title
stringlengths 7
246
| abstract
stringlengths 6
3.31k
|
|---|---|
Meta-Learner with Linear Nulling
|
We propose a meta-learning algorithm utilizing a linear transformer that
carries out null-space projection of neural network outputs. The main idea is
to construct an alternative classification space such that the error signals
during few-shot learning are quickly zero-forced on that space so that reliable
classification on low data is possible. The final decision on a query is
obtained utilizing a null-space-projected distance measure between the network
output and reference vectors, both of which have been trained in the initial
learning phase. Among the known methods with a given model size, our
meta-learner achieves the best or near-best image classification accuracies
with Omniglot and miniImageNet datasets.
|
Normative Modeling of Neuroimaging Data using Scalable Multi-Task
Gaussian Processes
|
Normative modeling has recently been proposed as an alternative for the
case-control approach in modeling heterogeneity within clinical cohorts.
Normative modeling is based on single-output Gaussian process regression that
provides coherent estimates of uncertainty required by the method but does not
consider spatial covariance structure. Here, we introduce a scalable multi-task
Gaussian process regression (S-MTGPR) approach to address this problem. To this
end, we exploit a combination of a low-rank approximation of the spatial
covariance matrix with algebraic properties of Kronecker product in order to
reduce the computational complexity of Gaussian process regression in
high-dimensional output spaces. On a public fMRI dataset, we show that S-MTGPR:
1) leads to substantial computational improvements that allow us to estimate
normative models for high-dimensional fMRI data whilst accounting for spatial
structure in data; 2) by modeling both spatial and across-sample variances, it
provides higher sensitivity in novelty detection scenarios.
|
Neural Network-Based Equations for Predicting PGA and PGV in Texas,
Oklahoma, and Kansas
|
Parts of Texas, Oklahoma, and Kansas have experienced increased rates of
seismicity in recent years, providing new datasets of earthquake recordings to
develop ground motion prediction models for this particular region of the
Central and Eastern North America (CENA). This paper outlines a framework for
using Artificial Neural Networks (ANNs) to develop attenuation models from the
ground motion recordings in this region. While attenuation models exist for the
CENA, concerns over the increased rate of seismicity in this region necessitate
investigation of ground motions prediction models particular to these states.
To do so, an ANN-based framework is proposed to predict peak ground
acceleration (PGA) and peak ground velocity (PGV) given magnitude, earthquake
source-to-site distance, and shear wave velocity. In this framework,
approximately 4,500 ground motions with magnitude greater than 3.0 recorded in
these three states (Texas, Oklahoma, and Kansas) since 2005 are considered.
Results from this study suggest that existing ground motion prediction models
developed for CENA do not accurately predict the ground motion intensity
measures for earthquakes in this region, especially for those with low
source-to-site distances or on very soft soil conditions. The proposed ANN
models provide much more accurate prediction of the ground motion intensity
measures at all distances and magnitudes. The proposed ANN models are also
converted to relatively simple mathematical equations so that engineers can
easily use them to predict the ground motion intensity measures for future
events. Finally, through a sensitivity analysis, the contributions of the
predictive parameters to the prediction of the considered intensity measures
are investigated.
|
iFair: Learning Individually Fair Data Representations for Algorithmic
Decision Making
|
People are rated and ranked, towards algorithmic decision making in an
increasing number of applications, typically based on machine learning.
Research on how to incorporate fairness into such tasks has prevalently pursued
the paradigm of group fairness: giving adequate success rates to specifically
protected groups. In contrast, the alternative paradigm of individual fairness
has received relatively little attention, and this paper advances this less
explored direction. The paper introduces a method for probabilistically mapping
user records into a low-rank representation that reconciles individual fairness
and the utility of classifiers and rankings in downstream applications. Our
notion of individual fairness requires that users who are similar in all
task-relevant attributes such as job qualification, and disregarding all
potentially discriminating attributes such as gender, should have similar
outcomes. We demonstrate the versatility of our method by applying it to
classification and learning-to-rank tasks on a variety of real-world datasets.
Our experiments show substantial improvements over the best prior work for this
setting.
|
A Highly Parallel FPGA Implementation of Sparse Neural Network Training
|
We demonstrate an FPGA implementation of a parallel and reconfigurable
architecture for sparse neural networks, capable of on-chip training and
inference. The network connectivity uses pre-determined, structured sparsity to
significantly reduce complexity by lowering memory and computational
requirements. The architecture uses a notion of edge-processing, leading to
efficient pipelining and parallelization. Moreover, the device can be
reconfigured to trade off resource utilization with training time to fit
networks and datasets of varying sizes. The combined effects of complexity
reduction and easy reconfigurability enable significantly greater exploration
of network hyperparameters and structures on-chip. As proof of concept, we show
implementation results on an Artix-7 FPGA.
|
Robustifying Independent Component Analysis by Adjusting for Group-Wise
Stationary Noise
|
We introduce coroICA, confounding-robust independent component analysis, a
novel ICA algorithm which decomposes linearly mixed multivariate observations
into independent components that are corrupted (and rendered dependent) by
hidden group-wise stationary confounding. It extends the ordinary ICA model in
a theoretically sound and explicit way to incorporate group-wise (or
environment-wise) confounding. We show that our proposed general noise model
allows to perform ICA in settings where other noisy ICA procedures fail.
Additionally, it can be used for applications with grouped data by adjusting
for different stationary noise within each group. Our proposed noise model has
a natural relation to causality and we explain how it can be applied in the
context of causal inference. In addition to our theoretical framework, we
provide an efficient estimation procedure and prove identifiability of the
unmixing matrix under mild assumptions. Finally, we illustrate the performance
and robustness of our method on simulated data, provide audible and visual
examples, and demonstrate the applicability to real-world scenarios by
experiments on publicly available Antarctic ice core data as well as two EEG
data sets. We provide a scikit-learn compatible pip-installable Python package
coroICA as well as R and Matlab implementations accompanied by a documentation
at https://sweichwald.de/coroICA/
|
GANAX: A Unified MIMD-SIMD Acceleration for Generative Adversarial
Networks
|
Generative Adversarial Networks (GANs) are one of the most recent deep
learning models that generate synthetic data from limited genuine datasets.
GANs are on the frontier as further extension of deep learning into many
domains (e.g., medicine, robotics, content synthesis) requires massive sets of
labeled data that is generally either unavailable or prohibitively costly to
collect. Although GANs are gaining prominence in various fields, there are no
accelerators for these new models. In fact, GANs leverage a new operator,
called transposed convolution, that exposes unique challenges for hardware
acceleration. This operator first inserts zeros within the multidimensional
input, then convolves a kernel over this expanded array to add information to
the embedded zeros. Even though there is a convolution stage in this operator,
the inserted zeros lead to underutilization of the compute resources when a
conventional convolution accelerator is employed. We propose the GANAX
architecture to alleviate the sources of inefficiency associated with the
acceleration of GANs using conventional convolution accelerators, making the
first GAN accelerator design possible. We propose a reorganization of the
output computations to allocate compute rows with similar patterns of zeros to
adjacent processing engines, which also avoids inconsequential multiply-adds on
the zeros. This compulsory adjacency reclaims data reuse across these
neighboring processing engines, which had otherwise diminished due to the
inserted zeros. The reordering breaks the full SIMD execution model, which is
prominent in convolution accelerators. Therefore, we propose a unified
MIMD-SIMD design for GANAX that leverages repeated patterns in the computation
to create distinct microprograms that execute concurrently in SIMD mode.
|
Machine Learning for Predictive Analytics of Compute Cluster Jobs
|
We address the problem of predicting whether sufficient memory and CPU
resources have been requested for jobs at submission time. For this purpose, we
examine the task of training a supervised machine learning system to predict
the outcome - whether the job will fail specifically due to insufficient
resources - as a classification task. Sufficiently high accuracy, precision,
and recall at this task facilitates more anticipatory decision support
applications in the domain of HPC resource allocation. Our preliminary results
using a new test bed show that the probability of failed jobs is associated
with information freely available at job submission time and may thus be usable
by a learning system for user modeling that gives personalized feedback to
users.
|
Text to brain: predicting the spatial distribution of neuroimaging
observations from text reports
|
Despite the digital nature of magnetic resonance imaging, the resulting
observations are most frequently reported and stored in text documents. There
is a trove of information untapped in medical health records, case reports, and
medical publications. In this paper, we propose to mine brain medical
publications to learn the spatial distribution associated with anatomical
terms. The problem is formulated in terms of minimization of a risk on
distributions which leads to a least-deviation cost function. An efficient
algorithm in the dual then learns the mapping from documents to brain
structures. Empirical results using coordinates extracted from the
brain-imaging literature show that i) models must adapt to semantic variation
in the terms used to describe a given anatomical structure, ii) voxel-wise
parameterization leads to higher likelihood of locations reported in unseen
documents, iii) least-deviation cost outperforms least-square. As a proof of
concept for our method, we use our model of spatial distributions to predict
the distribution of specific neurological conditions from text-only reports.
|
Machines hear better when they have ears
|
Deep-neural-network (DNN) based noise suppression systems yield significant
improvements over conventional approaches such as spectral subtraction and
non-negative matrix factorization, but do not generalize well to noise
conditions they were not trained for. In comparison to DNNs, humans show
remarkable noise suppression capabilities that yield successful speech
intelligibility under various adverse listening conditions and negative
signal-to-noise ratios (SNRs). Motivated by the excellent human performance,
this paper explores whether numerical models that simulate human cochlear
signal processing can be combined with DNNs to improve the robustness of DNN
based noise suppression systems. Five cochlear models were coupled to
fully-connected and recurrent NN-based noise suppression systems and were
trained and evaluated for a variety of noise conditions using objective
metrics: perceptual speech quality (PESQ), segmental SNR and cepstral distance.
The simulations show that biophysically-inspired cochlear models improve the
generalizability of DNN-based noise suppression systems for unseen noise and
negative SNRs. This approach thus leads to robust noise suppression systems
that are less sensitive to the noise type and noise level. Because cochlear
models capture the intrinsic nonlinearities and dynamics of peripheral auditory
processing, it is shown here that accounting for their deterministic signal
processing improves machine hearing and avoids overtraining of multi-layer
DNNs. We hence conclude that machines hear better when realistic cochlear
models are used at the input of DNNs.
|
Efficient and Scalable Batch Bayesian Optimization Using K-Means
|
We present K-Means Batch Bayesian Optimization (KMBBO), a novel batch
sampling algorithm for Bayesian Optimization (BO). KMBBO uses unsupervised
learning to efficiently estimate peaks of the model acquisition function. We
show in empirical experiments that our method outperforms the current
state-of-the-art batch allocation algorithms on a variety of test problems
including tuning of algorithm hyper-parameters and a challenging drug discovery
problem. In order to accommodate the real-world problem of high dimensional
data, we propose a modification to KMBBO by combining it with compressed
sensing to project the optimization into a lower dimensional subspace. We
demonstrate empirically that this 2-step method outperforms algorithms where no
dimensionality reduction has taken place.
|
TD or not TD: Analyzing the Role of Temporal Differencing in Deep
Reinforcement Learning
|
Our understanding of reinforcement learning (RL) has been shaped by
theoretical and empirical results that were obtained decades ago using tabular
representations and linear function approximators. These results suggest that
RL methods that use temporal differencing (TD) are superior to direct Monte
Carlo estimation (MC). How do these results hold up in deep RL, which deals
with perceptually complex environments and deep nonlinear models? In this
paper, we re-examine the role of TD in modern deep RL, using specially designed
environments that control for specific factors that affect performance, such as
reward sparsity, reward delay, and the perceptual complexity of the task. When
comparing TD with infinite-horizon MC, we are able to reproduce classic results
in modern settings. Yet we also find that finite-horizon MC is not inferior to
TD, even when rewards are sparse or delayed. This makes MC a viable alternative
to TD in deep RL.
|
Online Reciprocal Recommendation with Theoretical Performance Guarantees
|
A reciprocal recommendation problem is one where the goal of learning is not
just to predict a user's preference towards a passive item (e.g., a book), but
to recommend the targeted user on one side another user from the other side
such that a mutual interest between the two exists. The problem thus is sharply
different from the more traditional items-to-users recommendation, since a good
match requires meeting the preferences of both users. We initiate a rigorous
theoretical investigation of the reciprocal recommendation task in a specific
framework of sequential learning. We point out general limitations, formulate
reasonable assumptions enabling effective learning and, under these
assumptions, we design and analyze a computationally efficient algorithm that
uncovers mutual likes at a pace comparable to those achieved by a clearvoyant
algorithm knowing all user preferences in advance. Finally, we validate our
algorithm against synthetic and real-world datasets, showing improved empirical
performance over simple baselines.
|
Penalizing side effects using stepwise relative reachability
|
How can we design safe reinforcement learning agents that avoid unnecessary
disruptions to their environment? We show that current approaches to penalizing
side effects can introduce bad incentives, e.g. to prevent any irreversible
changes in the environment, including the actions of other agents. To isolate
the source of such undesirable incentives, we break down side effects penalties
into two components: a baseline state and a measure of deviation from this
baseline state. We argue that some of these incentives arise from the choice of
baseline, and others arise from the choice of deviation measure. We introduce a
new variant of the stepwise inaction baseline and a new deviation measure based
on relative reachability of states. The combination of these design choices
avoids the given undesirable incentives, while simpler baselines and the
unreachability measure fail. We demonstrate this empirically by comparing
different combinations of baseline and deviation measure choices on a set of
gridworld experiments designed to illustrate possible bad incentives.
|
Relational inductive bias for physical construction in humans and
machines
|
While current deep learning systems excel at tasks such as object
classification, language processing, and gameplay, few can construct or modify
a complex system such as a tower of blocks. We hypothesize that what these
systems lack is a "relational inductive bias": a capacity for reasoning about
inter-object relations and making choices over a structured description of a
scene. To test this hypothesis, we focus on a task that involves gluing pairs
of blocks together to stabilize a tower, and quantify how well humans perform.
We then introduce a deep reinforcement learning agent which uses object- and
relation-centric scene and policy representations and apply it to the task. Our
results show that these structured representations allow the agent to
outperform both humans and more naive approaches, suggesting that relational
inductive bias is an important component in solving structured reasoning
problems and for building more intelligent, flexible machines.
|
Deep Graphs
|
We propose an algorithm for deep learning on networks and graphs. It relies
on the notion that many graph algorithms, such as PageRank, Weisfeiler-Lehman,
or Message Passing can be expressed as iterative vertex updates. Unlike
previous methods which rely on the ingenuity of the designer, Deep Graphs are
adaptive to the estimation problem. Training and deployment are both efficient,
since the cost is $O(|E| + |V|)$, where $E$ and $V$ are the sets of edges and
vertices respectively. In short, we learn the recurrent update functions rather
than positing their specific functional form. This yields an algorithm that
achieves excellent accuracy on both graph labeling and regression tasks.
|
Diffeomorphic Learning
|
We introduce in this paper a learning paradigm in which the training data is
transformed by a diffeomorphic transformation before prediction. The learning
algorithm minimizes a cost function evaluating the prediction error on the
training set penalized by the distance between the diffeomorphism and the
identity. The approach borrows ideas from shape analysis where diffeomorphisms
are estimated for shape and image alignment, and brings them in a previously
unexplored setting, estimating, in particular diffeomorphisms in much larger
dimensions. After introducing the concept and describing a learning algorithm,
we present diverse applications, mostly with synthetic examples, demonstrating
the potential of the approach, as well as some insight on how it can be
improved.
|
Graph networks as learnable physics engines for inference and control
|
Understanding and interacting with everyday physical scenes requires rich
knowledge about the structure of the world, represented either implicitly in a
value or policy function, or explicitly in a transition model. Here we
introduce a new class of learnable models--based on graph networks--which
implement an inductive bias for object- and relation-centric representations of
complex, dynamical systems. Our results show that as a forward model, our
approach supports accurate predictions from real and simulated data, and
surprisingly strong and efficient generalization, across eight distinct
physical systems which we varied parametrically and structurally. We also found
that our inference model can perform system identification. Our models are also
differentiable, and support online planning via gradient-based trajectory
optimization, as well as offline policy optimization. Our framework offers new
opportunities for harnessing and exploiting rich knowledge about the world, and
takes a key step toward building machines with more human-like representations
of the world.
|
ML-Leaks: Model and Data Independent Membership Inference Attacks and
Defenses on Machine Learning Models
|
Machine learning (ML) has become a core component of many real-world
applications and training data is a key factor that drives current progress.
This huge success has led Internet companies to deploy machine learning as a
service (MLaaS). Recently, the first membership inference attack has shown that
extraction of information on the training set is possible in such MLaaS
settings, which has severe security and privacy implications.
However, the early demonstrations of the feasibility of such attacks have
many assumptions on the adversary, such as using multiple so-called shadow
models, knowledge of the target model structure, and having a dataset from the
same distribution as the target model's training data. We relax all these key
assumptions, thereby showing that such attacks are very broadly applicable at
low cost and thereby pose a more severe risk than previously thought. We
present the most comprehensive study so far on this emerging and developing
threat using eight diverse datasets which show the viability of the proposed
attacks across domains.
In addition, we propose the first effective defense mechanisms against such
broader class of membership inference attacks that maintain a high level of
utility of the ML model.
|
Dynamically Hierarchy Revolution: DirNet for Compressing Recurrent
Neural Network on Mobile Devices
|
Recurrent neural networks (RNNs) achieve cutting-edge performance on a
variety of problems. However, due to their high computational and memory
demands, deploying RNNs on resource constrained mobile devices is a challenging
task. To guarantee minimum accuracy loss with higher compression rate and
driven by the mobile resource requirement, we introduce a novel model
compression approach DirNet based on an optimized fast dictionary learning
algorithm, which 1) dynamically mines the dictionary atoms of the projection
dictionary matrix within layer to adjust the compression rate 2) adaptively
changes the sparsity of sparse codes cross the hierarchical layers.
Experimental results on language model and an ASR model trained with a 1000h
speech dataset demonstrate that our method significantly outperforms prior
approaches. Evaluated on off-the-shelf mobile devices, we are able to reduce
the size of original model by eight times with real-time model inference and
negligible accuracy loss.
|
Agreement-based Learning
|
Model selection is a problem that has occupied machine learning researchers
for a long time. Recently, its importance has become evident through
applications in deep learning. We propose an agreement-based learning framework
that prevents many of the pitfalls associated with model selection. It relies
on coupling the training of multiple models by encouraging them to agree on
their predictions while training. In contrast with other model selection and
combination approaches used in machine learning, the proposed framework is
inspired by human learning. We also propose a learning algorithm defined within
this framework which manages to significantly outperform alternatives in
practice, and whose performance improves further with the availability of
unlabeled data. Finally, we describe a number of potential directions for
developing more flexible agreement-based learning algorithms.
|
Learning a Code: Machine Learning for Approximate Non-Linear Coded
Computation
|
Machine learning algorithms are typically run on large scale, distributed
compute infrastructure that routinely face a number of unavailabilities such as
failures and temporary slowdowns. Adding redundant computations using
coding-theoretic tools called "codes" is an emerging technique to alleviate the
adverse effects of such unavailabilities. A code consists of an encoding
function that proactively introduces redundant computation and a decoding
function that reconstructs unavailable outputs using the available ones. Past
work focuses on using codes to provide resilience for linear computations and
specific iterative optimization algorithms. However, computations performed for
a variety of applications including inference on state-of-the-art machine
learning algorithms, such as neural networks, typically fall outside this
realm. In this paper, we propose taking a learning-based approach to designing
codes that can handle non-linear computations. We present carefully designed
neural network architectures and a training methodology for learning encoding
and decoding functions that produce approximate reconstructions of unavailable
computation results. We present extensive experimental results demonstrating
the effectiveness of the proposed approach: we show that the our learned codes
can accurately reconstruct $64 - 98\%$ of the unavailable predictions from
neural-network based image classifiers on the MNIST, Fashion-MNIST, and
CIFAR-10 datasets. To the best of our knowledge, this work proposes the first
learning-based approach for designing codes, and also presents the first
coding-theoretic solution that can provide resilience for any non-linear
(differentiable) computation. Our results show that learning can be an
effective technique for designing codes, and that learned codes are a highly
promising approach for bringing the benefits of coding to non-linear
computations.
|
Relational inductive biases, deep learning, and graph networks
|
Artificial intelligence (AI) has undergone a renaissance recently, making
major progress in key domains such as vision, language, control, and
decision-making. This has been due, in part, to cheap data and cheap compute
resources, which have fit the natural strengths of deep learning. However, many
defining characteristics of human intelligence, which developed under much
different pressures, remain out of reach for current approaches. In particular,
generalizing beyond one's experiences--a hallmark of human intelligence from
infancy--remains a formidable challenge for modern AI.
The following is part position paper, part review, and part unification. We
argue that combinatorial generalization must be a top priority for AI to
achieve human-like abilities, and that structured representations and
computations are key to realizing this objective. Just as biology uses nature
and nurture cooperatively, we reject the false choice between
"hand-engineering" and "end-to-end" learning, and instead advocate for an
approach which benefits from their complementary strengths. We explore how
using relational inductive biases within deep learning architectures can
facilitate learning about entities, relations, and rules for composing them. We
present a new building block for the AI toolkit with a strong relational
inductive bias--the graph network--which generalizes and extends various
approaches for neural networks that operate on graphs, and provides a
straightforward interface for manipulating structured knowledge and producing
structured behaviors. We discuss how graph networks can support relational
reasoning and combinatorial generalization, laying the foundation for more
sophisticated, interpretable, and flexible patterns of reasoning. As a
companion to this paper, we have released an open-source software library for
building graph networks, with demonstrations of how to use them in practice.
|
Equivalence Between Wasserstein and Value-Aware Loss for Model-based
Reinforcement Learning
|
Learning a generative model is a key component of model-based reinforcement
learning. Though learning a good model in the tabular setting is a simple task,
learning a useful model in the approximate setting is challenging. In this
context, an important question is the loss function used for model learning as
varying the loss function can have a remarkable impact on effectiveness of
planning. Recently Farahmand et al. (2017) proposed a value-aware model
learning (VAML) objective that captures the structure of value function during
model learning. Using tools from Asadi et al. (2018), we show that minimizing
the VAML objective is in fact equivalent to minimizing the Wasserstein metric.
This equivalence improves our understanding of value-aware models, and also
creates a theoretical foundation for applications of Wasserstein in model-based
reinforcement~learning.
|
Internal Model from Observations for Reward Shaping
|
Reinforcement learning methods require careful design involving a reward
function to obtain the desired action policy for a given task. In the absence
of hand-crafted reward functions, prior work on the topic has proposed several
methods for reward estimation by using expert state trajectories and action
pairs. However, there are cases where complete or good action information
cannot be obtained from expert demonstrations. We propose a novel reinforcement
learning method in which the agent learns an internal model of observation on
the basis of expert-demonstrated state trajectories to estimate rewards without
completely learning the dynamics of the external environment from state-action
pairs. The internal model is obtained in the form of a predictive model for the
given expert state distribution. During reinforcement learning, the agent
predicts the reward as a function of the difference between the actual state
and the state predicted by the internal model. We conducted multiple
experiments in environments of varying complexity, including the Super Mario
Bros and Flappy Bird games. We show our method successfully trains good
policies directly from expert game-play videos.
|
Universal Statistics of Fisher Information in Deep Neural Networks: Mean
Field Approach
|
The Fisher information matrix (FIM) is a fundamental quantity to represent
the characteristics of a stochastic model, including deep neural networks
(DNNs). The present study reveals novel statistics of FIM that are universal
among a wide class of DNNs. To this end, we use random weights and large width
limits, which enables us to utilize mean field theories. We investigate the
asymptotic statistics of the FIM's eigenvalues and reveal that most of them are
close to zero while the maximum eigenvalue takes a huge value. Because the
landscape of the parameter space is defined by the FIM, it is locally flat in
most dimensions, but strongly distorted in others. Moreover, we demonstrate the
potential usage of the derived statistics in learning strategies. First, small
eigenvalues that induce flatness can be connected to a norm-based capacity
measure of generalization ability. Second, the maximum eigenvalue that induces
the distortion enables us to quantitatively estimate an appropriately sized
learning rate for gradient methods to converge.
|
Data-driven Localization and Estimation of Disturbance in the
Interconnected Power System
|
Identifying the location of a disturbance and its magnitude is an important
component for stable operation of power systems. We study the problem of
localizing and estimating a disturbance in the interconnected power system. We
take a model-free approach to this problem by using frequency data from
generators. Specifically, we develop a logistic regression based method for
localization and a linear regression based method for estimation of the
magnitude of disturbance. Our model-free approach does not require the
knowledge of system parameters such as inertia constants and topology, and is
shown to achieve highly accurate localization and estimation performance even
in the presence of measurement noise and missing data.
|
Backdrop: Stochastic Backpropagation
|
We introduce backdrop, a flexible and simple-to-implement method, intuitively
described as dropout acting only along the backpropagation pipeline. Backdrop
is implemented via one or more masking layers which are inserted at specific
points along the network. Each backdrop masking layer acts as the identity in
the forward pass, but randomly masks parts of the backward gradient
propagation. Intuitively, inserting a backdrop layer after any convolutional
layer leads to stochastic gradients corresponding to features of that scale.
Therefore, backdrop is well suited for problems in which the data have a
multi-scale, hierarchical structure. Backdrop can also be applied to problems
with non-decomposable loss functions where standard SGD methods are not well
suited. We perform a number of experiments and demonstrate that backdrop leads
to significant improvements in generalization.
|
Importance Sampling Policy Evaluation with an Estimated Behavior Policy
|
We consider the problem of off-policy evaluation in Markov decision
processes. Off-policy evaluation is the task of evaluating the expected return
of one policy with data generated by a different, behavior policy. Importance
sampling is a technique for off-policy evaluation that re-weights off-policy
returns to account for differences in the likelihood of the returns between the
two policies. In this paper, we study importance sampling with an estimated
behavior policy where the behavior policy estimate comes from the same set of
data used to compute the importance sampling estimate. We find that this
estimator often lowers the mean squared error of off-policy evaluation compared
to importance sampling with the true behavior policy or using a behavior policy
that is estimated from a separate data set. Intuitively, estimating the
behavior policy in this way corrects for error due to sampling in the
action-space. Our empirical results also extend to other popular variants of
importance sampling and show that estimating a non-Markovian behavior policy
can further lower large-sample mean squared error even when the true behavior
policy is Markovian.
|
Natural Language Generation for Electronic Health Records
|
A variety of methods existing for generating synthetic electronic health
records (EHRs), but they are not capable of generating unstructured text, like
emergency department (ED) chief complaints, history of present illness or
progress notes. Here, we use the encoder-decoder model, a deep learning
algorithm that features in many contemporary machine translation systems, to
generate synthetic chief complaints from discrete variables in EHRs, like age
group, gender, and discharge diagnosis. After being trained end-to-end on
authentic records, the model can generate realistic chief complaint text that
preserves much of the epidemiological information in the original data. As a
side effect of the model's optimization goal, these synthetic chief complaints
are also free of relatively uncommon abbreviation and misspellings, and they
include none of the personally-identifiable information (PII) that was in the
training data, suggesting it may be used to support the de-identification of
text in EHRs. When combined with algorithms like generative adversarial
networks (GANs), our model could be used to generate fully-synthetic EHRs,
facilitating data sharing between healthcare providers and researchers and
improving our ability to develop machine learning methods tailored to the
information in healthcare data.
|
Playing Atari with Six Neurons
|
Deep reinforcement learning, applied to vision-based problems like Atari
games, maps pixels directly to actions; internally, the deep neural network
bears the responsibility of both extracting useful information and making
decisions based on it. By separating the image processing from decision-making,
one could better understand the complexity of each task, as well as potentially
find smaller policy representations that are easier for humans to understand
and may generalize better. To this end, we propose a new method for learning
policies and compact state representations separately but simultaneously for
policy approximation in reinforcement learning. State representations are
generated by an encoder based on two novel algorithms: Increasing Dictionary
Vector Quantization makes the encoder capable of growing its dictionary size
over time, to address new observations as they appear in an open-ended
online-learning context; Direct Residuals Sparse Coding encodes observations by
disregarding reconstruction error minimization, and aiming instead for highest
information inclusion. The encoder autonomously selects observations online to
train on, in order to maximize code sparsity. As the dictionary size increases,
the encoder produces increasingly larger inputs for the neural network: this is
addressed by a variation of the Exponential Natural Evolution Strategies
algorithm which adapts its probability distribution dimensionality along the
run. We test our system on a selection of Atari games using tiny neural
networks of only 6 to 18 neurons (depending on the game's controls). These are
still capable of achieving results comparable---and occasionally superior---to
state-of-the-art techniques which use two orders of magnitude more neurons.
|
Factorized Adversarial Networks for Unsupervised Domain Adaptation
|
In this paper, we propose Factorized Adversarial Networks (FAN) to solve
unsupervised domain adaptation problems for image classification tasks. Our
networks map the data distribution into a latent feature space, which is
factorized into a domain-specific subspace that contains domain-specific
characteristics and a task-specific subspace that retains category information,
for both source and target domains, respectively. Unsupervised domain
adaptation is achieved by adversarial training to minimize the discrepancy
between the distributions of two task-specific subspaces from source and target
domains. We demonstrate that the proposed approach outperforms state-of-the-art
methods on multiple benchmark datasets used in the literature for unsupervised
domain adaptation. Furthermore, we collect two real-world tagging datasets that
are much larger than existing benchmark datasets, and get significant
improvement upon baselines, proving the practical value of our approach.
|
A General Framework for Bandit Problems Beyond Cumulative Objectives
|
The stochastic multi-armed bandit (MAB) problem is a common model for
sequential decision problems. In the standard setup, a decision maker has to
choose at every instant between several competing arms, each of them provides a
scalar random variable, referred to as a "reward." Nearly all research on this
topic considers the total cumulative reward as the criterion of interest. This
work focuses on other natural objectives that cannot be cast as a sum over
rewards, but rather more involved functions of the reward stream. Unlike the
case of cumulative criteria, in the problems we study here the oracle policy,
that knows the problem parameters a priori and is used to "center" the regret,
is not trivial. We provide a systematic approach to such problems, and derive
general conditions under which the oracle policy is sufficiently tractable to
facilitate the design of optimism-based (upper confidence bound) learning
policies. These conditions elucidate an interesting interplay between the arm
reward distributions and the performance metric. Our main findings are
illustrated for several commonly used objectives such as conditional
value-at-risk, mean-variance trade-offs, Sharpe-ratio, and more.
|
FlowNet3D: Learning Scene Flow in 3D Point Clouds
|
Many applications in robotics and human-computer interaction can benefit from
understanding 3D motion of points in a dynamic environment, widely noted as
scene flow. While most previous methods focus on stereo and RGB-D images as
input, few try to estimate scene flow directly from point clouds. In this work,
we propose a novel deep neural network named $FlowNet3D$ that learns scene flow
from point clouds in an end-to-end fashion. Our network simultaneously learns
deep hierarchical features of point clouds and flow embeddings that represent
point motions, supported by two newly proposed learning layers for point sets.
We evaluate the network on both challenging synthetic data from FlyingThings3D
and real Lidar scans from KITTI. Trained on synthetic data only, our network
successfully generalizes to real scans, outperforming various baselines and
showing competitive results to the prior art. We also demonstrate two
applications of our scene flow output (scan registration and motion
segmentation) to show its potential wide use cases.
|
Composite Marginal Likelihood Methods for Random Utility Models
|
We propose a novel and flexible
rank-breaking-then-composite-marginal-likelihood (RBCML) framework for learning
random utility models (RUMs), which include the Plackett-Luce model. We
characterize conditions for the objective function of RBCML to be strictly
log-concave by proving that strict log-concavity is preserved under convolution
and marginalization. We characterize necessary and sufficient conditions for
RBCML to satisfy consistency and asymptotic normality. Experiments on synthetic
data show that RBCML for Gaussian RUMs achieves better statistical efficiency
and computational efficiency than the state-of-the-art algorithm and our RBCML
for the Plackett-Luce model provides flexible tradeoffs between running time
and statistical efficiency.
|
Analysis of DAWNBench, a Time-to-Accuracy Machine Learning Performance
Benchmark
|
Researchers have proposed hardware, software, and algorithmic optimizations
to improve the computational performance of deep learning. While some of these
optimizations perform the same operations faster (e.g., increasing GPU clock
speed), many others modify the semantics of the training procedure (e.g.,
reduced precision), and can impact the final model's accuracy on unseen data.
Due to a lack of standard evaluation criteria that considers these trade-offs,
it is difficult to directly compare these optimizations. To address this
problem, we recently introduced DAWNBench, a benchmark competition focused on
end-to-end training time to achieve near-state-of-the-art accuracy on an unseen
dataset---a combined metric called time-to-accuracy (TTA). In this work, we
analyze the entries from DAWNBench, which received optimized submissions from
multiple industrial groups, to investigate the behavior of TTA as a metric as
well as trends in the best-performing entries. We show that TTA has a low
coefficient of variation and that models optimized for TTA generalize nearly as
well as those trained using standard methods. Additionally, even though
DAWNBench entries were able to train ImageNet models in under 3 minutes, we
find they still underutilize hardware capabilities such as Tensor Cores.
Furthermore, we find that distributed entries can spend more than half of their
time on communication. We show similar findings with entries to the MLPERF v0.5
benchmark.
|
Embedding Logical Queries on Knowledge Graphs
|
Learning low-dimensional embeddings of knowledge graphs is a powerful
approach used to predict unobserved or missing edges between entities. However,
an open challenge in this area is developing techniques that can go beyond
simple edge prediction and handle more complex logical queries, which might
involve multiple unobserved edges, entities, and variables. For instance, given
an incomplete biological knowledge graph, we might want to predict "em what
drugs are likely to target proteins involved with both diseases X and Y?" -- a
query that requires reasoning about all possible proteins that {\em might}
interact with diseases X and Y. Here we introduce a framework to efficiently
make predictions about conjunctive logical queries -- a flexible but tractable
subset of first-order logic -- on incomplete knowledge graphs. In our approach,
we embed graph nodes in a low-dimensional space and represent logical operators
as learned geometric operations (e.g., translation, rotation) in this embedding
space. By performing logical operations within a low-dimensional embedding
space, our approach achieves a time complexity that is linear in the number of
query variables, compared to the exponential complexity required by a naive
enumeration-based approach. We demonstrate the utility of this framework in two
application studies on real-world datasets with millions of relations:
predicting logical relationships in a network of drug-gene-disease interactions
and in a graph-based representation of social interactions derived from a
popular web forum.
|
EigenNetworks
|
Many applications donot have the benefit of the laws of physics to derive
succinct descriptive models for observed data. In alternative,
interdependencies among $N$ time series $\{ x_{nk}, k>0 \}_{n=1}^{N}$ are
nowadays often captured by a graph or network $G$ that in practice may be very
large. The network itself may change over time as well (i.e., as $G_k$).
Tracking brute force the changes of time varying networks presents major
challenges, including the associated computational problems. Further, a large
set of networks may not lend itself to useful analysis. This paper approximates
the time varying networks $\left\{G_k\right\}$ as weighted linear combinations
of eigennetworks. The eigennetworks are fixed building blocks that are
estimated by first learning the time series of graphs $G_k$ from the data $\{
x_{nk}, k>0 \}_{n=1}^{N}$, followed by a Principal Network Analysis procedure.
The weights of the eigennetwork representation are eigenfeatures and the time
varying networks $\left\{G_k\right\}$ describe a trajectory in eigennetwork
space. These eigentrajectories should be smooth since the networks $G_k$ vary
at a much slower rate than the data $x_{nk}$, except when structural network
shifts occur reflecting potentially an abrupt change in the underlying
application and sources of the data. Algorithms for learning the time series of
graphs $\left\{G_k\right\}$, deriving the eigennetworks, eigenfeatures and
eigentrajectories, and detecting changepoints are presented. Experiments on
simulated data and with two real time series data (a voting record of the US
senate and genetic expression data for the \textit{Drosophila Melanogaster} as
it goes through its life cycle) demonstrate the performance of the learning and
provide interesting interpretations of the eigennetworks.
|
Informative Gene Selection for Microarray Classification via Adaptive
Elastic Net with Conditional Mutual Information
|
Due to the advantage of achieving a better performance under weak
regularization, elastic net has attracted wide attention in statistics, machine
learning, bioinformatics, and other fields. In particular, a variation of the
elastic net, adaptive elastic net (AEN), integrates the adaptive grouping
effect. In this paper, we aim to develop a new algorithm: Adaptive Elastic Net
with Conditional Mutual Information (AEN-CMI) that further improves AEN by
incorporating conditional mutual information into the gene selection process.
We apply this new algorithm to screen significant genes for two kinds of
cancers: colon cancer and leukemia. Compared with other algorithms including
Support Vector Machine, Classic Elastic Net and Adaptive Elastic Net, the
proposed algorithm, AEN-CMI, obtains the best classification performance using
the least number of genes.
|
Understanding Regularized Spectral Clustering via Graph Conductance
|
This paper uses the relationship between graph conductance and spectral
clustering to study (i) the failures of spectral clustering and (ii) the
benefits of regularization. The explanation is simple. Sparse and stochastic
graphs create a lot of small trees that are connected to the core of the graph
by only one edge. Graph conductance is sensitive to these noisy `dangling
sets'. Spectral clustering inherits this sensitivity. The second part of the
paper starts from a previously proposed form of regularized spectral clustering
and shows that it is related to the graph conductance on a `regularized graph'.
We call the conductance on the regularized graph CoreCut. Based upon previous
arguments that relate graph conductance to spectral clustering (e.g. Cheeger
inequality), minimizing CoreCut relaxes to regularized spectral clustering.
Simple inspection of CoreCut reveals why it is less sensitive to small cuts in
the graph. Together, these results show that unbalanced partitions from
spectral clustering can be understood as overfitting to noise in the periphery
of a sparse and stochastic graph. Regularization fixes this overfitting. In
addition to this statistical benefit, these results also demonstrate how
regularization can improve the computational speed of spectral clustering. We
provide simulations and data examples to illustrate these results.
|
PAC-learning in the presence of evasion adversaries
|
The existence of evasion attacks during the test phase of machine learning
algorithms represents a significant challenge to both their deployment and
understanding. These attacks can be carried out by adding imperceptible
perturbations to inputs to generate adversarial examples and finding effective
defenses and detectors has proven to be difficult. In this paper, we step away
from the attack-defense arms race and seek to understand the limits of what can
be learned in the presence of an evasion adversary. In particular, we extend
the Probably Approximately Correct (PAC)-learning framework to account for the
presence of an adversary. We first define corrupted hypothesis classes which
arise from standard binary hypothesis classes in the presence of an evasion
adversary and derive the Vapnik-Chervonenkis (VC)-dimension for these, denoted
as the adversarial VC-dimension. We then show that sample complexity upper
bounds from the Fundamental Theorem of Statistical learning can be extended to
the case of evasion adversaries, where the sample complexity is controlled by
the adversarial VC-dimension. We then explicitly derive the adversarial
VC-dimension for halfspace classifiers in the presence of a sample-wise
norm-constrained adversary of the type commonly studied for evasion attacks and
show that it is the same as the standard VC-dimension, closing an open
question. Finally, we prove that the adversarial VC-dimension can be either
larger or smaller than the standard VC-dimension depending on the hypothesis
class and adversary, making it an interesting object of study in its own right.
|
An Explainable Adversarial Robustness Metric for Deep Learning Neural
Networks
|
Deep Neural Networks(DNN) have excessively advanced the field of computer
vision by achieving state of the art performance in various vision tasks. These
results are not limited to the field of vision but can also be seen in speech
recognition and machine translation tasks. Recently, DNNs are found to poorly
fail when tested with samples that are crafted by making imperceptible changes
to the original input images. This causes a gap between the validation and
adversarial performance of a DNN. An effective and generalizable robustness
metric for evaluating the performance of DNN on these adversarial inputs is
still missing from the literature. In this paper, we propose Noise Sensitivity
Score (NSS), a metric that quantifies the performance of a DNN on a specific
input under different forms of fix-directional attacks. An insightful
mathematical explanation is provided for deeply understanding the proposed
metric. By leveraging the NSS, we also proposed a skewness based dataset
robustness metric for evaluating a DNN's adversarial performance on a given
dataset. Extensive experiments using widely used state of the art architectures
along with popular classification datasets, such as MNIST, CIFAR-10, CIFAR-100,
and ImageNet, are used to validate the effectiveness and generalization of our
proposed metrics. Instead of simply measuring a DNN's adversarial robustness in
the input domain, as previous works, the proposed NSS is built on top of
insightful mathematical understanding of the adversarial attack and gives a
more explicit explanation of the robustness.
|
Forecasting Crime with Deep Learning
|
The objective of this work is to take advantage of deep neural networks in
order to make next day crime count predictions in a fine-grain city partition.
We make predictions using Chicago and Portland crime data, which is augmented
with additional datasets covering weather, census data, and public
transportation. The crime counts are broken into 10 bins and our model predicts
the most likely bin for a each spatial region at a daily level. We train this
data using increasingly complex neural network structures, including variations
that are suited to the spatial and temporal aspects of the crime prediction
problem. With our best model we are able to predict the correct bin for overall
crime count with 75.6% and 65.3% accuracy for Chicago and Portland,
respectively. The results show the efficacy of neural networks for the
prediction problem and the value of using external datasets in addition to
standard crime data.
|
A Primer on Causal Analysis
|
We provide a conceptual map to navigate causal analysis problems. Focusing on
the case of discrete random variables, we consider the case of causal effect
estimation from observational data. The presented approaches apply also to
continuous variables, but the issue of estimation becomes more complex. We then
introduce the four schools of thought for causal analysis
|
Attention Based Fully Convolutional Network for Speech Emotion
Recognition
|
Speech emotion recognition is a challenging task for three main reasons: 1)
human emotion is abstract, which means it is hard to distinguish; 2) in
general, human emotion can only be detected in some specific moments during a
long utterance; 3) speech data with emotional labeling is usually limited. In
this paper, we present a novel attention based fully convolutional network for
speech emotion recognition. We employ fully convolutional network as it is able
to handle variable-length speech, free of the demand of segmentation to keep
critical information not lost. The proposed attention mechanism can make our
model be aware of which time-frequency region of speech spectrogram is more
emotion-relevant. Considering limited data, the transfer learning is also
adapted to improve the accuracy. Especially, it's interesting to observe
obvious improvement obtained with natural scene image based pre-trained model.
Validated on the publicly available IEMOCAP corpus, the proposed model
outperformed the state-of-the-art methods with a weighted accuracy of 70.4% and
an unweighted accuracy of 63.9% respectively.
|
The universal approximation power of finite-width deep ReLU networks
|
We show that finite-width deep ReLU neural networks yield rate-distortion
optimal approximation (B\"olcskei et al., 2018) of polynomials, windowed
sinusoidal functions, one-dimensional oscillatory textures, and the Weierstrass
function, a fractal function which is continuous but nowhere differentiable.
Together with their recently established universal approximation property of
affine function systems (B\"olcskei et al., 2018), this shows that deep neural
networks approximate vastly different signal structures generated by the affine
group, the Weyl-Heisenberg group, or through warping, and even certain
fractals, all with approximation error decaying exponentially in the number of
neurons. We also prove that in the approximation of sufficiently smooth
functions finite-width deep networks require strictly smaller connectivity than
finite-depth wide networks.
|
Combining Multiple Algorithms in Classifier Ensembles using Generalized
Mixture Functions
|
Classifier ensembles are pattern recognition structures composed of a set of
classification algorithms (members), organized in a parallel way, and a
combination method with the aim of increasing the classification accuracy of a
classification system. In this study, we investigate the application of a
generalized mixture (GM) functions as a new approach for providing an efficient
combination procedure for these systems through the use of dynamic weights in
the combination process. Therefore, we present three GM functions to be applied
as a combination method. The main advantage of these functions is that they can
define dynamic weights at the member outputs, making the combination process
more efficient. In order to evaluate the feasibility of the proposed approach,
an empirical analysis is conducted, applying classifier ensembles to 25
different classification data sets. In this analysis, we compare the use of the
proposed approaches to ensembles using traditional combination methods as well
as the state-of-the-art ensemble methods. Our findings indicated gains in terms
of performance when comparing the proposed approaches to the traditional ones
as well as comparable results with the state-of-the-art methods.
|
Semi-Supervised Clustering with Neural Networks
|
Clustering using neural networks has recently demonstrated promising
performance in machine learning and computer vision applications. However, the
performance of current approaches is limited either by unsupervised learning or
their dependence on large set of labeled data samples. In this paper, we
propose ClusterNet that uses pairwise semantic constraints from very few
labeled data samples (<5% of total data) and exploits the abundant unlabeled
data to drive the clustering approach. We define a new loss function that uses
pairwise semantic similarity between objects combined with constrained k-means
clustering to efficiently utilize both labeled and unlabeled data in the same
framework. The proposed network uses convolution autoencoder to learn a latent
representation that groups data into k specified clusters, while also learning
the cluster centers simultaneously. We evaluate and compare the performance of
ClusterNet on several datasets and state of the art deep clustering approaches.
|
Deep Mixed Effect Model using Gaussian Processes: A Personalized and
Reliable Prediction for Healthcare
|
We present a personalized and reliable prediction model for healthcare, which
can provide individually tailored medical services such as diagnosis, disease
treatment, and prevention. Our proposed framework targets at making
personalized and reliable predictions from time-series data, such as Electronic
Health Records (EHR), by modeling two complementary components: i) a shared
component that captures global trend across diverse patients and ii) a
patient-specific component that models idiosyncratic variability for each
patient. To this end, we propose a composite model of a deep neural network to
learn complex global trends from the large number of patients, and Gaussian
Processes (GP) to probabilistically model individual time-series given
relatively small number of visits per patient. We evaluate our model on diverse
and heterogeneous tasks from EHR datasets and show practical advantages over
standard time-series deep models such as pure Recurrent Neural Network (RNN).
|
A Visual Quality Index for Fuzzy C-Means
|
Cluster analysis is widely used in the areas of machine learning and data
mining. Fuzzy clustering is a particular method that considers that a data
point can belong to more than one cluster. Fuzzy clustering helps obtain
flexible clusters, as needed in such applications as text categorization. The
performance of a clustering algorithm critically depends on the number of
clusters, and estimating the optimal number of clusters is a challenging task.
Quality indices help estimate the optimal number of clusters. However, there is
no quality index that can obtain an accurate number of clusters for different
datasets. Thence, in this paper, we propose a new cluster quality index
associated with a visual, graph-based solution that helps choose the optimal
number of clusters in fuzzy partitions. Moreover, we validate our theoretical
results through extensive comparison experiments against state-of-the-art
quality indices on a variety of numerical real-world and artificial datasets.
|
GuideR: a guided separate-and-conquer rule learning in classification,
regression, and survival settings
|
This article presents GuideR, a user-guided rule induction algorithm, which
overcomes the largest limitation of the existing methods-the lack of the
possibility to introduce user's preferences or domain knowledge to the rule
learning process. Automatic selection of attributes and attribute ranges often
leads to the situation in which resulting rules do not contain interesting
information. We propose an induction algorithm which takes into account user's
requirements. Our method uses the sequential covering approach and is suitable
for classification, regression, and survival analysis problems. The
effectiveness of the algorithm in all these tasks has been verified
experimentally, confirming guided rule induction to be a powerful data analysis
tool.
|
Discovering and Removing Exogenous State Variables and Rewards for
Reinforcement Learning
|
Exogenous state variables and rewards can slow down reinforcement learning by
injecting uncontrolled variation into the reward signal. We formalize exogenous
state variables and rewards and identify conditions under which an MDP with
exogenous state can be decomposed into an exogenous Markov Reward Process
involving only the exogenous state+reward and an endogenous Markov Decision
Process defined with respect to only the endogenous rewards. We also derive a
variance-covariance condition under which Monte Carlo policy evaluation on the
endogenous MDP is accelerated compared to using the full MDP. Similar speedups
are likely to carry over to all RL algorithms. We develop two algorithms for
discovering the exogenous variables and test them on several MDPs. Results show
that the algorithms are practical and can significantly speed up reinforcement
learning.
|
Accelerated Randomized Coordinate Descent Algorithms for Stochastic
Optimization and Online Learning
|
We propose accelerated randomized coordinate descent algorithms for
stochastic optimization and online learning. Our algorithms have significantly
less per-iteration complexity than the known accelerated gradient algorithms.
The proposed algorithms for online learning have better regret performance than
the known randomized online coordinate descent algorithms. Furthermore, the
proposed algorithms for stochastic optimization exhibit as good convergence
rates as the best known randomized coordinate descent algorithms. We also show
simulation results to demonstrate performance of the proposed algorithms.
|
Layer rotation: a surprisingly powerful indicator of generalization in
deep networks?
|
Our work presents extensive empirical evidence that layer rotation, i.e. the
evolution across training of the cosine distance between each layer's weight
vector and its initialization, constitutes an impressively consistent indicator
of generalization performance. In particular, larger cosine distances between
final and initial weights of each layer consistently translate into better
generalization performance of the final model. Interestingly, this relation
admits a network independent optimum: training procedures during which all
layers' weights reach a cosine distance of 1 from their initialization
consistently outperform other configurations -by up to 30% test accuracy.
Moreover, we show that layer rotations are easily monitored and controlled
(helpful for hyperparameter tuning) and potentially provide a unified framework
to explain the impact of learning rate tuning, weight decay, learning rate
warmups and adaptive gradient methods on generalization and training speed. In
an attempt to explain the surprising properties of layer rotation, we show on a
1-layer MLP trained on MNIST that layer rotation correlates with the degree to
which features of intermediate layers have been trained.
|
Training Generative Reversible Networks
|
Generative models with an encoding component such as autoencoders currently
receive great interest. However, training of autoencoders is typically
complicated by the need to train a separate encoder and decoder model that have
to be enforced to be reciprocal to each other. To overcome this problem,
by-design reversible neural networks (RevNets) had been previously used as
generative models either directly optimizing the likelihood of the data under
the model or using an adversarial approach on the generated data. Here, we
instead investigate their performance using an adversary on the latent space in
the adversarial autoencoder framework. We investigate the generative
performance of RevNets on the CelebA dataset, showing that generative RevNets
can generate coherent faces with similar quality as Variational Autoencoders.
This first attempt to use RevNets inside the adversarial autoencoder framework
slightly underperformed relative to recent advanced generative models using an
autoencoder component on CelebA, but this gap may diminish with further
optimization of the training setup of generative RevNets. In addition to the
experiments on CelebA, we show a proof-of-principle experiment on the MNIST
dataset suggesting that adversary-free trained RevNets can discover meaningful
latent dimensions without pre-specifying the number of dimensions of the latent
sampling distribution. In summary, this study shows that RevNets can be
employed in different generative training settings.
Source code for this study is at
https://github.com/robintibor/generative-reversible
|
BOCK : Bayesian Optimization with Cylindrical Kernels
|
A major challenge in Bayesian Optimization is the boundary issue (Swersky,
2017) where an algorithm spends too many evaluations near the boundary of its
search space. In this paper, we propose BOCK, Bayesian Optimization with
Cylindrical Kernels, whose basic idea is to transform the ball geometry of the
search space using a cylindrical transformation. Because of the transformed
geometry, the Gaussian Process-based surrogate model spends less budget
searching near the boundary, while concentrating its efforts relatively more
near the center of the search region, where we expect the solution to be
located. We evaluate BOCK extensively, showing that it is not only more
accurate and efficient, but it also scales successfully to problems with a
dimensionality as high as 500. We show that the better accuracy and scalability
of BOCK even allows optimizing modestly sized neural network layers, as well as
neural network hyperparameters.
|
Deep Gaussian Processes with Convolutional Kernels
|
Deep Gaussian processes (DGPs) provide a Bayesian non-parametric alternative
to standard parametric deep learning models. A DGP is formed by stacking
multiple GPs resulting in a well-regularized composition of functions. The
Bayesian framework that equips the model with attractive properties, such as
implicit capacity control and predictive uncertainty, makes it at the same time
challenging to combine with a convolutional structure. This has hindered the
application of DGPs in computer vision tasks, an area where deep parametric
models (i.e. CNNs) have made breakthroughs. Standard kernels used in DGPs such
as radial basis functions (RBFs) are insufficient for handling pixel
variability in raw images. In this paper, we build on the recent convolutional
GP to develop Convolutional DGP (CDGP) models which effectively capture image
level features through the use of convolution kernels, therefore opening up the
way for applying DGPs to computer vision tasks. Our model learns local spatial
influence and outperforms strong GP based baselines on multi-class image
classification. We also consider various constructions of convolution kernel
over the image patches, analyze the computational trade-offs and provide an
efficient framework for convolutional DGP models. The experimental results on
image data such as MNIST, rectangles-image, CIFAR10 and Caltech101 demonstrate
the effectiveness of the proposed approaches.
|
Towards Understanding Acceleration Tradeoff between Momentum and
Asynchrony in Nonconvex Stochastic Optimization
|
Asynchronous momentum stochastic gradient descent algorithms (Async-MSGD) is
one of the most popular algorithms in distributed machine learning. However,
its convergence properties for these complicated nonconvex problems is still
largely unknown, because of the current technical limit. Therefore, in this
paper, we propose to analyze the algorithm through a simpler but nontrivial
nonconvex problem - streaming PCA, which helps us to understand Aync-MSGD
better even for more general problems. Specifically, we establish the
asymptotic rate of convergence of Async-MSGD for streaming PCA by diffusion
approximation. Our results indicate a fundamental tradeoff between asynchrony
and momentum: To ensure convergence and acceleration through asynchrony, we
have to reduce the momentum (compared with Sync-MSGD). To the best of our
knowledge, this is the first theoretical attempt on understanding Async-MSGD
for distributed nonconvex stochastic optimization. Numerical experiments on
both streaming PCA and training deep neural networks are provided to support
our findings for Async-MSGD.
|
On Latent Distributions Without Finite Mean in Generative Models
|
We investigate the properties of multidimensional probability distributions
in the context of latent space prior distributions of implicit generative
models. Our work revolves around the phenomena arising while decoding linear
interpolations between two random latent vectors -- regions of latent space in
close proximity to the origin of the space are sampled causing distribution
mismatch. We show that due to the Central Limit Theorem, this region is almost
never sampled during the training process. As a result, linear interpolations
may generate unrealistic data and their usage as a tool to check quality of the
trained model is questionable. We propose to use multidimensional Cauchy
distribution as the latent prior. Cauchy distribution does not satisfy the
assumptions of the CLT and has a number of properties that allow it to work
well in conjunction with linear interpolations. We also provide two general
methods of creating non-linear interpolations that are easily applicable to a
large family of common latent distributions. Finally we empirically analyze the
quality of data generated from low-probability-mass regions for the DCGAN model
on the CelebA dataset.
|
A Projection Method for Metric-Constrained Optimization
|
We outline a new approach for solving optimization problems which enforce
triangle inequalities on output variables. We refer to this as
metric-constrained optimization, and give several examples where problems of
this form arise in machine learning applications and theoretical approximation
algorithms for graph clustering. Although these problem are interesting from a
theoretical perspective, they are challenging to solve in practice due to the
high memory requirement of black-box solvers. In order to address this
challenge we first prove that the metric-constrained linear program relaxation
of correlation clustering is equivalent to a special case of the metric
nearness problem. We then developed a general solver for metric-constrained
linear and quadratic programs by generalizing and improving a simple projection
algorithm originally developed for metric nearness. We give several novel
approximation guarantees for using our framework to find lower bounds for
optimal solutions to several challenging graph clustering problems. We also
demonstrate the power of our framework by solving optimizing problems involving
up to 10^{8} variables and 10^{11} constraints.
|
Machine Learning for Yield Curve Feature Extraction: Application to
Illiquid Corporate Bonds (Preliminary Draft)
|
This paper studies the application of machine learning in extracting the
market implied features from historical risk neutral corporate bond yields. We
consider the example of a hypothetical illiquid fixed income market. After
choosing a surrogate liquid market, we apply the Denoising Autoencoder
algorithm from the field of computer vision and pattern recognition to learn
the features of the missing yield parameters from the historically implied data
of the instruments traded in the chosen liquid market. The results of the
trained machine learning algorithm are compared with the outputs of a point in-
time 2 dimensional interpolation algorithm known as the Thin Plate Spline.
Finally, the performances of the two algorithms are compared.
|
Disease Prediction using Graph Convolutional Networks: Application to
Autism Spectrum Disorder and Alzheimer's Disease
|
Graphs are widely used as a natural framework that captures interactions
between individual elements represented as nodes in a graph. In medical
applications, specifically, nodes can represent individuals within a
potentially large population (patients or healthy controls) accompanied by a
set of features, while the graph edges incorporate associations between
subjects in an intuitive manner. This representation allows to incorporate the
wealth of imaging and non-imaging information as well as individual subject
features simultaneously in disease classification tasks. Previous graph-based
approaches for supervised or unsupervised learning in the context of disease
prediction solely focus on pairwise similarities between subjects, disregarding
individual characteristics and features, or rather rely on subject-specific
imaging feature vectors and fail to model interactions between them. In this
paper, we present a thorough evaluation of a generic framework that leverages
both imaging and non-imaging information and can be used for brain analysis in
large populations. This framework exploits Graph Convolutional Networks (GCNs)
and involves representing populations as a sparse graph, where its nodes are
associated with imaging-based feature vectors, while phenotypic information is
integrated as edge weights. The extensive evaluation explores the effect of
each individual component of this framework on disease prediction performance
and further compares it to different baselines. The framework performance is
tested on two large datasets with diverse underlying data, ABIDE and ADNI, for
the prediction of Autism Spectrum Disorder and conversion to Alzheimer's
disease, respectively. Our analysis shows that our novel framework can improve
over state-of-the-art results on both databases, with 70.4% classification
accuracy for ABIDE and 80.0% for ADNI.
|
A Machine Learning Framework for Stock Selection
|
This paper demonstrates how to apply machine learning algorithms to
distinguish good stocks from the bad stocks. To this end, we construct 244
technical and fundamental features to characterize each stock, and label stocks
according to their ranking with respect to the return-to-volatility ratio.
Algorithms ranging from traditional statistical learning methods to recently
popular deep learning method, e.g. Logistic Regression (LR), Random Forest
(RF), Deep Neural Network (DNN), and the Stacking, are trained to solve the
classification task. Genetic Algorithm (GA) is also used to implement feature
selection. The effectiveness of the stock selection strategy is validated in
Chinese stock market in both statistical and practical aspects, showing that:
1) Stacking outperforms other models reaching an AUC score of 0.972; 2) Genetic
Algorithm picks a subset of 114 features and the prediction performances of all
models remain almost unchanged after the selection procedure, which suggests
some features are indeed redundant; 3) LR and DNN are radical models; RF is
risk-neutral model; Stacking is somewhere between DNN and RF. 4) The portfolios
constructed by our models outperform market average in back tests.
|
Neural-Kernelized Conditional Density Estimation
|
Conditional density estimation is a general framework for solving various
problems in machine learning. Among existing methods, non-parametric and/or
kernel-based methods are often difficult to use on large datasets, while
methods based on neural networks usually make restrictive parametric
assumptions on the probability densities. Here, we propose a novel method for
estimating the conditional density based on score matching. In contrast to
existing methods, we employ scalable neural networks, but do not make explicit
parametric assumptions on densities. The key challenge in applying score
matching to neural networks is computation of the first- and second-order
derivatives of a model for the log-density. We tackle this challenge by
developing a new neural-kernelized approach, which can be applied on large
datasets with stochastic gradient descent, while the reproducing kernels allow
for easy computation of the derivatives needed in score matching. We show that
the neural-kernelized function approximator has universal approximation
capability and that our method is consistent in conditional density estimation.
We numerically demonstrate that our method is useful in high-dimensional
conditional density estimation, and compares favourably with existing methods.
Finally, we prove that the proposed method has interesting connections to two
probabilistically principled frameworks of representation learning: Nonlinear
sufficient dimension reduction and nonlinear independent component analysis.
|
Evidential Deep Learning to Quantify Classification Uncertainty
|
Deterministic neural nets have been shown to learn effective predictors on a
wide range of machine learning problems. However, as the standard approach is
to train the network to minimize a prediction loss, the resultant model remains
ignorant to its prediction confidence. Orthogonally to Bayesian neural nets
that indirectly infer prediction uncertainty through weight uncertainties, we
propose explicit modeling of the same using the theory of subjective logic. By
placing a Dirichlet distribution on the class probabilities, we treat
predictions of a neural net as subjective opinions and learn the function that
collects the evidence leading to these opinions by a deterministic neural net
from data. The resultant predictor for a multi-class classification problem is
another Dirichlet distribution whose parameters are set by the continuous
output of a neural net. We provide a preliminary analysis on how the
peculiarities of our new loss function drive improved uncertainty estimation.
We observe that our method achieves unprecedented success on detection of
out-of-distribution queries and endurance against adversarial perturbations.
|
Cycle-Consistent Adversarial Learning as Approximate Bayesian Inference
|
We formalize the problem of learning interdomain correspondences in the
absence of paired data as Bayesian inference in a latent variable model (LVM),
where one seeks the underlying hidden representations of entities from one
domain as entities from the other domain. First, we introduce implicit latent
variable models, where the prior over hidden representations can be specified
flexibly as an implicit distribution. Next, we develop a new variational
inference (VI) algorithm for this model based on minimization of the symmetric
Kullback-Leibler (KL) divergence between a variational joint and the exact
joint distribution. Lastly, we demonstrate that the state-of-the-art
cycle-consistent adversarial learning (CYCLEGAN) models can be derived as a
special case within our proposed VI framework, thus establishing its connection
to approximate Bayesian inference methods.
|
Mix&Match - Agent Curricula for Reinforcement Learning
|
We introduce Mix&Match (M&M) - a training framework designed to facilitate
rapid and effective learning in RL agents, especially those that would be too
slow or too challenging to train otherwise. The key innovation is a procedure
that allows us to automatically form a curriculum over agents. Through such a
curriculum we can progressively train more complex agents by, effectively,
bootstrapping from solutions found by simpler agents. In contradistinction to
typical curriculum learning approaches, we do not gradually modify the tasks or
environments presented, but instead use a process to gradually alter how the
policy is represented internally. We show the broad applicability of our method
by demonstrating significant performance gains in three different experimental
setups: (1) We train an agent able to control more than 700 actions in a
challenging 3D first-person task; using our method to progress through an
action-space curriculum we achieve both faster training and better final
performance than one obtains using traditional methods. (2) We further show
that M&M can be used successfully to progress through a curriculum of
architectural variants defining an agents internal state. (3) Finally, we
illustrate how a variant of our method can be used to improve agent performance
in a multitask setting.
|
Adaptive System Identification Using LMS Algorithm Integrated with
Evolutionary Computation
|
System identification is an exceptionally expansive topic and of remarkable
significance in the discipline of signal processing and communication. Our goal
in this paper is to show how simple adaptive FIR and IIR filters can be used in
system modeling and demonstrating the application of adaptive system
identification. The main objective of our research is to study the LMS
algorithm and its improvement by the genetic search approach, namely, LMS-GA,
to search the multi-modal error surface of the IIR filter to avoid local minima
and finding the optimal weight vector when only measured or estimated data are
available. Convergence analysis of the LMS algorithm in the case of coloured
input signal, i.e., correlated input signal is demonstrated on adaptive FIR
filter via power spectral density of the input signals and Fourier transform of
the autocorrelation matrix of the input signal. Simulations have been carried
out on adaptive filtering of FIR and IIR filters and tested on white and
coloured input signals to validate the powerfulness of the genetic-based LMS
algorithm.
|
Gradient-based Filter Design for the Dual-tree Wavelet Transform
|
The wavelet transform has seen success when incorporated into neural network
architectures, such as in wavelet scattering networks. More recently, it has
been shown that the dual-tree complex wavelet transform can provide better
representations than the standard transform. With this in mind, we extend our
previous method for learning filters for the 1D and 2D wavelet transforms into
the dual-tree domain. We show that with few modifications to our original
model, we can learn directional filters that leverage the properties of the
dual-tree wavelet transform.
|
Sequential Attend, Infer, Repeat: Generative Modelling of Moving Objects
|
We present Sequential Attend, Infer, Repeat (SQAIR), an interpretable deep
generative model for videos of moving objects. It can reliably discover and
track objects throughout the sequence of frames, and can also generate future
frames conditioning on the current frame, thereby simulating expected motion of
objects. This is achieved by explicitly encoding object presence, locations and
appearances in the latent variables of the model. SQAIR retains all strengths
of its predecessor, Attend, Infer, Repeat (AIR, Eslami et. al., 2016),
including learning in an unsupervised manner, and addresses its shortcomings.
We use a moving multi-MNIST dataset to show limitations of AIR in detecting
overlapping or partially occluded objects, and show how SQAIR overcomes them by
leveraging temporal consistency of objects. Finally, we also apply SQAIR to
real-world pedestrian CCTV data, where it learns to reliably detect, track and
generate walking pedestrians with no supervision.
|
Stochastic Gradient Descent on Separable Data: Exact Convergence with a
Fixed Learning Rate
|
Stochastic Gradient Descent (SGD) is a central tool in machine learning. We
prove that SGD converges to zero loss, even with a fixed (non-vanishing)
learning rate - in the special case of homogeneous linear classifiers with
smooth monotone loss functions, optimized on linearly separable data. Previous
works assumed either a vanishing learning rate, iterate averaging, or loss
assumptions that do not hold for monotone loss functions used for
classification, such as the logistic loss. We prove our result on a fixed
dataset, both for sampling with or without replacement. Furthermore, for
logistic loss (and similar exponentially-tailed losses), we prove that with SGD
the weight vector converges in direction to the $L_2$ max margin vector as
$O(1/\log(t))$ for almost all separable datasets, and the loss converges as
$O(1/t)$ - similarly to gradient descent. Lastly, we examine the case of a
fixed learning rate proportional to the minibatch size. We prove that in this
case, the asymptotic convergence rate of SGD (with replacement) does not depend
on the minibatch size in terms of epochs, if the support vectors span the data.
These results may suggest an explanation to similar behaviors observed in deep
networks, when trained with SGD.
|
AdaGrad stepsizes: Sharp convergence over nonconvex landscapes
|
Adaptive gradient methods such as AdaGrad and its variants update the
stepsize in stochastic gradient descent on the fly according to the gradients
received along the way; such methods have gained widespread use in large-scale
optimization for their ability to converge robustly, without the need to
fine-tune the stepsize schedule. Yet, the theoretical guarantees to date for
AdaGrad are for online and convex optimization. We bridge this gap by providing
theoretical guarantees for the convergence of AdaGrad for smooth, nonconvex
functions. We show that the norm version of AdaGrad (AdaGrad-Norm) converges to
a stationary point at the $\mathcal{O}(\log(N)/\sqrt{N})$ rate in the
stochastic setting, and at the optimal $\mathcal{O}(1/N)$ rate in the batch
(non-stochastic) setting -- in this sense, our convergence guarantees are
'sharp'. In particular, the convergence of AdaGrad-Norm is robust to the choice
of all hyper-parameters of the algorithm, in contrast to stochastic gradient
descent whose convergence depends crucially on tuning the step-size to the
(generally unknown) Lipschitz smoothness constant and level of stochastic noise
on the gradient. Extensive numerical experiments are provided to corroborate
our theory; moreover, the experiments suggest that the robustness of
AdaGrad-Norm extends to state-of-the-art models in deep learning, without
sacrificing generalization.
|
Perturbative Neural Networks
|
Convolutional neural networks are witnessing wide adoption in computer vision
systems with numerous applications across a range of visual recognition tasks.
Much of this progress is fueled through advances in convolutional neural
network architectures and learning algorithms even as the basic premise of a
convolutional layer has remained unchanged. In this paper, we seek to revisit
the convolutional layer that has been the workhorse of state-of-the-art visual
recognition models. We introduce a very simple, yet effective, module called a
perturbation layer as an alternative to a convolutional layer. The perturbation
layer does away with convolution in the traditional sense and instead computes
its response as a weighted linear combination of non-linearly activated
additive noise perturbed inputs. We demonstrate both analytically and
empirically that this perturbation layer can be an effective replacement for a
standard convolutional layer. Empirically, deep neural networks with
perturbation layers, called Perturbative Neural Networks (PNNs), in lieu of
convolutional layers perform comparably with standard CNNs on a range of visual
datasets (MNIST, CIFAR-10, PASCAL VOC, and ImageNet) with fewer parameters.
|
LSTM Benchmarks for Deep Learning Frameworks
|
This study provides benchmarks for different implementations of LSTM units
between the deep learning frameworks PyTorch, TensorFlow, Lasagne and Keras.
The comparison includes cuDNN LSTMs, fused LSTM variants and less optimized,
but more flexible LSTM implementations. The benchmarks reflect two typical
scenarios for automatic speech recognition, notably continuous speech
recognition and isolated digit recognition. These scenarios cover input
sequences of fixed and variable length as well as the loss functions CTC and
cross entropy. Additionally, a comparison between four different PyTorch
versions is included. The code is available online
https://github.com/stefbraun/rnn_benchmarks.
|
Relational recurrent neural networks
|
Memory-based neural networks model temporal data by leveraging an ability to
remember information for long periods. It is unclear, however, whether they
also have an ability to perform complex relational reasoning with the
information they remember. Here, we first confirm our intuitions that standard
memory architectures may struggle at tasks that heavily involve an
understanding of the ways in which entities are connected -- i.e., tasks
involving relational reasoning. We then improve upon these deficits by using a
new memory module -- a \textit{Relational Memory Core} (RMC) -- which employs
multi-head dot product attention to allow memories to interact. Finally, we
test the RMC on a suite of tasks that may profit from more capable relational
reasoning across sequential information, and show large gains in RL domains
(e.g. Mini PacMan), program evaluation, and language modeling, achieving
state-of-the-art results on the WikiText-103, Project Gutenberg, and GigaWord
datasets.
|
Performance Metric Elicitation from Pairwise Classifier Comparisons
|
Given a binary prediction problem, which performance metric should the
classifier optimize? We address this question by formalizing the problem of
Metric Elicitation. The goal of metric elicitation is to discover the
performance metric of a practitioner, which reflects her innate rewards (costs)
for correct (incorrect) classification. In particular, we focus on eliciting
binary classification performance metrics from pairwise feedback, where a
practitioner is queried to provide relative preference between two classifiers.
By exploiting key geometric properties of the space of confusion matrices, we
obtain provably query efficient algorithms for eliciting linear and
linear-fractional performance metrics. We further show that our method is
robust to feedback and finite sample noise.
|
Relational Deep Reinforcement Learning
|
We introduce an approach for deep reinforcement learning (RL) that improves
upon the efficiency, generalization capacity, and interpretability of
conventional approaches through structured perception and relational reasoning.
It uses self-attention to iteratively reason about the relations between
entities in a scene and to guide a model-free policy. Our results show that in
a novel navigation and planning task called Box-World, our agent finds
interpretable solutions that improve upon baselines in terms of sample
complexity, ability to generalize to more complex scenes than experienced
during training, and overall performance. In the StarCraft II Learning
Environment, our agent achieves state-of-the-art performance on six mini-games
-- surpassing human grandmaster performance on four. By considering
architectural inductive biases, our work opens new directions for overcoming
important, but stubborn, challenges in deep RL.
|
SBAF: A New Activation Function for Artificial Neural Net based
Habitability Classification
|
We explore the efficacy of using a novel activation function in Artificial
Neural Networks (ANN) in characterizing exoplanets into different classes. We
call this Saha-Bora Activation Function (SBAF) as the motivation is derived
from long standing understanding of using advanced calculus in modeling
habitability score of Exoplanets. The function is demonstrated to possess nice
analytical properties and doesn't seem to suffer from local oscillation
problems. The manuscript presents the analytical properties of the activation
function and the architecture implemented on the function. Keywords:
Astroinformatics, Machine Learning, Exoplanets, ANN, Activation Function.
|
Deep Neural Networks with Multi-Branch Architectures Are Less Non-Convex
|
Several recently proposed architectures of neural networks such as ResNeXt,
Inception, Xception, SqueezeNet and Wide ResNet are based on the designing idea
of having multiple branches and have demonstrated improved performance in many
applications. We show that one cause for such success is due to the fact that
the multi-branch architecture is less non-convex in terms of duality gap. The
duality gap measures the degree of intrinsic non-convexity of an optimization
problem: smaller gap in relative value implies lower degree of intrinsic
non-convexity. The challenge is to quantitatively measure the duality gap of
highly non-convex problems such as deep neural networks. In this work, we
provide strong guarantees of this quantity for two classes of network
architectures. For the neural networks with arbitrary activation functions,
multi-branch architecture and a variant of hinge loss, we show that the duality
gap of both population and empirical risks shrinks to zero as the number of
branches increases. This result sheds light on better understanding the power
of over-parametrization where increasing the network width tends to make the
loss surface less non-convex. For the neural networks with linear activation
function and $\ell_2$ loss, we show that the duality gap of empirical risk is
zero. Our two results work for arbitrary depths and adversarial data, while the
analytical techniques might be of independent interest to non-convex
optimization more broadly. Experiments on both synthetic and real-world
datasets validate our results.
|
Pathwise Derivatives Beyond the Reparameterization Trick
|
We observe that gradients computed via the reparameterization trick are in
direct correspondence with solutions of the transport equation in the formalism
of optimal transport. We use this perspective to compute (approximate) pathwise
gradients for probability distributions not directly amenable to the
reparameterization trick: Gamma, Beta, and Dirichlet. We further observe that
when the reparameterization trick is applied to the Cholesky-factorized
multivariate Normal distribution, the resulting gradients are suboptimal in the
sense of optimal transport. We derive the optimal gradients and show that they
have reduced variance in a Gaussian Process regression task. We demonstrate
with a variety of synthetic experiments and stochastic variational inference
tasks that our pathwise gradients are competitive with other methods.
|
Pathwise Derivatives for Multivariate Distributions
|
We exploit the link between the transport equation and derivatives of
expectations to construct efficient pathwise gradient estimators for
multivariate distributions. We focus on two main threads. First, we use null
solutions of the transport equation to construct adaptive control variates that
can be used to construct gradient estimators with reduced variance. Second, we
consider the case of multivariate mixture distributions. In particular we show
how to compute pathwise derivatives for mixtures of multivariate Normal
distributions with arbitrary means and diagonal covariances. We demonstrate in
a variety of experiments in the context of variational inference that our
gradient estimators can outperform other methods, especially in high
dimensions.
|
EEG-GAN: Generative adversarial networks for electroencephalograhic
(EEG) brain signals
|
Generative adversarial networks (GANs) are recently highly successful in
generative applications involving images and start being applied to time series
data. Here we describe EEG-GAN as a framework to generate
electroencephalographic (EEG) brain signals. We introduce a modification to the
improved training of Wasserstein GANs to stabilize training and investigate a
range of architectural choices critical for time series generation (most
notably up- and down-sampling). For evaluation we consider and compare
different metrics such as Inception score, Frechet inception distance and
sliced Wasserstein distance, together showing that our EEG-GAN framework
generated naturalistic EEG examples. It thus opens up a range of new generative
application scenarios in the neuroscientific and neurological context, such as
data augmentation in brain-computer interfacing tasks, EEG super-sampling, or
restoration of corrupted data segments. The possibility to generate signals of
a certain class and/or with specific properties may also open a new avenue for
research into the underlying structure of brain signals.
|
An explicit analysis of the entropic penalty in linear programming
|
Solving linear programs by using entropic penalization has recently attracted
new interest in the optimization community, since this strategy forms the basis
for the fastest-known algorithms for the optimal transport problem, with many
applications in modern large-scale machine learning. Crucial to these
applications has been an analysis of how quickly solutions to the penalized
program approach true optima to the original linear program. More than 20 years
ago, Cominetti and San Mart\'in showed that this convergence is exponentially
fast; however, their proof is asymptotic and does not give any indication of
how accurately the entropic program approximates the original program for any
particular choice of the penalization parameter. We close this long-standing
gap in the literature regarding entropic penalization by giving a new proof of
the exponential convergence, valid for any linear program. Our proof is
non-asymptotic, yields explicit constants, and has the virtue of being
extremely simple. We provide matching lower bounds and show that the entropic
approach does not lead to a near-linear time approximation scheme for the
linear assignment problem.
|
Performance Evaluation of Deep Learning Networks for Semantic
Segmentation of Traffic Stereo-Pair Images
|
Semantic image segmentation is one the most demanding task, especially for
analysis of traffic conditions for self-driving cars. Here the results of
application of several deep learning architectures (PSPNet and ICNet) for
semantic image segmentation of traffic stereo-pair images are presented. The
images from Cityscapes dataset and custom urban images were analyzed as to the
segmentation accuracy and image inference time. For the models pre-trained on
Cityscapes dataset, the inference time was equal in the limits of standard
deviation, but the segmentation accuracy was different for various cities and
stereo channels even. The distributions of accuracy (mean intersection over
union - mIoU) values for each city and channel are asymmetric, long-tailed, and
have many extreme outliers, especially for PSPNet network in comparison to
ICNet network. Some statistical properties of these distributions (skewness,
kurtosis) allow us to distinguish these two networks and open the question
about relations between architecture of deep learning networks and statistical
distribution of the predicted results (mIoU here). The results obtained
demonstrated the different sensitivity of these networks to: (1) the local
street view peculiarities in different cities that should be taken into account
during the targeted fine tuning the models before their practical applications,
(2) the right and left data channels in stereo-pairs. For both networks, the
difference in the predicted results (mIoU here) for the right and left data
channels in stereo-pairs is out of the limits of statistical error in relation
to mIoU values. It means that the traffic stereo pairs can be effectively used
not only for depth calculations (as it is usually used), but also as an
additional data channel that can provide much more information about scene
objects than simple duplication of the same street view images.
|
MRPC: An R package for accurate inference of causal graphs
|
We present MRPC, an R package that learns causal graphs with improved
accuracy over existing packages, such as pcalg and bnlearn. Our algorithm
builds on the powerful PC algorithm, the canonical algorithm in computer
science for learning directed acyclic graphs. The improvement in accuracy
results from online control of the false discovery rate (FDR) that reduces
false positive edges, a more accurate approach to identifying v-structures
(i.e., $T_1 \rightarrow T_2 \leftarrow T_3$), and robust estimation of the
correlation matrix among nodes. For genomic data that contain genotypes and
gene expression for each sample, MRPC incorporates the principle of Mendelian
randomization to orient the edges. Our package can be applied to continuous and
discrete data.
|
Probabilistic Deep Learning using Random Sum-Product Networks
|
The need for consistent treatment of uncertainty has recently triggered
increased interest in probabilistic deep learning methods. However, most
current approaches have severe limitations when it comes to inference, since
many of these models do not even permit to evaluate exact data likelihoods.
Sum-product networks (SPNs), on the other hand, are an excellent architecture
in that regard, as they allow to efficiently evaluate likelihoods, as well as
arbitrary marginalization and conditioning tasks. Nevertheless, SPNs have not
been fully explored as serious deep learning models, likely due to their
special structural requirements, which complicate learning. In this paper, we
make a drastic simplification and use random SPN structures which are trained
in a "classical deep learning manner", i.e. employing automatic
differentiation, SGD, and GPU support. The resulting models, called RAT-SPNs,
yield prediction results comparable to deep neural networks, while still being
interpretable as generative model and maintaining well-calibrated
uncertainties. This property makes them highly robust under missing input
features and enables them to naturally detect outliers and peculiar samples.
|
A Framework for the construction of upper bounds on the number of affine
linear regions of ReLU feed-forward neural networks
|
We present a framework to derive upper bounds on the number of regions that
feed-forward neural networks with ReLU activation functions are affine linear
on. It is based on an inductive analysis that keeps track of the number of such
regions per dimensionality of their images within the layers. More precisely,
the information about the number regions per dimensionality is pushed through
the layers starting with one region of the input dimension of the neural
network and using a recursion based on an analysis of how many regions per
output dimensionality a subsequent layer with a certain width can induce on an
input region with a given dimensionality. The final bound on the number of
regions depends on the number and widths of the layers of the neural network
and on some additional parameters that were used for the recursion. It is
stated in terms of the $L1$-norm of the last column of a product of matrices
and provides a unifying treatment of several previously known bounds: Depending
on the choice of the recursion parameters that determine these matrices, it is
possible to obtain the bounds from Mont\'{u}far (2014), (2017) and Serra et.
al. (2017) as special cases. For the latter, which is the strongest of these
bounds, the formulation in terms of matrices provides new insight. In
particular, by using explicit formulas for a Jordan-like decomposition of the
involved matrices, we achieve new tighter results for the asymptotic setting,
where the number of layers of the same fixed width tends to infinity.
|
Explainable Neural Networks based on Additive Index Models
|
Machine Learning algorithms are increasingly being used in recent years due
to their flexibility in model fitting and increased predictive performance.
However, the complexity of the models makes them hard for the data analyst to
interpret the results and explain them without additional tools. This has led
to much research in developing various approaches to understand the model
behavior. In this paper, we present the Explainable Neural Network (xNN), a
structured neural network designed especially to learn interpretable features.
Unlike fully connected neural networks, the features engineered by the xNN can
be extracted from the network in a relatively straightforward manner and the
results displayed. With appropriate regularization, the xNN provides a
parsimonious explanation of the relationship between the features and the
output. We illustrate this interpretable feature--engineering property on
simulated examples.
|
Learning to Understand Goal Specifications by Modelling Reward
|
Recent work has shown that deep reinforcement-learning agents can learn to
follow language-like instructions from infrequent environment rewards. However,
this places on environment designers the onus of designing language-conditional
reward functions which may not be easily or tractably implemented as the
complexity of the environment and the language scales. To overcome this
limitation, we present a framework within which instruction-conditional RL
agents are trained using rewards obtained not from the environment, but from
reward models which are jointly trained from expert examples. As reward models
improve, they learn to accurately reward agents for completing tasks for
environment configurations---and for instructions---not present amongst the
expert data. This framework effectively separates the representation of what
instructions require from how they can be executed. In a simple grid world, it
enables an agent to learn a range of commands requiring interaction with blocks
and understanding of spatial relations and underspecified abstract
arrangements. We further show the method allows our agent to adapt to changes
in the environment without requiring new expert examples.
|
A linear time method for the detection of point and collective anomalies
|
The challenge of efficiently identifying anomalies in data sequences is an
important statistical problem that now arises in many applications. Whilst
there has been substantial work aimed at making statistical analyses robust to
outliers, or point anomalies, there has been much less work on detecting
anomalous segments, or collective anomalies, particularly in those settings
where point anomalies might also occur. In this article, we introduce
Collective And Point Anomalies (CAPA), a computationally efficient approach
that is suitable when collective anomalies are characterised by either a change
in mean, variance, or both, and distinguishes them from point anomalies.
Theoretical results establish the consistency of CAPA at detecting collective
anomalies and, as a by-product, the consistency of a popular penalised cost
based change in mean and variance detection method. Empirical results show that
CAPA has close to linear computational cost as well as being more accurate at
detecting and locating collective anomalies than other approaches. We
demonstrate the utility of CAPA through its ability to detect exoplanets from
light curve data from the Kepler telescope.
|
Reverse iterative volume sampling for linear regression
|
We study the following basic machine learning task: Given a fixed set of
$d$-dimensional input points for a linear regression problem, we wish to
predict a hidden response value for each of the points. We can only afford to
attain the responses for a small subset of the points that are then used to
construct linear predictions for all points in the dataset. The performance of
the predictions is evaluated by the total square loss on all responses (the
attained as well as the hidden ones). We show that a good approximate solution
to this least squares problem can be obtained from just dimension $d$ many
responses by using a joint sampling technique called volume sampling. Moreover,
the least squares solution obtained for the volume sampled subproblem is an
unbiased estimator of optimal solution based on all n responses. This
unbiasedness is a desirable property that is not shared by other common subset
selection techniques.
Motivated by these basic properties, we develop a theoretical framework for
studying volume sampling, resulting in a number of new matrix expectation
equalities and statistical guarantees which are of importance not only to least
squares regression but also to numerical linear algebra in general. Our methods
also lead to a regularized variant of volume sampling, and we propose the first
efficient algorithms for volume sampling which make this technique a practical
tool in the machine learning toolbox. Finally, we provide experimental evidence
which confirms our theoretical findings.
|
Graph Convolutional Neural Networks for Web-Scale Recommender Systems
|
Recent advancements in deep neural networks for graph-structured data have
led to state-of-the-art performance on recommender system benchmarks. However,
making these methods practical and scalable to web-scale recommendation tasks
with billions of items and hundreds of millions of users remains a challenge.
Here we describe a large-scale deep recommendation engine that we developed and
deployed at Pinterest. We develop a data-efficient Graph Convolutional Network
(GCN) algorithm PinSage, which combines efficient random walks and graph
convolutions to generate embeddings of nodes (i.e., items) that incorporate
both graph structure as well as node feature information. Compared to prior GCN
approaches, we develop a novel method based on highly efficient random walks to
structure the convolutions and design a novel training strategy that relies on
harder-and-harder training examples to improve robustness and convergence of
the model. We also develop an efficient MapReduce model inference algorithm to
generate embeddings using a trained model. We deploy PinSage at Pinterest and
train it on 7.5 billion examples on a graph with 3 billion nodes representing
pins and boards, and 18 billion edges. According to offline metrics, user
studies and A/B tests, PinSage generates higher-quality recommendations than
comparable deep learning and graph-based alternatives. To our knowledge, this
is the largest application of deep graph embeddings to date and paves the way
for a new generation of web-scale recommender systems based on graph
convolutional architectures.
|
Learning to rank for censored survival data
|
Survival analysis is a type of semi-supervised ranking task where the target
output (the survival time) is often right-censored. Utilizing this information
is a challenge because it is not obvious how to correctly incorporate these
censored examples into a model. We study how three categories of loss
functions, namely partial likelihood methods, rank methods, and our
classification method based on a Wasserstein metric (WM) and the non-parametric
Kaplan Meier estimate of the probability density to impute the labels of
censored examples, can take advantage of this information. The proposed method
allows us to have a model that predict the probability distribution of an
event. If a clinician had access to the detailed probability of an event over
time this would help in treatment planning. For example, determining if the
risk of kidney graft rejection is constant or peaked after some time. Also, we
demonstrate that this approach directly optimizes the expected C-index which is
the most common evaluation metric for ranking survival models.
|
Beyond Trees: Classification with Sparse Pairwise Dependencies
|
Several classification methods assume that the underlying distributions
follow tree-structured graphical models. Indeed, trees capture statistical
dependencies between pairs of variables, which may be crucial to attain low
classification errors. The resulting classifier is linear in the
log-transformed univariate and bivariate densities that correspond to the tree
edges. In practice, however, observed data may not be well approximated by
trees. Yet, motivated by the importance of pairwise dependencies for accurate
classification, here we propose to approximate the optimal decision boundary by
a sparse linear combination of the univariate and bivariate log-transformed
densities. Our proposed approach is semi-parametric in nature: we
non-parametrically estimate the univariate and bivariate densities, remove
pairs of variables that are nearly independent using the Hilbert-Schmidt
independence criteria, and finally construct a linear SVM on the retained
log-transformed densities. We demonstrate using both synthetic and real data
that our resulting classifier, denoted SLB (Sparse Log-Bivariate density), is
competitive with popular classification methods.
|
Deep Algorithms: designs for networks
|
A new design methodology for neural networks that is guided by traditional
algorithm design is presented. To prove our point, we present two heuristics
and demonstrate an algorithmic technique for incorporating additional weights
in their signal-flow graphs. We show that with training the performance of
these networks can not only exceed the performance of the initial network, but
can match the performance of more-traditional neural network architectures. A
key feature of our approach is that these networks are initialized with
parameters that provide a known performance threshold for the architecture on a
given task.
|
A Peek Into the Hidden Layers of a Convolutional Neural Network Through
a Factorization Lens
|
Despite their increasing popularity and success in a variety of supervised
learning problems, deep neural networks are extremely hard to interpret and
debug: Given and already trained Deep Neural Net, and a set of test inputs, how
can we gain insight into how those inputs interact with different layers of the
neural network? Furthermore, can we characterize a given deep neural network
based on it's observed behavior on different inputs? In this paper we propose a
novel factorization based approach on understanding how different deep neural
networks operate. In our preliminary results, we identify fascinating patterns
that link the factorization rank (typically used as a measure of
interestingness in unsupervised data analysis) with how well or poorly the deep
network has been trained. Finally, our proposed approach can help provide
visual insights on how high-level. interpretable patterns of the network's
input behave inside the hidden layers of the deep network.
|
Discrete-Continuous Mixtures in Probabilistic Programming: Generalized
Semantics and Inference Algorithms
|
Despite the recent successes of probabilistic programming languages (PPLs) in
AI applications, PPLs offer only limited support for random variables whose
distributions combine discrete and continuous elements. We develop the notion
of measure-theoretic Bayesian networks (MTBNs) and use it to provide more
general semantics for PPLs with arbitrarily many random variables defined over
arbitrary measure spaces. We develop two new general sampling algorithms that
are provably correct under the MTBN framework: the lexicographic likelihood
weighting (LLW) for general MTBNs and the lexicographic particle filter (LPF),
a specialized algorithm for state-space models. We further integrate MTBNs into
a widely used PPL system, BLOG, and verify the effectiveness of the new
inference algorithms through representative examples.
|
Killing four birds with one Gaussian process: the relation between
different test-time attacks
|
In machine learning (ML) security, attacks like evasion, model stealing or
membership inference are generally studied in individually. Previous work has
also shown a relationship between some attacks and decision function curvature
of the targeted model. Consequently, we study an ML model allowing direct
control over the decision surface curvature: Gaussian Process classifiers
(GPCs). For evasion, we find that changing GPC's curvature to be robust against
one attack algorithm boils down to enabling a different norm or attack
algorithm to succeed. This is backed up by our formal analysis showing that
static security guarantees are opposed to learning. Concerning intellectual
property, we show formally that lazy learning does not necessarily leak all
information when applied. In practice, often a seemingly secure curvature can
be found. For example, we are able to secure GPC against empirical membership
inference by proper configuration. In this configuration, however, the GPC's
hyper-parameters are leaked, e.g. model reverse engineering succeeds. We
conclude that attacks on classification should not be studied in isolation, but
in relation to each other.
|
Degrees of Freedom and Model Selection for k-means Clustering
|
This paper investigates the model degrees of freedom in k-means clustering.
An extension of Stein's lemma provides an expression for the effective degrees
of freedom in the k-means model. Approximating the degrees of freedom in
practice requires simplifications of this expression, however empirical studies
evince the appropriateness of our proposed approach. The practical relevance of
this new degrees of freedom formulation for k-means is demonstrated through
model selection using the Bayesian Information Criterion. The reliability of
this method is validated through experiments on simulated data as well as on a
large collection of publicly available benchmark data sets from diverse
application areas. Comparisons with popular existing techniques indicate that
this approach is extremely competitive for selecting high quality clustering
solutions. Code to implement the proposed approach is available in the form of
an R package from https://github.com/DavidHofmeyr/edfkmeans.
|
Implicit regularization and solution uniqueness in over-parameterized
matrix sensing
|
We consider whether algorithmic choices in over-parameterized linear matrix
factorization introduce implicit regularization. We focus on noiseless matrix
sensing over rank-$r$ positive semi-definite (PSD) matrices in $\mathbb{R}^{n
\times n}$, with a sensing mechanism that satisfies restricted isometry
properties (RIP). The algorithm we study is \emph{factored gradient descent},
where we model the low-rankness and PSD constraints with the factorization
$UU^\top$, for $U \in \mathbb{R}^{n \times r}$. Surprisingly, recent work
argues that the choice of $r \leq n$ is not pivotal: even setting $U \in
\mathbb{R}^{n \times n}$ is sufficient for factored gradient descent to find
the rank-$r$ solution, which suggests that operating over the factors leads to
an implicit regularization. In this contribution, we provide a different
perspective to the problem of implicit regularization. We show that under
certain conditions, the PSD constraint by itself is sufficient to lead to a
unique rank-$r$ matrix recovery, without implicit or explicit low-rank
regularization. \emph{I.e.}, under assumptions, the set of PSD matrices, that
are consistent with the observed data, is a singleton, regardless of the
algorithm used.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.