adaptive training strategy, we
further demonstrate that the concept drift adaptation capability of HLFR can be
significantly boosted. The theoretical analysis on the Type-I and Type-II
errors of HLFR is also performed. Experiments on both simulated and real-world
datasets illustrate that our methods outperform state-of-the-art methods in
terms of detection precision, detection delay as well as the adaptability
across different concept drift types.
|
Linear Discriminant Generative Adversarial Networks | We develop a novel method for training of GANs for unsupervised and class
conditional generation of images, called Linear Discriminant GAN (LD-GAN). The
discriminator of an LD-GAN is trained to maximize the linear separability
between distributions of hidden representations of generated and targeted
samples, while the generator is updated based on the decision hyper-planes
computed by performing LDA over the hidden representations. LD-GAN provides a
concrete metric of separation capacity for the discriminator, and we
experimentally show that it is possible to stabilize the training of LD-GAN
simply by calibrating the update frequencies between generators and
discriminators in the unsupervised case, without employment of normalization
methods and constraints on weights. In the class conditional generation tasks,
the proposed method shows improved training stability together with better
generalization performance compared to WGAN that employs an auxiliary
classifier.
|
Partial Transfer Learning with Selective Adversarial Networks | Adversarial learning has been successfully embedded into deep networks to
learn transferable features, which reduce distribution discrepancy between the
source and target domains. Existing domain adversarial networks assume fully
shared label space across domains. In the presence of big data, there is strong
motivation of transferring both classification and representation models from
existing big domains to unknown small domains. This paper introduces partial
transfer learning, which relaxes the shared label space assumption to that the
target label space is only a subspace of the source label space. Previous
methods typically match the whole source domain to the target domain, which are
prone to negative transfer for the partial transfer problem. We present
Selective Adversarial Network (SAN), which simultaneously circumvents negative
transfer by selecting out the outlier source classes and promotes positive
transfer by maximally matching the data distributions in the shared label
space. Experiments demonstrate that our models exceed state-of-the-art results
for partial transfer learning tasks on several benchmark datasets.
|
Error Bounds for Piecewise Smooth and Switching Regression | The paper deals with regression problems, in which the nonsmooth target is
assumed to switch between different operating modes. Specifically, piecewise
smooth (PWS) regression considers target functions switching deterministically
via a partition of the input space, while switching regression considers
arbitrary switching laws. The paper derives generalization error bounds in
these two settings by following the approach based on Rademacher complexities.
For PWS regression, our derivation involves a chaining argument and a
decomposition of the covering numbers of PWS classes in terms of the ones of
their component functions and the capacity of the classifier partitioning the
input space. This yields error bounds with a radical dependency on the number
of modes. For switching regression, the decomposition can be performed directly
at the level of the Rademacher complexities, which yields bounds with a linear
dependency on the number of modes. By using once more chaining and a
decomposition at the level of covering numbers, we show how to recover a
radical dependency. Examples of applications are given in particular for PWS
and swichting regression with linear and kernel-based component functions.
|
Towards Evolutional Compression | Compressing convolutional neural networks (CNNs) is essential for
transferring the success of CNNs to a wide variety of applications to mobile
devices. In contrast to directly recognizing subtle weights or filters as
redundant in a given CNN, this paper presents an evolutionary method to
automatically eliminate redundant convolution filters. We represent each
compressed network as a binary individual of specific fitness. Then, the
population is upgraded at each evolutionary iteration using genetic operations.
As a result, an extremely compact CNN is generated using the fittest
individual. In this approach, either large or small convolution filters can be
redundant, and filters in the compressed network are more distinct. In
addition, since the number of filters in each convolutional layer is reduced,
the number of filter channels and the size of feature maps are also decreased,
naturally improving both the compression and speed-up ratios. Experiments on
benchmark deep CNN models suggest the superiority of the proposed algorithm
over the state-of-the-art compression methods.
|
Boosted Zero-Shot Learning with Semantic Correlation Regularization | We study zero-shot learning (ZSL) as a transfer learning problem, and focus
on the two key aspects of ZSL, model effectiveness and model adaptation. For
effective modeling, we adopt the boosting strategy to learn a zero-shot
classifier from weak models to a strong model. For adaptable knowledge
transfer, we devise a Semantic Correlation Regularization (SCR) approach to
regularize the boosted model to be consistent with the inter-class semantic
correlations. With SCR embedded in the boosting objective, and with a
self-controlled sample selection for learning robustness, we propose a unified
framework, Boosted Zero-shot classification with Semantic Correlation
Regularization (BZ-SCR). By balancing the SCR-regularized boosted model
selection and the self-controlled sample selection, BZ-SCR is capable of
capturing both discriminative and adaptable feature-to-class semantic
alignments, while ensuring the reliability and adaptability of the learned
samples. The experiments on two ZSL datasets show the superiority of BZ-SCR
over the state-of-the-arts.
|
A Simple Exponential Family Framework for Zero-Shot Learning | We present a simple generative framework for learning to predict previously
unseen classes, based on estimating class-attribute-gated class-conditional
distributions. We model each class-conditional distribution as an exponential
family distribution and the parameters of the distribution of each seen/unseen
class are defined as functions of the respective observed class attributes.
These functions can be learned using only the seen class data and can be used
to predict the parameters of the class-conditional distribution of each unseen
class. Unlike most existing methods for zero-shot learning that represent
classes as fixed embeddings in some vector space, our generative model
naturally represents each class as a probability distribution. It is simple to
implement and also allows leveraging additional unlabeled data from unseen
classes to improve the estimates of their class-conditional distributions using
transductive/semi-supervised learning. Moreover, it extends seamlessly to
few-shot learning by easily updating these distributions when provided with a
small number of additional labelled examples from unseen classes. Through a
comprehensive set of experiments on several benchmark data sets, we demonstrate
the efficacy of our framework.
|
Learning to Singulate Objects using a Push Proposal Network | Learning to act in unstructured environments, such as cluttered piles of
objects, poses a substantial challenge for manipulation robots. We present a
novel neural network-based approach that separates unknown objects in clutter
by selecting favourable push actions. Our network is trained from data
collected through autonomous interaction of a PR2 robot with randomly organized
tabletop scenes. The model is designed to propose meaningful push actions based
on over-segmented RGB-D images. We evaluate our approach by singulating up to 8
unknown objects in clutter. We demonstrate that our method enables the robot to
perform the task with a high success rate and a low number of required push
actions. Our results based on real-world experiments show that our network is
able to generalize to novel objects of various sizes and shapes, as well as to
arbitrary object configurations. Videos of our experiments can be viewed at
http://robotpush.cs.uni-freiburg.de
|
Deep Forecast: Deep Learning-based Spatio-Temporal Forecasting | The paper presents a spatio-temporal wind speed forecasting algorithm using
Deep Learning (DL)and in particular, Recurrent Neural Networks(RNNs). Motivated
by recent advances in renewable energy integration and smart grids, we apply
our proposed algorithm for wind speed forecasting. Renewable energy resources
(wind and solar)are random in nature and, thus, their integration is
facilitated with accurate short-term forecasts. In our proposed framework, we
model the spatiotemporal information by a graph whose nodes are data generating
entities and its edges basically model how these nodes are interacting with
each other. One of the main contributions of our work is the fact that we
obtain forecasts of all nodes of the graph at the same time based on one
framework. Results of a case study on recorded time series data from a
collection of wind mills in the north-east of the U.S. show that the proposed
DL-based forecasting algorithm significantly improves the short-term forecasts
compared to a set of widely-used benchmarks models.
|
A Survey on Multi-Task Learning | Multi-Task Learning (MTL) is a learning paradigm in machine learning and its
aim is to leverage useful information contained in multiple related tasks to
help improve the generalization performance of all the tasks. In this paper, we
give a survey for MTL from the perspective of algorithmic modeling,
applications and theoretical analyses. For algorithmic modeling, we give a
definition of MTL and then classify different MTL algorithms into five
categories, including feature learning approach, low-rank approach, task
clustering approach, task relation learning approach and decomposition approach
as well as discussing the characteristics of each approach. In order to improve
the performance of learning tasks further, MTL can be combined with other
learning paradigms including semi-supervised learning, active learning,
unsupervised learning, reinforcement learning, multi-view learning and
graphical models. When the number of tasks is large or the data dimensionality
is high, we review online, parallel and distributed MTL models as well as
dimensionality reduction and feature hashing to reveal their computational and
storage advantages. Many real-world applications use MTL to boost their
performance and we review representative works in this paper. Finally, we
present theoretical analyses and discuss several future directions for MTL.
|
Proxy Non-Discrimination in Data-Driven Systems | Machine learnt systems inherit biases against protected classes, historically
disparaged groups, from training data. Usually, these biases are not explicit,
they rely on subtle correlations discovered by training algorithms, and are
therefore difficult to detect. We formalize proxy discrimination in data-driven
systems, a class of properties indicative of bias, as the presence of protected
class correlates that have causal influence on the system's output. We evaluate
an implementation on a corpus of social datasets, demonstrating how to validate
systems against these properties and to repair violations where they occur.
|
On The Robustness of a Neural Network | With the development of neural networks based machine learning and their
usage in mission critical applications, voices are rising against the
\textit{black box} aspect of neural networks as it becomes crucial to
understand their limits and capabilities. With the rise of neuromorphic
hardware, it is even more critical to understand how a neural network, as a
distributed system, tolerates the failures of its computing nodes, neurons, and
its communication channels, synapses. Experimentally assessing the robustness
of neural networks involves the quixotic venture of testing all the possible
failures, on all the possible inputs, which ultimately hits a combinatorial
explosion for the first, and the impossibility to gather all the possible
inputs for the second.
In this paper, we prove an upper bound on the expected error of the output
when a subset of neurons crashes. This bound involves dependencies on the
network parameters that can be seen as being too pessimistic in the average
case. It involves a polynomial dependency on the Lipschitz coefficient of the
neurons activation function, and an exponential dependency on the depth of the
layer where a failure occurs. We back up our theoretical results with
experiments illustrating the extent to which our prediction matches the
dependencies between the network parameters and robustness. Our results show
that the robustness of neural networks to the average crash can be estimated
without the need to neither test the network on all failure configurations, nor
access the training set used to train the network, both of which are
practically impossible requirements.
|
Efficient Low Rank Tensor Ring Completion | Using the matrix product state (MPS) representation of the recently proposed
tensor ring decompositions, in this paper we propose a tensor completion
algorithm, which is an alternating minimization algorithm that alternates over
the factors in the MPS representation. This development is motivated in part by
the success of matrix completion algorithms that alternate over the (low-rank)
factors. In this paper, we propose a spectral initialization for the tensor
ring completion algorithm and analyze the computational complexity of the
proposed algorithm. We numerically compare it with existing methods that employ
a low rank tensor train approximation for data completion and show that our
method outperforms the existing ones for a variety of real computer vision
settings, and thus demonstrate the improved expressive power of tensor ring as
compared to tensor train.
|
Dual Rectified Linear Units (DReLUs): A Replacement for Tanh Activation
Functions in Quasi-Recurrent Neural Networks | In this paper, we introduce a novel type of Rectified Linear Unit (ReLU),
called a Dual Rectified Linear Unit (DReLU). A DReLU, which comes with an
unbounded positive and negative image, can be used as a drop-in replacement for
a tanh activation function in the recurrent step of Quasi-Recurrent Neural
Networks (QRNNs) (Bradbury et al. (2017)). Similar to ReLUs, DReLUs are less
prone to the vanishing gradient problem, they are noise robust, and they induce
sparse activations.
We independently reproduce the QRNN experiments of Bradbury et al. (2017) and
compare our DReLU-based QRNNs with the original tanh-based QRNNs and Long
Short-Term Memory networks (LSTMs) on sentiment classification and word-level
language modeling. Additionally, we evaluate on character-level language
modeling, showing that we are able to stack up to eight QRNN layers with
DReLUs, thus making it possible to improve the current state-of-the-art in
character-level language modeling over shallow architectures based on LSTMs.
|
SLEEPNET: Automated Sleep Staging System via Deep Learning | Sleep disorders, such as sleep apnea, parasomnias, and hypersomnia, affect
50-70 million adults in the United States (Hillman et al., 2006). Overnight
polysomnography (PSG), including brain monitoring using electroencephalography
(EEG), is a central component of the diagnostic evaluation for sleep disorders.
While PSG is conventionally performed by trained technologists, the recent rise
of powerful neural network learning algorithms combined with large
physiological datasets offers the possibility of automation, potentially making
expert-level sleep analysis more widely available. We propose SLEEPNET (Sleep
EEG neural network), a deployed annotation tool for sleep staging. SLEEPNET
uses a deep recurrent neural network trained on the largest sleep physiology
database assembled to date, consisting of PSGs from over 10,000 patients from
the Massachusetts General Hospital (MGH) Sleep Laboratory. SLEEPNET achieves
human-level annotation performance on an independent test set of 1,000 EEGs,
with an average accuracy of 85.76% and algorithm-expert inter-rater agreement
(IRA) of kappa = 79.46%, comparable to expert-expert IRA.
|
Dragon: A Computation Graph Virtual Machine Based Deep Learning
Framework | Deep Learning has made a great progress for these years. However, it is still
difficult to master the implement of various models because different
researchers may release their code based on different frameworks or interfaces.
In this paper, we proposed a computation graph based framework which only aims
to introduce well-known interfaces. It will help a lot when reproducing a newly
model or transplanting models that were implemented by other frameworks.
Additionally, we implement numerous recent models covering both Computer Vision
and Nature Language Processing. We demonstrate that our framework will not
suffer from model-starving because it is much easier to make full use of the
works that are already done.
|
MMGAN: Manifold Matching Generative Adversarial Network | It is well-known that GANs are difficult to train, and several different
techniques have been proposed in order to stabilize their training. In this
paper, we propose a novel training method called manifold-matching, and a new
GAN model called manifold-matching GAN (MMGAN). MMGAN finds two manifolds
representing the vector representations of real and fake images. If these two
manifolds match, it means that real and fake images are statistically
identical. To assist the manifold-matching task, we also use i) kernel tricks
to find better manifold structures, ii) moving-averaged manifolds across
mini-batches, and iii) a regularizer based on correlation matrix to suppress
mode collapse.
We conduct in-depth experiments with three image datasets and compare with
several state-of-the-art GAN models. 32.4% of images generated by the proposed
MMGAN are recognized as fake images during our user study (16% enhancement
compared to other state-of-the-art model). MMGAN achieved an unsupervised
inception score of 7.8 for CIFAR-10.
|
Unsupervised Motion Artifact Detection in Wrist-Measured Electrodermal
Activity Data | One of the main benefits of a wrist-worn computer is its ability to collect a
variety of physiological data in a minimally intrusive manner. Among these
data, electrodermal activity (EDA) is readily collected and provides a window
into a person's emotional and sympathetic responses. EDA data collected using a
wearable wristband are easily influenced by motion artifacts (MAs) that may
significantly distort the data and degrade the quality of analyses performed on
the data if not identified and removed. Prior work has demonstrated that MAs
can be successfully detected using supervised machine learning algorithms on a
small data set collected in a lab setting. In this paper, we demonstrate that
unsupervised learning algorithms perform competitively with supervised
algorithms for detecting MAs on EDA data collected in both a lab-based setting
and a real-world setting comprising about 23 hours of data. We also find,
somewhat surprisingly, that incorporating accelerometer data as well as EDA
improves detection accuracy only slightly for supervised algorithms and
significantly degrades the accuracy of unsupervised algorithms.
|
Graph-Based Classification of Omnidirectional Images | Omnidirectional cameras are widely used in such areas as robotics and virtual
reality as they provide a wide field of view. Their images are often processed
with classical methods, which might unfortunately lead to non-optimal solutions
as these methods are designed for planar images that have different geometrical
properties than omnidirectional ones. In this paper we study image
classification task by taking into account the specific geometry of
omnidirectional cameras with graph-based representations. In particular, we
extend deep learning architectures to data on graphs; we propose a principled
way of graph construction such that convolutional filters respond similarly for
the same pattern on different positions of the image regardless of lens
distortions. Our experiments show that the proposed method outperforms current
techniques for the omnidirectional image classification problem.
|
Tensor Regression Networks | Convolutional neural networks typically consist of many convolutional layers
followed by one or more fully connected layers. While convolutional layers map
between high-order activation tensors, the fully connected layers operate on
flattened activation vectors. Despite empirical success, this approach has
notable drawbacks. Flattening followed by fully connected layers discards
multilinear structure in the activations and requires many parameters. We
address these problems by incorporating tensor algebraic operations that
preserve multilinear structure at every layer. First, we introduce Tensor
Contraction Layers (TCLs) that reduce the dimensionality of their input while
preserving their multilinear structure using tensor contraction. Next, we
introduce Tensor Regression Layers (TRLs), which express outputs through a
low-rank multilinear mapping from a high-order activation tensor to an output
tensor of arbitrary order. We learn the contraction and regression factors
end-to-end, and produce accurate nets with fewer parameters. Additionally, our
layers regularize networks by imposing low-rank constraints on the activations
(TCL) and regression weights (TRL). Experiments on ImageNet show that, applied
to VGG and ResNet architectures, TCLs and TRLs reduce the number of parameters
compared to fully connected layers by more than 65% while maintaining or
increasing accuracy. In addition to the space savings, our approach's ability
to leverage topological structure can be crucial for structured data such as
MRI. In particular, we demonstrate significant performance improvements over
comparable architectures on three tasks associated with the UK Biobank dataset.
|
Learning Sparse Representations in Reinforcement Learning with Sparse
Coding | A variety of representation learning approaches have been investigated for
reinforcement learning; much less attention, however, has been given to
investigating the utility of sparse coding. Outside of reinforcement learning,
sparse coding representations have been widely used, with non-convex objectives
that result in discriminative representations. In this work, we develop a
supervised sparse coding objective for policy evaluation. Despite the
non-convexity of this objective, we prove that all local minima are global
minima, making the approach amenable to simple optimization strategies. We
empirically show that it is key to use a supervised objective, rather than the
more straightforward unsupervised sparse coding approach. We compare the
learned representations to a canonical fixed sparse representation, called
tile-coding, demonstrating that the sparse coding representation outperforms a
wide variety of tilecoding representations.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.