title
stringlengths 5
246
| categories
stringlengths 5
94
⌀ | abstract
stringlengths 54
5.03k
| authors
stringlengths 0
6.72k
| doi
stringlengths 12
54
⌀ | id
stringlengths 6
10
⌀ | year
float64 2.02k
2.02k
⌀ | venue
stringclasses 13
values |
|---|---|---|---|---|---|---|---|
Full-Capacity Unitary Recurrent Neural Networks | stat.ML cs.LG cs.NE | Recurrent neural networks are powerful models for processing sequential data,
but they are generally plagued by vanishing and exploding gradient problems.
Unitary recurrent neural networks (uRNNs), which use unitary recurrence
matrices, have recently been proposed as a means to avoid these issues.
However, in previous experiments, the recurrence matrices were restricted to be
a product of parameterized unitary matrices, and an open question remains: when
does such a parameterization fail to represent all unitary matrices, and how
does this restricted representational capacity limit what can be learned? To
address this question, we propose full-capacity uRNNs that optimize their
recurrence matrix over all unitary matrices, leading to significantly improved
performance over uRNNs that use a restricted-capacity recurrence matrix. Our
contribution consists of two main components. First, we provide a theoretical
argument to determine if a unitary parameterization has restricted capacity.
Using this argument, we show that a recently proposed unitary parameterization
has restricted capacity for hidden state dimension greater than 7. Second, we
show how a complete, full-capacity unitary recurrence matrix can be optimized
over the differentiable manifold of unitary matrices. The resulting
multiplicative gradient step is very simple and does not require gradient
clipping or learning rate adaptation. We confirm the utility of our claims by
empirically evaluating our new full-capacity uRNNs on both synthetic and
natural data, achieving superior performance compared to both LSTMs and the
original restricted-capacity uRNNs.
| Scott Wisdom, Thomas Powers, John R. Hershey, Jonathan Le Roux, and
Les Atlas | null | 1611.00035 | null | null |
Exploiting Spatio-Temporal Structure with Recurrent Winner-Take-All
Networks | cs.LG cs.CV | We propose a convolutional recurrent neural network, with Winner-Take-All
dropout for high dimensional unsupervised feature learning in multi-dimensional
time series. We apply the proposedmethod for object recognition with temporal
context in videos and obtain better results than comparable methods in the
literature, including the Deep Predictive Coding Networks previously proposed
by Chalasani and Principe.Our contributions can be summarized as a scalable
reinterpretation of the Deep Predictive Coding Networks trained end-to-end with
backpropagation through time, an extension of the previously proposed
Winner-Take-All Autoencoders to sequences in time, and a new technique for
initializing and regularizing convolutional-recurrent neural networks.
| Eder Santana, Matthew Emigh, Pablo Zegers, Jose C Principe | null | 1611.0005 | null | null |
Kernel Bandwidth Selection for SVDD: Peak Criterion Approach for Large
Data | cs.LG stat.ML | Support Vector Data Description (SVDD) provides a useful approach to
construct a description of multivariate data for single-class classification
and outlier detection with various practical applications. Gaussian kernel used
in SVDD formulation allows flexible data description defined by observations
designated as support vectors. The data boundary of such description is
non-spherical and conforms to the geometric features of the data. By varying
the Gaussian kernel bandwidth parameter, the SVDD-generated boundary can be
made either smoother (more spherical) or tighter/jagged. The former case may
lead to under-fitting, whereas the latter may result in overfitting. Peak
criterion has been proposed to select an optimal value of the kernel bandwidth
to strike the balance between the data boundary smoothness and its ability to
capture the general geometric shape of the data. Peak criterion involves
training SVDD at various values of the kernel bandwidth parameter. When
training datasets are large, the time required to obtain the optimal value of
the Gaussian kernel bandwidth parameter according to Peak method can become
prohibitively large. This paper proposes an extension of Peak method for the
case of large data. The proposed method gives good results when applied to
several datasets. Two existing alternative methods of computing the Gaussian
kernel bandwidth parameter (Coefficient of Variation and Distance to the
Farthest Neighbor) were modified to allow comparison with the proposed method
on convergence. Empirical comparison demonstrates the advantage of the proposed
method.
| Sergiy Peredriy, Deovrat Kakde, Arin Chaudhuri | 10.1109/BigData.2017.8258344 | 1611.00058 | null | null |
Bayesian Adaptive Data Analysis Guarantees from Subgaussianity | cs.LG math.PR stat.ML | The new field of adaptive data analysis seeks to provide algorithms and
provable guarantees for models of machine learning that allow researchers to
reuse their data, which normally falls outside of the usual statistical
paradigm of static data analysis. In 2014, Dwork, Feldman, Hardt, Pitassi,
Reingold and Roth introduced one potential model and proposed several solutions
based on differential privacy. In previous work in 2016, we described a problem
with this model and instead proposed a Bayesian variant, but also found that
the analogous Bayesian methods cannot achieve the same statistical guarantees
as in the static case.
In this paper, we prove the first positive results for the Bayesian model,
showing that with a Dirichlet prior, the posterior mean algorithm indeed
matches the statistical guarantees of the static case. The main ingredient is a
new theorem showing that the $\mathrm{Beta}(\alpha,\beta)$ distribution is
subgaussian with variance proxy $O(1/(\alpha+\beta+1))$, a concentration result
also of independent interest. We provide two proofs of this result: a
probabilistic proof utilizing a simple condition for the raw moments of a
positive random variable and a learning-theoretic proof based on considering
the beta distribution as a posterior, both of which have implications to other
related problems.
| Sam Elder | null | 1611.00065 | null | null |
Embedding Deep Metric for Person Re-identication A Study Against Large
Variations | cs.CV cs.LG | Person re-identification is challenging due to the large variations of pose,
illumination, occlusion and camera view. Owing to these variations, the
pedestrian data is distributed as highly-curved manifolds in the feature space,
despite the current convolutional neural networks (CNN)'s capability of feature
extraction. However, the distribution is unknown, so it is difficult to use the
geodesic distance when comparing two samples. In practice, the current deep
embedding methods use the Euclidean distance for the training and test. On the
other hand, the manifold learning methods suggest to use the Euclidean distance
in the local range, combining with the graphical relationship between samples,
for approximating the geodesic distance. From this point of view, selecting
suitable positive i.e. intra-class) training samples within a local range is
critical for training the CNN embedding, especially when the data has large
intra-class variations. In this paper, we propose a novel moderate positive
sample mining method to train robust CNN for person re-identification, dealing
with the problem of large variation. In addition, we improve the learning by a
metric weight constraint, so that the learned metric has a better
generalization ability. Experiments show that these two strategies are
effective in learning robust deep metrics for person re-identification, and
accordingly our deep model significantly outperforms the state-of-the-art
methods on several benchmarks of person re-identification. Therefore, the study
presented in this paper may be useful in inspiring new designs of deep models
for person re-identification.
| Hailin Shi, Yang Yang, Xiangyu Zhu, Shengcai Liao, Zhen Lei, Weishi
Zheng, Stan Z. Li | null | 1611.00137 | null | null |
MusicMood: Predicting the mood of music from song lyrics using machine
learning | cs.LG cs.CL cs.IR | Sentiment prediction of contemporary music can have a wide-range of
applications in modern society, for instance, selecting music for public
institutions such as hospitals or restaurants to potentially improve the
emotional well-being of personnel, patients, and customers, respectively. In
this project, music recommendation system built upon on a naive Bayes
classifier, trained to predict the sentiment of songs based on song lyrics
alone. The experimental results show that music corresponding to a happy mood
can be detected with high precision based on text features obtained from song
lyrics.
| Sebastian Raschka | null | 1611.00138 | null | null |
Product-based Neural Networks for User Response Prediction | cs.LG cs.IR | Predicting user responses, such as clicks and conversions, is of great
importance and has found its usage in many Web applications including
recommender systems, web search and online advertising. The data in those
applications is mostly categorical and contains multiple fields; a typical
representation is to transform it into a high-dimensional sparse binary feature
representation via one-hot encoding. Facing with the extreme sparsity,
traditional models may limit their capacity of mining shallow patterns from the
data, i.e. low-order feature combinations. Deep models like deep neural
networks, on the other hand, cannot be directly applied for the
high-dimensional input because of the huge feature space. In this paper, we
propose a Product-based Neural Networks (PNN) with an embedding layer to learn
a distributed representation of the categorical data, a product layer to
capture interactive patterns between inter-field categories, and further fully
connected layers to explore high-order feature interactions. Our experimental
results on two large-scale real-world ad click datasets demonstrate that PNNs
consistently outperform the state-of-the-art models on various metrics.
| Yanru Qu, Han Cai, Kan Ren, Weinan Zhang, Yong Yu, Ying Wen, Jun Wang | null | 1611.00144 | null | null |
Robust Spectral Inference for Joint Stochastic Matrix Factorization | cs.LG cs.AI | Spectral inference provides fast algorithms and provable optimality for
latent topic analysis. But for real data these algorithms require additional
ad-hoc heuristics, and even then often produce unusable results. We explain
this poor performance by casting the problem of topic inference in the
framework of Joint Stochastic Matrix Factorization (JSMF) and showing that
previous methods violate the theoretical conditions necessary for a good
solution to exist. We then propose a novel rectification method that learns
high quality topics and their interactions even on small, noisy data. This
method achieves results comparable to probabilistic techniques in several
domains while maintaining scalability and provable optimality.
| Moontae Lee, David Bindel, David Mimno | null | 1611.00175 | null | null |
Application Specific Instrumentation (ASIN): A Bio-inspired Paradigm to
Instrumentation using recognition before detection | cs.OH cs.LG | In this paper we present a new scheme for instrumentation, which has been
inspired by the way small mammals sense their environment. We call this scheme
Application Specific Instrumentation (ASIN). A conventional instrumentation
system focuses on gathering as much information about the scene as possible.
This, usually, is a generic system whose data can be used by another system to
take a specific action. ASIN fuses these two steps into one. The major merit of
the proposed scheme is that it uses low resolution sensors and much less
computational overhead to give good performance for a highly specialised
application
| Amit Kumar Mishra | null | 1611.00228 | null | null |
Improving a Credit Scoring Model by Incorporating Bank Statement Derived
Features | cs.LG | In this paper, we investigate the extent to which features derived from bank
statements provided by loan applicants, and which are not declared on an
application form, can enhance a credit scoring model for a New Zealand lending
company. Exploring the potential of such information to improve credit scoring
models in this manner has not been studied previously. We construct a baseline
model based solely on the existing scoring features obtained from the loan
application form, and a second baseline model based solely on the new bank
statement-derived features. A combined feature model is then created by
augmenting the application form features with the new bank statement derived
features. Our experimental results using ROC analysis show that a combined
feature model performs better than both of the two baseline models, and show
that a number of the bank statement-derived features have value in improving
the credit scoring model. The target data set used for modelling was highly
imbalanced, and Naive Bayes was found to be the best performing model, and
outperformed a number of other classifiers commonly used in credit scoring,
suggesting its potential for future use on highly imbalanced data sets.
| Rory P. Bunker, Wenjun Zhang, M. Asif Naeem | null | 1611.00252 | null | null |
Stationary time-vertex signal processing | cs.LG cs.DS stat.ML | This paper considers regression tasks involving high-dimensional multivariate
processes whose structure is dependent on some {known} graph topology. We put
forth a new definition of time-vertex wide-sense stationarity, or joint
stationarity for short, that goes beyond product graphs. Joint stationarity
helps by reducing the estimation variance and recovery complexity. In
particular, for any jointly stationary process (a) one reliably learns the
covariance structure from as little as a single realization of the process, and
(b) solves MMSE recovery problems, such as interpolation and denoising, in
computational time nearly linear on the number of edges and timesteps.
Experiments with three datasets suggest that joint stationarity can yield
accuracy improvements in the recovery of high-dimensional processes evolving
over a graph, even when the latter is only approximately known, or the process
is not strictly stationary.
| Andreas Loukas and Nathana\"el Perraudin | null | 1611.00255 | null | null |
Recurrent Neural Radio Anomaly Detection | cs.LG | We introduce a powerful recurrent neural network based method for novelty
detection to the application of detecting radio anomalies. This approach holds
promise in significantly increasing the ability of naive anomaly detection to
detect small anomalies in highly complex complexity multi-user radio bands. We
demonstrate the efficacy of this approach on a number of common real over the
air radio communications bands of interest and quantify detection performance
in terms of probability of detection an false alarm rates across a range of
interference to band power ratios and compare to baseline methods.
| Timothy J O'Shea, T. Charles Clancy, Robert W. McGwier | null | 1611.00301 | null | null |
Semi-Supervised Radio Signal Identification | cs.LG cs.IT math.IT stat.ML | Radio emitter recognition in dense multi-user environments is an important
tool for optimizing spectrum utilization, identifying and minimizing
interference, and enforcing spectrum policy. Radio data is readily available
and easy to obtain from an antenna, but labeled and curated data is often
scarce making supervised learning strategies difficult and time consuming in
practice. We demonstrate that semi-supervised learning techniques can be used
to scale learning beyond supervised datasets, allowing for discerning and
recalling new radio signals by using sparse signal representations based on
both unsupervised and supervised methods for nonlinear feature learning and
clustering methods.
| Timothy J. O'Shea, Nathan West, Matthew Vondal, T. Charles Clancy | null | 1611.00303 | null | null |
Enhanced Factored Three-Way Restricted Boltzmann Machines for Speech
Detection | cs.SD cs.LG stat.ML | In this letter, we propose enhanced factored three way restricted Boltzmann
machines (EFTW-RBMs) for speech detection. The proposed model incorporates
conditional feature learning by multiplying the dynamical state of the third
unit, which allows a modulation over the visible-hidden node pairs. Instead of
stacking previous frames of speech as the third unit in a recursive manner, the
correlation related weighting coefficients are assigned to the contextual
neighboring frames. Specifically, a threshold function is designed to capture
the long-term features and blend the globally stored speech structure. A
factored low rank approximation is introduced to reduce the parameters of the
three-dimensional interaction tensor, on which non-negative constraint is
imposed to address the sparsity characteristic. The validations through the
area-under-ROC-curve (AUC) and signal distortion ratio (SDR) show that our
approach outperforms several existing 1D and 2D (i.e., time and time-frequency
domain) speech detection algorithms in various noisy environments.
| Pengfei Sun and Jun Qin | null | 1611.00326 | null | null |
Variational Inference via $\chi$-Upper Bound Minimization | stat.ML cs.LG stat.CO stat.ME | Variational inference (VI) is widely used as an efficient alternative to
Markov chain Monte Carlo. It posits a family of approximating distributions $q$
and finds the closest member to the exact posterior $p$. Closeness is usually
measured via a divergence $D(q || p)$ from $q$ to $p$. While successful, this
approach also has problems. Notably, it typically leads to underestimation of
the posterior variance. In this paper we propose CHIVI, a black-box variational
inference algorithm that minimizes $D_{\chi}(p || q)$, the $\chi$-divergence
from $p$ to $q$. CHIVI minimizes an upper bound of the model evidence, which we
term the $\chi$ upper bound (CUBO). Minimizing the CUBO leads to improved
posterior uncertainty, and it can also be used with the classical VI lower
bound (ELBO) to provide a sandwich estimate of the model evidence. We study
CHIVI on three models: probit regression, Gaussian process classification, and
a Cox process model of basketball plays. When compared to expectation
propagation and classical VI, CHIVI produces better error rates and more
accurate estimates of posterior variance.
| Adji B. Dieng, Dustin Tran, Rajesh Ranganath, John Paisley, David M.
Blei | null | 1611.00328 | null | null |
Stochastic Variational Deep Kernel Learning | stat.ML cs.LG stat.ME | Deep kernel learning combines the non-parametric flexibility of kernel
methods with the inductive biases of deep learning architectures. We propose a
novel deep kernel learning model and stochastic variational inference procedure
which generalizes deep kernel learning approaches to enable classification,
multi-task learning, additive covariance structures, and stochastic gradient
training. Specifically, we apply additive base kernels to subsets of output
features from deep neural architectures, and jointly learn the parameters of
the base kernels and deep network through a Gaussian process marginal
likelihood objective. Within this framework, we derive an efficient form of
stochastic variational inference which leverages local kernel interpolation,
inducing points, and structure exploiting algebra. We show improved performance
over stand alone deep networks, SVMs, and state of the art scalable Gaussian
processes on several classification benchmarks, including an airline delay
dataset containing 6 million training points, CIFAR, and ImageNet.
| Andrew Gordon Wilson, Zhiting Hu, Ruslan Salakhutdinov, Eric P. Xing | null | 1611.00336 | null | null |
Surpassing Gradient Descent Provably: A Cyclic Incremental Method with
Linear Convergence Rate | math.OC cs.LG | Recently, there has been growing interest in developing optimization methods
for solving large-scale machine learning problems. Most of these problems boil
down to the problem of minimizing an average of a finite set of smooth and
strongly convex functions where the number of functions $n$ is large. Gradient
descent method (GD) is successful in minimizing convex problems at a fast
linear rate; however, it is not applicable to the considered large-scale
optimization setting because of the high computational complexity. Incremental
methods resolve this drawback of gradient methods by replacing the required
gradient for the descent direction with an incremental gradient approximation.
They operate by evaluating one gradient per iteration and executing the average
of the $n$ available gradients as a gradient approximate. Although, incremental
methods reduce the computational cost of GD, their convergence rates do not
justify their advantage relative to GD in terms of the total number of gradient
evaluations until convergence. In this paper, we introduce a Double Incremental
Aggregated Gradient method (DIAG) that computes the gradient of only one
function at each iteration, which is chosen based on a cyclic scheme, and uses
the aggregated average gradient of all the functions to approximate the full
gradient. The iterates of the proposed DIAG method uses averages of both
iterates and gradients in oppose to classic incremental methods that utilize
gradient averages but do not utilize iterate averages. We prove that not only
the proposed DIAG method converges linearly to the optimal solution, but also
its linear convergence factor justifies the advantage of incremental methods on
GD. In particular, we prove that the worst case performance of DIAG is better
than the worst case performance of GD.
| Aryan Mokhtari and Mert G\"urb\"uzbalaban and Alejandro Ribeiro | null | 1611.00347 | null | null |
Adversarial Influence Maximization | cs.SI cs.LG stat.ML | We consider the problem of influence maximization in fixed networks for
contagion models in an adversarial setting. The goal is to select an optimal
set of nodes to seed the influence process, such that the number of influenced
nodes at the conclusion of the campaign is as large as possible. We formulate
the problem as a repeated game between a player and adversary, where the
adversary specifies the edges along which the contagion may spread, and the
player chooses sets of nodes to influence in an online fashion. We establish
upper and lower bounds on the minimax pseudo-regret in both undirected and
directed networks.
| Justin Khim, Varun Jog, Po-Ling Loh | null | 1611.0035 | null | null |
Using Artificial Intelligence to Identify State Secrets | cs.CY cs.CL cs.LG | Whether officials can be trusted to protect national security information has
become a matter of great public controversy, reigniting a long-standing debate
about the scope and nature of official secrecy. The declassification of
millions of electronic records has made it possible to analyze these issues
with greater rigor and precision. Using machine-learning methods, we examined
nearly a million State Department cables from the 1970s to identify features of
records that are more likely to be classified, such as international
negotiations, military operations, and high-level communications. Even with
incomplete data, algorithms can use such features to identify 90% of classified
cables with <11% false positives. But our results also show that there are
longstanding problems in the identification of sensitive information. Error
analysis reveals many examples of both overclassification and
underclassification. This indicates both the need for research on inter-coder
reliability among officials as to what constitutes classified material and the
opportunity to develop recommender systems to better manage both classification
and declassification.
| Renato Rocha Souza, Flavio Codeco Coelho, Rohan Shah, Matthew Connelly | null | 1611.00356 | null | null |
The Machine Learning Algorithm as Creative Musical Tool | cs.HC cs.LG | Machine learning is the capacity of a computational system to learn
structures from datasets in order to make predictions on newly seen data. Such
an approach offers a significant advantage in music scenarios in which
musicians can teach the system to learn an idiosyncratic style, or can break
the rules to explore the system's capacity in unexpected ways. In this chapter
we draw on music, machine learning, and human-computer interaction to elucidate
an understanding of machine learning algorithms as creative tools for music and
the sonic arts. We motivate a new understanding of learning algorithms as
human-computer interfaces. We show that, like other interfaces, learning
algorithms can be characterised by the ways their affordances intersect with
goals of human users. We also argue that the nature of interaction between
users and algorithms impacts the usability and usefulness of those algorithms
in profound ways. This human-centred view of machine learning motivates our
concluding discussion of what it means to employ machine learning as a creative
tool.
| Rebecca Fiebrink, Baptiste Caramiaux | null | 1611.00379 | null | null |
CB2CF: A Neural Multiview Content-to-Collaborative Filtering Model for
Completely Cold Item Recommendations | cs.IR cs.CL cs.LG | In Recommender Systems research, algorithms are often characterized as either
Collaborative Filtering (CF) or Content Based (CB). CF algorithms are trained
using a dataset of user preferences while CB algorithms are typically based on
item profiles. These approaches harness different data sources and therefore
the resulting recommended items are generally very different. This paper
presents the CB2CF, a deep neural multiview model that serves as a bridge from
items content into their CF representations. CB2CF is a real-world algorithm
designed for Microsoft Store services that handle around a billion users
worldwide. CB2CF is demonstrated on movies and apps recommendations, where it
is shown to outperform an alternative CB model on completely cold items.
| Oren Barkan, Noam Koenigstein, Eylon Yogev and Ori Katz | null | 1611.00384 | null | null |
Distributed Mean Estimation with Limited Communication | cs.LG | Motivated by the need for distributed learning and optimization algorithms
with low communication cost, we study communication efficient algorithms for
distributed mean estimation. Unlike previous works, we make no probabilistic
assumptions on the data. We first show that for $d$ dimensional data with $n$
clients, a naive stochastic binary rounding approach yields a mean squared
error (MSE) of $\Theta(d/n)$ and uses a constant number of bits per dimension
per client. We then extend this naive algorithm in two ways: we show that
applying a structured random rotation before quantization reduces the error to
$\mathcal{O}((\log d)/n)$ and a better coding strategy further reduces the
error to $\mathcal{O}(1/n)$ and uses a constant number of bits per dimension
per client. We also show that the latter coding strategy is optimal up to a
constant in the minimax sense i.e., it achieves the best MSE for a given
communication cost. We finally demonstrate the practicality of our algorithms
by applying them to distributed Lloyd's algorithm for k-means and power
iteration for PCA.
| Ananda Theertha Suresh, Felix X. Yu, Sanjiv Kumar, H. Brendan McMahan | null | 1611.00429 | null | null |
Natural-Parameter Networks: A Class of Probabilistic Neural Networks | cs.LG cs.AI cs.CL cs.CV stat.ML | Neural networks (NN) have achieved state-of-the-art performance in various
applications. Unfortunately in applications where training data is
insufficient, they are often prone to overfitting. One effective way to
alleviate this problem is to exploit the Bayesian approach by using Bayesian
neural networks (BNN). Another shortcoming of NN is the lack of flexibility to
customize different distributions for the weights and neurons according to the
data, as is often done in probabilistic graphical models. To address these
problems, we propose a class of probabilistic neural networks, dubbed
natural-parameter networks (NPN), as a novel and lightweight Bayesian treatment
of NN. NPN allows the usage of arbitrary exponential-family distributions to
model the weights and neurons. Different from traditional NN and BNN, NPN takes
distributions as input and goes through layers of transformation before
producing distributions to match the target output distributions. As a Bayesian
treatment, efficient backpropagation (BP) is performed to learn the natural
parameters for the distributions over both the weights and neurons. The output
distributions of each layer, as byproducts, may be used as second-order
representations for the associated tasks such as link prediction. Experiments
on real-world datasets show that NPN can achieve state-of-the-art performance.
| Hao Wang, Xingjian Shi, Dit-Yan Yeung | null | 1611.00448 | null | null |
Collaborative Recurrent Autoencoder: Recommend while Learning to Fill in
the Blanks | cs.LG cs.AI cs.CL cs.CV stat.ML | Hybrid methods that utilize both content and rating information are commonly
used in many recommender systems. However, most of them use either handcrafted
features or the bag-of-words representation as a surrogate for the content
information but they are neither effective nor natural enough. To address this
problem, we develop a collaborative recurrent autoencoder (CRAE) which is a
denoising recurrent autoencoder (DRAE) that models the generation of content
sequences in the collaborative filtering (CF) setting. The model generalizes
recent advances in recurrent deep learning from i.i.d. input to non-i.i.d.
(CF-based) input and provides a new denoising scheme along with a novel
learnable pooling scheme for the recurrent autoencoder. To do this, we first
develop a hierarchical Bayesian model for the DRAE and then generalize it to
the CF setting. The synergy between denoising and CF enables CRAE to make
accurate recommendations while learning to fill in the blanks in sequences.
Experiments on real-world datasets from different domains (CiteULike and
Netflix) show that, by jointly modeling the order-aware generation of sequences
for the content information and performing CF for the ratings, CRAE is able to
significantly outperform the state of the art on both the recommendation task
based on ratings and the sequence generation task based on content information.
| Hao Wang, Xingjian Shi, Dit-Yan Yeung | null | 1611.00454 | null | null |
Online Multi-view Clustering with Incomplete Views | cs.LG | In the era of big data, it is common to have data with multiple modalities or
coming from multiple sources, known as "multi-view data". Multi-view clustering
provides a natural way to generate clusters from such data. Since different
views share some consistency and complementary information, previous works on
multi-view clustering mainly focus on how to combine various numbers of views
to improve clustering performance. However, in reality, each view may be
incomplete, i.e., instances missing in the view. Furthermore, the size of data
could be extremely huge. It is unrealistic to apply multi-view clustering in
large real-world applications without considering the incompleteness of views
and the memory requirement. None of previous works have addressed all these
challenges simultaneously. In this paper, we propose an online multi-view
clustering algorithm, OMVC, which deals with large-scale incomplete views. We
model the multi-view clustering problem as a joint weighted nonnegative matrix
factorization problem and process the multi-view data chunk by chunk to reduce
the memory requirement. OMVC learns the latent feature matrices for all the
views and pushes them towards a consensus. We further increase the robustness
of the learned latent feature matrices in OMVC via lasso regularization. To
minimize the influence of incompleteness, dynamic weight setting is introduced
to give lower weights to the incoming missing instances in different views.
More importantly, to reduce the computational time, we incorporate a faster
projected gradient descent by utilizing the Hessian matrices in OMVC. Extensive
experiments conducted on four real data demonstrate the effectiveness of the
proposed OMVC method.
| Weixiang Shao, Lifang He, Chun-Ta Lu, Philip S. Yu | null | 1611.00481 | null | null |
Deep Neural Networks for HDR imaging | cs.CV cs.LG cs.NE | We propose novel methods of solving two tasks using Convolutional Neural
Networks, firstly the task of generating HDR map of a static scene using
differently exposed LDR images of the scene captured using conventional cameras
and secondly the task of finding an optimal tone mapping operator that would
give a better score on the TMQI metric compared to the existing methods. We
quantitatively show the performance of our networks and illustrate the cases
where our networks performs good as well as bad.
| Kshiteej Sheth | null | 1611.00591 | null | null |
TorchCraft: a Library for Machine Learning Research on Real-Time
Strategy Games | cs.LG cs.AI | We present TorchCraft, a library that enables deep learning research on
Real-Time Strategy (RTS) games such as StarCraft: Brood War, by making it
easier to control these games from a machine learning framework, here Torch.
This white paper argues for using RTS games as a benchmark for AI research, and
describes the design and components of TorchCraft.
| Gabriel Synnaeve, Nantas Nardelli, Alex Auvolat, Soumith Chintala,
Timoth\'ee Lacroix, Zeming Lin, Florian Richoux, Nicolas Usunier | null | 1611.00625 | null | null |
Deep counter networks for asynchronous event-based processing | cs.NE cs.LG | Despite their advantages in terms of computational resources, latency, and
power consumption, event-based implementations of neural networks have not been
able to achieve the same performance figures as their equivalent
state-of-the-art deep network models. We propose counter neurons as minimal
spiking neuron models which only require addition and comparison operations,
thus avoiding costly multiplications. We show how inference carried out in deep
counter networks converges to the same accuracy levels as are achieved with
state-of-the-art conventional networks. As their event-based style of
computation leads to reduced latency and sparse updates, counter networks are
ideally suited for efficient compact and low-power hardware implementation. We
present theory and training methods for counter networks, and demonstrate on
the MNIST benchmark that counter networks converge quickly, both in terms of
time and number of operations required, to state-of-the-art classification
accuracy.
| Jonathan Binas, Giacomo Indiveri, Michael Pfeiffer | null | 1611.0071 | null | null |
The Concrete Distribution: A Continuous Relaxation of Discrete Random
Variables | cs.LG stat.ML | The reparameterization trick enables optimizing large scale stochastic
computation graphs via gradient descent. The essence of the trick is to
refactor each stochastic node into a differentiable function of its parameters
and a random variable with fixed distribution. After refactoring, the gradients
of the loss propagated by the chain rule through the graph are low variance
unbiased estimators of the gradients of the expected loss. While many
continuous random variables have such reparameterizations, discrete random
variables lack useful reparameterizations due to the discontinuous nature of
discrete states. In this work we introduce Concrete random
variables---continuous relaxations of discrete random variables. The Concrete
distribution is a new family of distributions with closed form densities and a
simple reparameterization. Whenever a discrete stochastic node of a computation
graph can be refactored into a one-hot bit representation that is treated
continuously, Concrete stochastic nodes can be used with automatic
differentiation to produce low-variance biased gradients of objectives
(including objectives that depend on the log-probability of latent stochastic
nodes) on the corresponding discrete graph. We demonstrate the effectiveness of
Concrete relaxations on density estimation and structured prediction tasks
using neural networks.
| Chris J. Maddison, Andriy Mnih, Yee Whye Teh | null | 1611.00712 | null | null |
Scalable Semi-Supervised Learning over Networks using Nonsmooth Convex
Optimization | cs.LG cs.DC | We propose a scalable method for semi-supervised (transductive) learning from
massive network-structured datasets. Our approach to semi-supervised learning
is based on representing the underlying hypothesis as a graph signal with small
total variation. Requiring a small total variation of the graph signal
representing the underlying hypothesis corresponds to the central smoothness
assumption that forms the basis for semi-supervised learning, i.e., input
points forming clusters have similar output values or labels. We formulate the
learning problem as a nonsmooth convex optimization problem which we solve by
appealing to Nesterovs optimal first-order method for nonsmooth optimization.
We also provide a message passing formulation of the learning method which
allows for a highly scalable implementation in big data frameworks.
| Alexander Jung and Alfred O. Hero III and Alexandru Mara and Sabeur
Aridhi | null | 1611.00714 | null | null |
Why and When Can Deep -- but Not Shallow -- Networks Avoid the Curse of
Dimensionality: a Review | cs.LG | The paper characterizes classes of functions for which deep learning can be
exponentially better than shallow learning. Deep convolutional networks are a
special case of these conditions, though weight sharing is not the main reason
for their exponential advantage.
| Tomaso Poggio, Hrushikesh Mhaskar, Lorenzo Rosasco, Brando Miranda,
Qianli Liao | null | 1611.0074 | null | null |
Quantum Laplacian Eigenmap | quant-ph cs.LG | Laplacian eigenmap algorithm is a typical nonlinear model for dimensionality
reduction in classical machine learning. We propose an efficient quantum
Laplacian eigenmap algorithm to exponentially speed up the original
counterparts. In our work, we demonstrate that the Hermitian chain product
proposed in quantum linear discriminant analysis (arXiv:1510.00113,2015) can be
applied to implement quantum Laplacian eigenmap algorithm. While classical
Laplacian eigenmap algorithm requires polynomial time to solve the eigenvector
problem, our algorithm is able to exponentially speed up nonlinear
dimensionality reduction.
| Yiming Huang, Xiaoyu Li | null | 1611.0076 | null | null |
Temporal Matrix Completion with Locally Linear Latent Factors for
Medical Applications | cs.LG cs.CV stat.ML | Regular medical records are useful for medical practitioners to analyze and
monitor patient health status especially for those with chronic disease, but
such records are usually incomplete due to unpunctuality and absence of
patients. In order to resolve the missing data problem over time, tensor-based
model is suggested for missing data imputation in recent papers because this
approach makes use of low rank tensor assumption for highly correlated data.
However, when the time intervals between records are long, the data correlation
is not high along temporal direction and such assumption is not valid. To
address this problem, we propose to decompose a matrix with missing data into
its latent factors. Then, the locally linear constraint is imposed on these
factors for matrix completion in this paper. By using a publicly available
dataset and two medical datasets collected from hospital, experimental results
show that the proposed algorithm achieves the best performance by comparing
with the existing methods.
| Frodo Kin Sun Chan, Andy J Ma, Pong C Yuen, Terry Cheuk-Fung Yip,
Yee-Kit Tse, Vincent Wai-Sun Wong and Grace Lai-Hung Wong | null | 1611.008 | null | null |
Multidimensional Binary Search for Contextual Decision-Making | cs.DS cs.LG | We consider a multidimensional search problem that is motivated by questions
in contextual decision-making, such as dynamic pricing and personalized
medicine. Nature selects a state from a $d$-dimensional unit ball and then
generates a sequence of $d$-dimensional directions. We are given access to the
directions, but not access to the state. After receiving a direction, we have
to guess the value of the dot product between the state and the direction. Our
goal is to minimize the number of times when our guess is more than $\epsilon$
away from the true answer. We construct a polynomial time algorithm that we
call Projected Volume achieving regret $O(d\log(d/\epsilon))$, which is optimal
up to a $\log d$ factor. The algorithm combines a volume cutting strategy with
a new geometric technique that we call cylindrification.
| Ilan Lobel, Renato Paes Leme, Adrian Vladu | null | 1611.00829 | null | null |
Initialization and Coordinate Optimization for Multi-way Matching | stat.ML cs.CV cs.LG | We consider the problem of consistently matching multiple sets of elements to
each other, which is a common task in fields such as computer vision. To solve
the underlying NP-hard objective, existing methods often relax or approximate
it, but end up with unsatisfying empirical performance due to a misaligned
objective. We propose a coordinate update algorithm that directly optimizes the
target objective. By using pairwise alignment information to build an
undirected graph and initializing the permutation matrices along the edges of
its Maximum Spanning Tree, our algorithm successfully avoids bad local optima.
Theoretically, with high probability our algorithm guarantees an optimal
solution under reasonable noise assumptions. Empirically, our algorithm
consistently and significantly outperforms existing methods on several
benchmark tasks on real datasets.
| Da Tang and Tony Jebara | null | 1611.00838 | null | null |
Deep Convolutional Neural Network Design Patterns | cs.LG cs.CV cs.NE | Recent research in the deep learning field has produced a plethora of new
architectures. At the same time, a growing number of groups are applying deep
learning to new applications. Some of these groups are likely to be composed of
inexperienced deep learning practitioners who are baffled by the dizzying array
of architecture choices and therefore opt to use an older architecture (i.e.,
Alexnet). Here we attempt to bridge this gap by mining the collective knowledge
contained in recent deep learning research to discover underlying principles
for designing neural network architectures. In addition, we describe several
architectural innovations, including Fractal of FractalNet network, Stagewise
Boosting Networks, and Taylor Series Networks (our Caffe code and prototxt
files is available at https://github.com/iPhysicist/CNNDesignPatterns). We hope
others are inspired to build on our preliminary work.
| Leslie N. Smith and Nicholay Topin | null | 1611.00847 | null | null |
Quantile Reinforcement Learning | cs.LG cs.AI | In reinforcement learning, the standard criterion to evaluate policies in a
state is the expectation of (discounted) sum of rewards. However, this
criterion may not always be suitable, we consider an alternative criterion
based on the notion of quantiles. In the case of episodic reinforcement
learning problems, we propose an algorithm based on stochastic approximation
with two timescales. We evaluate our proposition on a simple model of the TV
show, Who wants to be a millionaire.
| Hugo Gilbert and Paul Weng | null | 1611.00862 | null | null |
Extracting Actionability from Machine Learning Models by Sub-optimal
Deterministic Planning | cs.AI cs.LG | A main focus of machine learning research has been improving the
generalization accuracy and efficiency of prediction models. Many models such
as SVM, random forest, and deep neural nets have been proposed and achieved
great success. However, what emerges as missing in many applications is
actionability, i.e., the ability to turn prediction results into actions. For
example, in applications such as customer relationship management, clinical
prediction, and advertisement, the users need not only accurate prediction, but
also actionable instructions which can transfer an input to a desirable goal
(e.g., higher profit repays, lower morbidity rates, higher ads hit rates).
Existing effort in deriving such actionable knowledge is few and limited to
simple action models which restricted to only change one attribute for each
action. The dilemma is that in many real applications those action models are
often more complex and harder to extract an optimal solution.
In this paper, we propose a novel approach that achieves actionability by
combining learning with planning, two core areas of AI. In particular, we
propose a framework to extract actionable knowledge from random forest, one of
the most widely used and best off-the-shelf classifiers. We formulate the
actionability problem to a sub-optimal action planning (SOAP) problem, which is
to find a plan to alter certain features of a given input so that the random
forest would yield a desirable output, while minimizing the total costs of
actions. Technically, the SOAP problem is formulated in the SAS+ planning
formalism, and solved using a Max-SAT based approach. Our experimental results
demonstrate the effectiveness and efficiency of the proposed approach on a
personal credit dataset and other benchmarks. Our work represents a new
application of automated planning on an emerging and challenging machine
learning paradigm.
| Qiang Lyu, Yixin Chen, Zhaorong Li, Zhicheng Cui, Ling Chen, Xing
Zhang, Haihua Shen | null | 1611.00873 | null | null |
Low Rank Approximation with Entrywise $\ell_1$-Norm Error | cs.DS cs.CC cs.LG | We study the $\ell_1$-low rank approximation problem, where for a given $n
\times d$ matrix $A$ and approximation factor $\alpha \geq 1$, the goal is to
output a rank-$k$ matrix $\widehat{A}$ for which
$$\|A-\widehat{A}\|_1 \leq \alpha \cdot \min_{\textrm{rank-}k\textrm{
matrices}~A'}\|A-A'\|_1,$$ where for an $n \times d$ matrix $C$, we let
$\|C\|_1 = \sum_{i=1}^n \sum_{j=1}^d |C_{i,j}|$. This error measure is known to
be more robust than the Frobenius norm in the presence of outliers and is
indicated in models where Gaussian assumptions on the noise may not apply. The
problem was shown to be NP-hard by Gillis and Vavasis and a number of
heuristics have been proposed. It was asked in multiple places if there are any
approximation algorithms.
We give the first provable approximation algorithms for $\ell_1$-low rank
approximation, showing that it is possible to achieve approximation factor
$\alpha = (\log d) \cdot \mathrm{poly}(k)$ in $\mathrm{nnz}(A) + (n+d)
\mathrm{poly}(k)$ time, where $\mathrm{nnz}(A)$ denotes the number of non-zero
entries of $A$. If $k$ is constant, we further improve the approximation ratio
to $O(1)$ with a $\mathrm{poly}(nd)$-time algorithm. Under the Exponential Time
Hypothesis, we show there is no $\mathrm{poly}(nd)$-time algorithm achieving a
$(1+\frac{1}{\log^{1+\gamma}(nd)})$-approximation, for $\gamma > 0$ an
arbitrarily small constant, even when $k = 1$.
We give a number of additional results for $\ell_1$-low rank approximation:
nearly tight upper and lower bounds for column subset selection, CUR
decompositions, extensions to low rank approximation with respect to
$\ell_p$-norms for $1 \leq p < 2$ and earthmover distance, low-communication
distributed protocols and low-memory streaming algorithms, algorithms with
limited randomness, and bicriteria algorithms. We also give a preliminary
empirical evaluation.
| Zhao Song, David P. Woodruff, Peilin Zhong | null | 1611.00898 | null | null |
Fast Eigenspace Approximation using Random Signals | cs.DS cs.LG stat.ML | We focus in this work on the estimation of the first $k$ eigenvectors of any
graph Laplacian using filtering of Gaussian random signals. We prove that we
only need $k$ such signals to be able to exactly recover as many of the
smallest eigenvectors, regardless of the number of nodes in the graph. In
addition, we address key issues in implementing the theoretical concepts in
practice using accurate approximated methods. We also propose fast algorithms
both for eigenspace approximation and for the determination of the $k$th
smallest eigenvalue $\lambda_k$. The latter proves to be extremely efficient
under the assumption of locally uniform distribution of the eigenvalue over the
spectrum. Finally, we present experiments which show the validity of our method
in practice and compare it to state-of-the-art methods for clustering and
visualization both on synthetic small-scale datasets and larger real-world
problems of millions of nodes. We show that our method allows a better scaling
with the number of nodes than all previous methods while achieving an almost
perfect reconstruction of the eigenspace formed by the first $k$ eigenvectors.
| Johan Paratte and Lionel Martin | null | 1611.00938 | null | null |
Multitask Protein Function Prediction Through Task Dissimilarity | stat.ML cs.LG q-bio.QM | Automated protein function prediction is a challenging problem with
distinctive features, such as the hierarchical organization of protein
functions and the scarcity of annotated proteins for most biological functions.
We propose a multitask learning algorithm addressing both issues. Unlike
standard multitask algorithms, which use task (protein functions) similarity
information as a bias to speed up learning, we show that dissimilarity
information enforces separation of rare class labels from frequent class
labels, and for this reason is better suited for solving unbalanced protein
function prediction problems. We support our claim by showing that a multitask
extension of the label propagation algorithm empirically works best when the
task relatedness information is represented using a dissimilarity matrix as
opposed to a similarity matrix. Moreover, the experimental comparison carried
out on three model organism shows that our method has a more stable performance
in both "protein-centric" and "function-centric" evaluation settings.
| Marco Frasca and Nicol\`o Cesa Bianchi | 10.1109/TCBB.2017.2684127 | 1611.00962 | null | null |
Learning to Pivot with Adversarial Networks | stat.ML cs.LG cs.NE physics.data-an stat.ME | Several techniques for domain adaptation have been proposed to account for
differences in the distribution of the data used for training and testing. The
majority of this work focuses on a binary domain label. Similar problems occur
in a scientific context where there may be a continuous family of plausible
data generation processes associated to the presence of systematic
uncertainties. Robust inference is possible if it is based on a pivot -- a
quantity whose distribution does not depend on the unknown values of the
nuisance parameters that parametrize this family of data generation processes.
In this work, we introduce and derive theoretical results for a training
procedure based on adversarial networks for enforcing the pivotal property (or,
equivalently, fairness with respect to continuous attributes) on a predictive
model. The method includes a hyperparameter to control the trade-off between
accuracy and robustness. We demonstrate the effectiveness of this approach with
a toy example and examples from particle physics.
| Gilles Louppe, Michael Kagan, Kyle Cranmer | null | 1611.01046 | null | null |
Learning Locomotion Skills Using DeepRL: Does the Choice of Action Space
Matter? | cs.LG cs.GR cs.RO | The use of deep reinforcement learning allows for high-dimensional state
descriptors, but little is known about how the choice of action representation
impacts the learning difficulty and the resulting performance. We compare the
impact of four different action parameterizations (torques, muscle-activations,
target joint angles, and target joint-angle velocities) in terms of learning
time, policy robustness, motion quality, and policy query rates. Our results
are evaluated on a gait-cycle imitation task for multiple planar articulated
figures and multiple gaits. We demonstrate that the local feedback provided by
higher-level action parameterizations can significantly impact the learning,
robustness, and quality of the resulting policies.
| Xue Bin Peng, Michiel van de Panne | 10.1145/3099564.3099567 | 1611.01055 | null | null |
A-Ward_p\b{eta}: Effective hierarchical clustering using the Minkowski
metric and a fast k -means initialisation | cs.LG stat.ML | In this paper we make two novel contributions to hierarchical clustering.
First, we introduce an anomalous pattern initialisation method for hierarchical
clustering algorithms, called A-Ward, capable of substantially reducing the
time they take to converge. This method generates an initial partition with a
sufficiently large number of clusters. This allows the cluster merging process
to start from this partition rather than from a trivial partition composed
solely of singletons. Our second contribution is an extension of the Ward and
Ward p algorithms to the situation where the feature weight exponent can differ
from the exponent of the Minkowski distance. This new method, called A-Ward
p\b{eta} , is able to generate a much wider variety of clustering solutions. We
also demonstrate that its parameters can be estimated reasonably well by using
a cluster validity index. We perform numerous experiments using data sets with
two types of noise, insertion of noise features and blurring within-cluster
values of some features. These experiments allow us to conclude: (i) our
anomalous pattern initialisation method does indeed reduce the time a
hierarchical clustering algorithm takes to complete, without negatively
impacting its cluster recovery ability; (ii) A-Ward p\b{eta} provides better
cluster recovery than both Ward and Ward p.
| Renato Cordeiro de Amorim, Vladimir Makarenkov, Boris Mirkin | 10.1016/j.ins.2016.07.076 | 1611.0106 | null | null |
Cross: Efficient Low-rank Tensor Completion | stat.ME cs.LG math.ST stat.ML stat.TH | The completion of tensors, or high-order arrays, attracts significant
attention in recent research. Current literature on tensor completion primarily
focuses on recovery from a set of uniformly randomly measured entries, and the
required number of measurements to achieve recovery is not guaranteed to be
optimal. In addition, the implementation of some previous methods is NP-hard.
In this article, we propose a framework for low-rank tensor completion via a
novel tensor measurement scheme we name Cross. The proposed procedure is
efficient and easy to implement. In particular, we show that a third order
tensor of Tucker rank-$(r_1, r_2, r_3)$ in $p_1$-by-$p_2$-by-$p_3$ dimensional
space can be recovered from as few as $r_1r_2r_3 + r_1(p_1-r_1) + r_2(p_2-r_2)
+ r_3(p_3-r_3)$ noiseless measurements, which matches the sample complexity
lower-bound. In the case of noisy measurements, we also develop a theoretical
upper bound and the matching minimax lower bound for recovery error over
certain classes of low-rank tensors for the proposed procedure. The results can
be further extended to fourth or higher-order tensors. Simulation studies show
that the method performs well under a variety of settings. Finally, the
procedure is illustrated through a real dataset in neuroimaging.
| Anru Zhang | null | 1611.01129 | null | null |
Using a Deep Reinforcement Learning Agent for Traffic Signal Control | cs.LG cs.SY | Ensuring transportation systems are efficient is a priority for modern
society. Technological advances have made it possible for transportation
systems to collect large volumes of varied data on an unprecedented scale. We
propose a traffic signal control system which takes advantage of this new, high
quality data, with minimal abstraction compared to other proposed systems. We
apply modern deep reinforcement learning methods to build a truly adaptive
traffic signal control agent in the traffic microsimulator SUMO. We propose a
new state space, the discrete traffic state encoding, which is information
dense. The discrete traffic state encoding is used as input to a deep
convolutional neural network, trained using Q-learning with experience replay.
Our agent was compared against a one hidden layer neural network traffic signal
control agent and reduces average cumulative delay by 82%, average queue length
by 66% and average travel time by 20%.
| Wade Genders, Saiedeh Razavi | null | 1611.01142 | null | null |
Categorical Reparameterization with Gumbel-Softmax | stat.ML cs.LG | Categorical variables are a natural choice for representing discrete
structure in the world. However, stochastic neural networks rarely use
categorical latent variables due to the inability to backpropagate through
samples. In this work, we present an efficient gradient estimator that replaces
the non-differentiable sample from a categorical distribution with a
differentiable sample from a novel Gumbel-Softmax distribution. This
distribution has the essential property that it can be smoothly annealed into a
categorical distribution. We show that our Gumbel-Softmax estimator outperforms
state-of-the-art gradient estimators on structured output prediction and
unsupervised generative modeling tasks with categorical latent variables, and
enables large speedups on semi-supervised classification.
| Eric Jang, Shixiang Gu, Ben Poole | null | 1611.01144 | null | null |
PrivLogit: Efficient Privacy-preserving Logistic Regression by Tailoring
Numerical Optimizers | cs.LG cs.CR stat.ML | Safeguarding privacy in machine learning is highly desirable, especially in
collaborative studies across many organizations. Privacy-preserving distributed
machine learning (based on cryptography) is popular to solve the problem.
However, existing cryptographic protocols still incur excess computational
overhead. Here, we make a novel observation that this is partially due to naive
adoption of mainstream numerical optimization (e.g., Newton method) and failing
to tailor for secure computing. This work presents a contrasting perspective:
customizing numerical optimization specifically for secure settings. We propose
a seemingly less-favorable optimization method that can in fact significantly
accelerate privacy-preserving logistic regression. Leveraging this new method,
we propose two new secure protocols for conducting logistic regression in a
privacy-preserving and distributed manner. Extensive theoretical and empirical
evaluations prove the competitive performance of our two secure proposals while
without compromising accuracy or privacy: with speedup up to 2.3x and 8.1x,
respectively, over state-of-the-art; and even faster as data scales up. Such
drastic speedup is on top of and in addition to performance improvements from
existing (and future) state-of-the-art cryptography. Our work provides a new
way towards efficient and practical privacy-preserving logistic regression for
large-scale studies which are common for modern science.
| Wei Xie, Yang Wang, Steven M. Boker, Donald E. Brown | null | 1611.0117 | null | null |
Demystifying ResNet | cs.NE cs.LG stat.ML | The Residual Network (ResNet), proposed in He et al. (2015), utilized
shortcut connections to significantly reduce the difficulty of training, which
resulted in great performance boosts in terms of both training and
generalization error.
It was empirically observed in He et al. (2015) that stacking more layers of
residual blocks with shortcut 2 results in smaller training error, while it is
not true for shortcut of length 1 or 3. We provide a theoretical explanation
for the uniqueness of shortcut 2.
We show that with or without nonlinearities, by adding shortcuts that have
depth two, the condition number of the Hessian of the loss function at the zero
initial point is depth-invariant, which makes training very deep models no more
difficult than shallow ones. Shortcuts of higher depth result in an extremely
flat (high-order) stationary point initially, from which the optimization
algorithm is hard to escape. The shortcut 1, however, is essentially equivalent
to no shortcuts, which has a condition number exploding to infinity as the
number of layers grows. We further argue that as the number of layers tends to
infinity, it suffices to only look at the loss function at the zero initial
point.
Extensive experiments are provided accompanying our theoretical results. We
show that initializing the network to small weights with shortcut 2 achieves
significantly better results than random Gaussian (Xavier) initialization,
orthogonal initialization, and shortcuts of deeper depth, from various
perspectives ranging from final loss, learning dynamics and stability, to the
behavior of the Hessian along the learning process.
| Sihan Li, Jiantao Jiao, Yanjun Han, Tsachy Weissman | null | 1611.01186 | null | null |
Conspiracies between Learning Algorithms, Circuit Lower Bounds and
Pseudorandomness | cs.CC cs.CR cs.DS cs.LG | We prove several results giving new and stronger connections between
learning, circuit lower bounds and pseudorandomness. Among other results, we
show a generic learning speedup lemma, equivalences between various learning
models in the exponential time and subexponential time regimes, a dichotomy
between learning and pseudorandomness, consequences of non-trivial learning for
circuit lower bounds, Karp-Lipton theorems for probabilistic exponential time,
and NC$^1$-hardness for the Minimum Circuit Size Problem.
| Igor C. Oliveira, Rahul Santhanam | null | 1611.0119 | null | null |
Combating Reinforcement Learning's Sisyphean Curse with Intrinsic Fear | cs.LG cs.NE stat.ML | Many practical environments contain catastrophic states that an optimal agent
would visit infrequently or never. Even on toy problems, Deep Reinforcement
Learning (DRL) agents tend to periodically revisit these states upon forgetting
their existence under a new policy. We introduce intrinsic fear (IF), a learned
reward shaping that guards DRL agents against periodic catastrophes. IF agents
possess a fear model trained to predict the probability of imminent
catastrophe. This score is then used to penalize the Q-learning objective. Our
theoretical analysis bounds the reduction in average return due to learning on
the perturbed objective. We also prove robustness to classification errors. As
a bonus, IF models tend to learn faster, owing to reward shaping. Experiments
demonstrate that intrinsic-fear DQNs solve otherwise pathological environments
and improve on several Atari games.
| Zachary C. Lipton, Kamyar Azizzadenesheli, Abhishek Kumar, Lihong Li,
Jianfeng Gao, Li Deng | null | 1611.01211 | null | null |
Sample Efficient Actor-Critic with Experience Replay | cs.LG | This paper presents an actor-critic deep reinforcement learning agent with
experience replay that is stable, sample efficient, and performs remarkably
well on challenging environments, including the discrete 57-game Atari domain
and several continuous control problems. To achieve this, the paper introduces
several innovations, including truncated importance sampling with bias
correction, stochastic dueling network architectures, and a new trust region
policy optimization method.
| Ziyu Wang, Victor Bapst, Nicolas Heess, Volodymyr Mnih, Remi Munos,
Koray Kavukcuoglu, Nando de Freitas | null | 1611.01224 | null | null |
Deep Information Propagation | stat.ML cs.LG | We study the behavior of untrained neural networks whose weights and biases
are randomly distributed using mean field theory. We show the existence of
depth scales that naturally limit the maximum depth of signal propagation
through these random networks. Our main practical result is to show that random
networks may be trained precisely when information can travel through them.
Thus, the depth scales that we identify provide bounds on how deep a network
may be trained for a specific choice of hyperparameters. As a corollary to
this, we argue that in networks at the edge of chaos, one of these depth scales
diverges. Thus arbitrarily deep networks may be trained only sufficiently close
to criticality. We show that the presence of dropout destroys the
order-to-chaos critical point and therefore strongly limits the maximum
trainable depth for random networks. Finally, we develop a mean field theory
for backpropagation and we show that the ordered and chaotic phases correspond
to regions of vanishing and exploding gradient respectively.
| Samuel S. Schoenholz, Justin Gilmer, Surya Ganguli and Jascha
Sohl-Dickstein | null | 1611.01232 | null | null |
Adversarial Machine Learning at Scale | cs.CV cs.CR cs.LG stat.ML | Adversarial examples are malicious inputs designed to fool machine learning
models. They often transfer from one model to another, allowing attackers to
mount black box attacks without knowledge of the target model's parameters.
Adversarial training is the process of explicitly training a model on
adversarial examples, in order to make it more robust to attack or to reduce
its test error on clean inputs. So far, adversarial training has primarily been
applied to small problems. In this research, we apply adversarial training to
ImageNet. Our contributions include: (1) recommendations for how to succesfully
scale adversarial training to large models and datasets, (2) the observation
that adversarial training confers robustness to single-step attack methods, (3)
the finding that multi-step attack methods are somewhat less transferable than
single-step attack methods, so single-step attacks are the best for mounting
black-box attacks, and (4) resolution of a "label leaking" effect that causes
adversarially trained models to perform better on adversarial examples than on
clean examples, because the adversarial example construction process uses the
true label and the model can learn to exploit regularities in the construction
process.
| Alexey Kurakin, Ian Goodfellow, Samy Bengio | null | 1611.01236 | null | null |
Reparameterization trick for discrete variables | stat.ML cs.LG | Low-variance gradient estimation is crucial for learning directed graphical
models parameterized by neural networks, where the reparameterization trick is
widely used for those with continuous variables. While this technique gives
low-variance gradient estimates, it has not been directly applicable to
discrete variables, the sampling of which inherently requires discontinuous
operations. We argue that the discontinuity can be bypassed by marginalizing
out the variable of interest, which results in a new reparameterization trick
for discrete variables. This reparameterization greatly reduces the variance,
which is understood by regarding the method as an application of common random
numbers to the estimation. The resulting estimator is theoretically guaranteed
to have a variance not larger than that of the likelihood-ratio method with the
optimal input-dependent baseline. We give empirical results for variational
learning of sigmoid belief networks.
| Seiya Tokui and Issei sato | null | 1611.01239 | null | null |
Generalized Topic Modeling | cs.LG cs.CL cs.DS cs.IR | Recently there has been significant activity in developing algorithms with
provable guarantees for topic modeling. In standard topic models, a topic (such
as sports, business, or politics) is viewed as a probability distribution $\vec
a_i$ over words, and a document is generated by first selecting a mixture $\vec
w$ over topics, and then generating words i.i.d. from the associated mixture
$A{\vec w}$. Given a large collection of such documents, the goal is to recover
the topic vectors and then to correctly classify new documents according to
their topic mixture.
In this work we consider a broad generalization of this framework in which
words are no longer assumed to be drawn i.i.d. and instead a topic is a complex
distribution over sequences of paragraphs. Since one could not hope to even
represent such a distribution in general (even if paragraphs are given using
some natural feature representation), we aim instead to directly learn a
document classifier. That is, we aim to learn a predictor that given a new
document, accurately predicts its topic mixture, without learning the
distributions explicitly. We present several natural conditions under which one
can do this efficiently and discuss issues such as noise tolerance and sample
complexity in this model. More generally, our model can be viewed as a
generalization of the multi-view or co-training setting in machine learning.
| Avrim Blum, Nika Haghtalab | null | 1611.01259 | null | null |
Learning Identity Mappings with Residual Gates | cs.CV cs.LG | We propose a new layer design by adding a linear gating mechanism to shortcut
connections. By using a scalar parameter to control each gate, we provide a way
to learn identity mappings by optimizing only one parameter. We build upon the
motivation behind Residual Networks, where a layer is reformulated in order to
make learning identity mappings less problematic to the optimizer. The
augmentation introduces only one extra parameter per layer, and provides easier
optimization by making degeneration into identity mappings simpler. We propose
a new model, the Gated Residual Network, which is the result when augmenting
Residual Networks. Experimental results show that augmenting layers provides
better optimization, increased performance, and more layer independence. We
evaluate our method on MNIST using fully-connected networks, showing empirical
indications that our augmentation facilitates the optimization of deep models,
and that it provides high tolerance to full layer removal: the model retains
over 90% of its performance even after half of its layers have been randomly
removed. We also evaluate our model on CIFAR-10 and CIFAR-100 using Wide Gated
ResNets, achieving 3.65% and 18.27% error, respectively.
| Pedro H. P. Savarese and Leonardo O. Mazza and Daniel R. Figueiredo | null | 1611.0126 | null | null |
Semantic Noise Modeling for Better Representation Learning | cs.LG cs.NE | Latent representation learned from multi-layered neural networks via
hierarchical feature abstraction enables recent success of deep learning. Under
the deep learning framework, generalization performance highly depends on the
learned latent representation which is obtained from an appropriate training
scenario with a task-specific objective on a designed network model. In this
work, we propose a novel latent space modeling method to learn better latent
representation. We designed a neural network model based on the assumption that
good base representation can be attained by maximizing the total correlation
between the input, latent, and output variables. From the base model, we
introduce a semantic noise modeling method which enables class-conditional
perturbation on latent space to enhance the representational power of learned
latent feature. During training, latent vector representation can be
stochastically perturbed by a modeled class-conditional additive noise while
maintaining its original semantic feature. It implicitly brings the effect of
semantic augmentation on the latent space. The proposed model can be easily
learned by back-propagation with common gradient-based optimization algorithms.
Experimental results show that the proposed method helps to achieve performance
benefits against various previous approaches. We also provide the empirical
analyses for the proposed class-conditional perturbation process including
t-SNE visualization.
| Hyo-Eun Kim, Sangheum Hwang, Kyunghyun Cho | null | 1611.01268 | null | null |
A Communication-Efficient Parallel Algorithm for Decision Tree | cs.LG | Decision tree (and its extensions such as Gradient Boosting Decision Trees
and Random Forest) is a widely used machine learning algorithm, due to its
practical effectiveness and model interpretability. With the emergence of big
data, there is an increasing need to parallelize the training process of
decision tree. However, most existing attempts along this line suffer from high
communication costs. In this paper, we propose a new algorithm, called
\emph{Parallel Voting Decision Tree (PV-Tree)}, to tackle this challenge. After
partitioning the training data onto a number of (e.g., $M$) machines, this
algorithm performs both local voting and global voting in each iteration. For
local voting, the top-$k$ attributes are selected from each machine according
to its local data. Then, globally top-$2k$ attributes are determined by a
majority voting among these local candidates. Finally, the full-grained
histograms of the globally top-$2k$ attributes are collected from local
machines in order to identify the best (most informative) attribute and its
split point. PV-Tree can achieve a very low communication cost (independent of
the total number of attributes) and thus can scale out very well. Furthermore,
theoretical analysis shows that this algorithm can learn a near optimal
decision tree, since it can find the best attribute with a large probability.
Our experiments on real-world datasets show that PV-Tree significantly
outperforms the existing parallel decision tree algorithms in the trade-off
between accuracy and efficiency.
| Qi Meng, Guolin Ke, Taifeng Wang, Wei Chen, Qiwei Ye, Zhi-Ming Ma and
Tie-Yan Liu | null | 1611.01276 | null | null |
Information Dropout: Learning Optimal Representations Through Noisy
Computation | stat.ML cs.LG stat.CO | The cross-entropy loss commonly used in deep learning is closely related to
the defining properties of optimal representations, but does not enforce some
of the key properties. We show that this can be solved by adding a
regularization term, which is in turn related to injecting multiplicative noise
in the activations of a Deep Neural Network, a special case of which is the
common practice of dropout. We show that our regularized loss function can be
efficiently minimized using Information Dropout, a generalization of dropout
rooted in information theoretic principles that automatically adapts to the
data and can better exploit architectures of limited capacity. When the task is
the reconstruction of the input, we show that our loss function yields a
Variational Autoencoder as a special case, thus providing a link between
representation learning, information theory and variational inference. Finally,
we prove that we can promote the creation of disentangled representations
simply by enforcing a factorized prior, a fact that has been observed
empirically in recent work. Our experiments validate the theoretical intuitions
behind our method, and we find that information dropout achieves a comparable
or better generalization performance than binary dropout, especially on smaller
models, since it can automatically adapt the noise to the structure of the
network, as well as to the test sample.
| Alessandro Achille, Stefano Soatto | null | 1611.01353 | null | null |
Learning to Rank Scientific Documents from the Crowd | cs.IR cs.CL cs.DL cs.LG cs.SI | Finding related published articles is an important task in any science, but
with the explosion of new work in the biomedical domain it has become
especially challenging. Most existing methodologies use text similarity metrics
to identify whether two articles are related or not. However biomedical
knowledge discovery is hypothesis-driven. The most related articles may not be
ones with the highest text similarities. In this study, we first develop an
innovative crowd-sourcing approach to build an expert-annotated
document-ranking corpus. Using this corpus as the gold standard, we then
evaluate the approaches of using text similarity to rank the relatedness of
articles. Finally, we develop and evaluate a new supervised model to
automatically rank related scientific articles. Our results show that authors'
ranking differ significantly from rankings by text-similarity-based models. By
training a learning-to-rank model on a subset of the annotated corpus, we found
the best supervised learning-to-rank model (SVM-Rank) significantly surpassed
state-of-the-art baseline systems.
| Jesse M Lingeman, Hong Yu | null | 1611.014 | null | null |
Information-Theoretic Bounds and Approximations in Neural Population
Coding | cs.IT cs.LG math.IT | While Shannon's mutual information has widespread applications in many
disciplines, for practical applications it is often difficult to calculate its
value accurately for high-dimensional variables because of the curse of
dimensionality. This paper is focused on effective approximation methods for
evaluating mutual information in the context of neural population coding. For
large but finite neural populations, we derive several information-theoretic
asymptotic bounds and approximation formulas that remain valid in
high-dimensional spaces. We prove that optimizing the population density
distribution based on these approximation formulas is a convex optimization
problem which allows efficient numerical solutions. Numerical simulation
results confirmed that our asymptotic formulas were highly accurate for
approximating mutual information for large neural populations. In special
cases, the approximation formulas are exactly equal to the true mutual
information. We also discuss techniques of variable transformation and
dimensionality reduction to facilitate computation of the approximations.
| Wentao Huang and Kechen Zhang | 10.1162/NECO_a_01056 | 1611.01414 | null | null |
Learning Continuous Semantic Representations of Symbolic Expressions | cs.LG cs.AI | Combining abstract, symbolic reasoning with continuous neural reasoning is a
grand challenge of representation learning. As a step in this direction, we
propose a new architecture, called neural equivalence networks, for the problem
of learning continuous semantic representations of algebraic and logical
expressions. These networks are trained to represent semantic equivalence, even
of expressions that are syntactically very different. The challenge is that
semantic representations must be computed in a syntax-directed manner, because
semantics is compositional, but at the same time, small changes in syntax can
lead to very large changes in semantics, which can be difficult for continuous
neural architectures. We perform an exhaustive evaluation on the task of
checking equivalence on a highly diverse class of symbolic algebraic and
boolean expression types, showing that our model significantly outperforms
existing architectures.
| Miltiadis Allamanis, Pankajan Chanthirasegaran, Pushmeet Kohli,
Charles Sutton | null | 1611.01423 | null | null |
Sparsely-Connected Neural Networks: Towards Efficient VLSI
Implementation of Deep Neural Networks | cs.NE cs.LG | Recently deep neural networks have received considerable attention due to
their ability to extract and represent high-level abstractions in data sets.
Deep neural networks such as fully-connected and convolutional neural networks
have shown excellent performance on a wide range of recognition and
classification tasks. However, their hardware implementations currently suffer
from large silicon area and high power consumption due to the their high degree
of complexity. The power/energy consumption of neural networks is dominated by
memory accesses, the majority of which occur in fully-connected networks. In
fact, they contain most of the deep neural network parameters. In this paper,
we propose sparsely-connected networks, by showing that the number of
connections in fully-connected networks can be reduced by up to 90% while
improving the accuracy performance on three popular datasets (MNIST, CIFAR10
and SVHN). We then propose an efficient hardware architecture based on
linear-feedback shift registers to reduce the memory requirements of the
proposed sparsely-connected networks. The proposed architecture can save up to
90% of memory compared to the conventional implementations of fully-connected
neural networks. Moreover, implementation results show up to 84% reduction in
the energy consumption of a single neuron of the proposed sparsely-connected
networks compared to a single neuron of fully-connected neural networks.
| Arash Ardakani, Carlo Condo and Warren J. Gross | null | 1611.01427 | null | null |
Semi-supervised deep learning by metric embedding | cs.LG | Deep networks are successfully used as classification models yielding
state-of-the-art results when trained on a large number of labeled samples.
These models, however, are usually much less suited for semi-supervised
problems because of their tendency to overfit easily when trained on small
amounts of data. In this work we will explore a new training objective that is
targeting a semi-supervised regime with only a small subset of labeled data.
This criterion is based on a deep metric embedding over distance relations
within the set of labeled samples, together with constraints over the
embeddings of the unlabeled set. The final learned representations are
discriminative in euclidean space, and hence can be used with subsequent
nearest-neighbor classification using the labeled samples.
| Elad Hoffer, Nir Ailon | null | 1611.01449 | null | null |
Ways of Conditioning Generative Adversarial Networks | cs.LG cs.AI stat.ML | The GANs are generative models whose random samples realistically reflect
natural images. It also can generate samples with specific attributes by
concatenating a condition vector into the input, yet research on this field is
not well studied. We propose novel methods of conditioning generative
adversarial networks (GANs) that achieve state-of-the-art results on MNIST and
CIFAR-10. We mainly introduce two models: an information retrieving model that
extracts conditional information from the samples, and a spatial bilinear
pooling model that forms bilinear features derived from the spatial cross
product of an image and a condition vector. These methods significantly enhance
log-likelihood of test data under the conditional distributions compared to the
methods of concatenation.
| Hanock Kwak and Byoung-Tak Zhang | null | 1611.01455 | null | null |
Learning heat diffusion graphs | cs.LG cs.SI stat.ML | Effective information analysis generally boils down to properly identifying
the structure or geometry of the data, which is often represented by a graph.
In some applications, this structure may be partly determined by design
constraints or pre-determined sensing arrangements, like in road transportation
networks for example. In general though, the data structure is not readily
available and becomes pretty difficult to define. In particular, the global
smoothness assumptions, that most of the existing works adopt, are often too
general and unable to properly capture localized properties of data. In this
paper, we go beyond this classical data model and rather propose to represent
information as a sparse combination of localized functions that live on a data
structure represented by a graph. Based on this model, we focus on the problem
of inferring the connectivity that best explains the data samples at different
vertices of a graph that is a priori unknown. We concentrate on the case where
the observed data is actually the sum of heat diffusion processes, which is a
quite common model for data on networks or other irregular structures. We cast
a new graph learning problem and solve it with an efficient nonconvex
optimization algorithm. Experiments on both synthetic and real world data
finally illustrate the benefits of the proposed graph learning framework and
confirm that the data structure can be efficiently learned from data
observations only. We believe that our algorithm will help solving key
questions in diverse application domains such as social and biological network
analysis where it is crucial to unveil proper geometry for data understanding
and inference.
| Dorina Thanou, Xiaowen Dong, Daniel Kressner, and Pascal Frossard | null | 1611.01456 | null | null |
Multi-task learning with deep model based reinforcement learning | cs.LG | In recent years, model-free methods that use deep learning have achieved
great success in many different reinforcement learning environments. Most
successful approaches focus on solving a single task, while multi-task
reinforcement learning remains an open problem. In this paper, we present a
model based approach to deep reinforcement learning which we use to solve
different tasks simultaneously. We show that our approach not only does not
degrade but actually benefits from learning multiple tasks. For our model, we
also present a new kind of recurrent neural network inspired by residual
networks that decouples memory from computation allowing to model complex
environments that do not require lots of memory.
| Asier Mujika | null | 1611.01457 | null | null |
Tying Word Vectors and Word Classifiers: A Loss Framework for Language
Modeling | cs.LG cs.CL stat.ML | Recurrent neural networks have been very successful at predicting sequences
of words in tasks such as language modeling. However, all such models are based
on the conventional classification framework, where the model is trained
against one-hot targets, and each word is represented both as an input and as
an output in isolation. This causes inefficiencies in learning both in terms of
utilizing all of the information and in terms of the number of parameters
needed to train. We introduce a novel theoretical framework that facilitates
better learning in language modeling, and show that our framework leads to
tying together the input embedding and the output projection matrices, greatly
reducing the number of trainable variables. Our framework leads to state of the
art performance on the Penn Treebank with a variety of network models.
| Hakan Inan, Khashayar Khosravi, Richard Socher | null | 1611.01462 | null | null |
Understanding Deep Neural Networks with Rectified Linear Units | cs.LG cond-mat.dis-nn cs.AI cs.CC stat.ML | In this paper we investigate the family of functions representable by deep
neural networks (DNN) with rectified linear units (ReLU). We give an algorithm
to train a ReLU DNN with one hidden layer to *global optimality* with runtime
polynomial in the data size albeit exponential in the input dimension. Further,
we improve on the known lower bounds on size (from exponential to super
exponential) for approximating a ReLU deep net function by a shallower ReLU
net. Our gap theorems hold for smoothly parametrized families of "hard"
functions, contrary to countable, discrete families known in the literature. An
example consequence of our gap theorems is the following: for every natural
number $k$ there exists a function representable by a ReLU DNN with $k^2$
hidden layers and total size $k^3$, such that any ReLU DNN with at most $k$
hidden layers will require at least $\frac{1}{2}k^{k+1}-1$ total nodes.
Finally, for the family of $\mathbb{R}^n\to \mathbb{R}$ DNNs with ReLU
activations, we show a new lowerbound on the number of affine pieces, which is
larger than previous constructions in certain regimes of the network
architecture and most distinctively our lowerbound is demonstrated by an
explicit construction of a *smoothly parameterized* family of functions
attaining this scaling. Our construction utilizes the theory of zonotopes from
polyhedral theory.
| Raman Arora, Amitabh Basu, Poorya Mianjy and Anirbit Mukherjee | null | 1611.01491 | null | null |
Protein Secondary Structure Prediction Using Deep Multi-scale
Convolutional Neural Networks and Next-Step Conditioning | cs.LG q-bio.BM | Recently developed deep learning techniques have significantly improved the
accuracy of various speech and image recognition systems. In this paper we
adapt some of these techniques for protein secondary structure prediction. We
first train a series of deep neural networks to predict eight-class secondary
structure labels given a protein's amino acid sequence information and find
that using recent methods for regularization, such as dropout and weight-norm
constraining, leads to measurable gains in accuracy. We then adapt recent
convolutional neural network architectures--Inception, ReSNet, and DenseNet
with Batch Normalization--to the problem of protein structure prediction. These
convolutional architectures make heavy use of multi-scale filter layers that
simultaneously compute features on several scales, and use residual connections
to prevent underfitting. Using a carefully modified version of these
architectures, we achieve state-of-the-art performance of 70.0% per amino acid
accuracy on the public CB513 benchmark dataset. Finally, we explore additions
from sequence-to-sequence learning, altering the model to make its predictions
conditioned on both the protein's amino acid sequence and its past secondary
structure labels. We introduce a new method of ensembling such a conditional
model with our convolutional model, an approach which reaches 70.6% Q8 accuracy
on CB513. We argue that these results can be further refined for larger boosts
in prediction accuracy through more sophisticated attempts to control
overfitting of conditional models. We aim to release the code for these
experiments as part of the TensorFlow repository.
| Akosua Busia, Jasmine Collins, Navdeep Jaitly | null | 1611.01503 | null | null |
Estimating Causal Direction and Confounding of Two Discrete Variables | stat.ML cs.AI cs.LG | We propose a method to classify the causal relationship between two discrete
variables given only the joint distribution of the variables, acknowledging
that the method is subject to an inherent baseline error. We assume that the
causal system is acyclicity, but we do allow for hidden common causes. Our
algorithm presupposes that the probability distributions $P(C)$ of a cause $C$
is independent from the probability distribution $P(E\mid C)$ of the
cause-effect mechanism. While our classifier is trained with a Bayesian
assumption of flat hyperpriors, we do not make this assumption about our test
data. This work connects to recent developments on the identifiability of
causal models over continuous variables under the assumption of "independent
mechanisms". Carefully-commented Python notebooks that reproduce all our
experiments are available online at
http://vision.caltech.edu/~kchalupk/code.html.
| Krzysztof Chalupka, Frederick Eberhardt and Pietro Perona | null | 1611.01504 | null | null |
Eve: A Gradient Based Optimization Method with Locally and Globally
Adaptive Learning Rates | cs.LG | Adaptive gradient methods for stochastic optimization adjust the learning
rate for each parameter locally. However, there is also a global learning rate
which must be tuned in order to get the best performance. In this paper, we
present a new algorithm that adapts the learning rate locally for each
parameter separately, and also globally for all parameters together.
Specifically, we modify Adam, a popular method for training deep learning
models, with a coefficient that captures properties of the objective function.
Empirically, we show that our method, which we call Eve, outperforms Adam and
other popular methods in training deep neural networks, like convolutional
neural networks for image classification, and recurrent neural networks for
language tasks.
| Hiroaki Hayashi, Jayanth Koushik, Graham Neubig | null | 1611.01505 | null | null |
Topology and Geometry of Half-Rectified Network Optimization | stat.ML cs.LG | The loss surface of deep neural networks has recently attracted interest in
the optimization and machine learning communities as a prime example of
high-dimensional non-convex problem. Some insights were recently gained using
spin glass models and mean-field approximations, but at the expense of strongly
simplifying the nonlinear nature of the model.
In this work, we do not make any such assumption and study conditions on the
data distribution and model architecture that prevent the existence of bad
local minima. Our theoretical work quantifies and formalizes two important
\emph{folklore} facts: (i) the landscape of deep linear networks has a
radically different topology from that of deep half-rectified ones, and (ii)
that the energy landscape in the non-linear case is fundamentally controlled by
the interplay between the smoothness of the data distribution and model
over-parametrization. Our main theoretical contribution is to prove that
half-rectified single layer networks are asymptotically connected, and we
provide explicit bounds that reveal the aforementioned interplay.
The conditioning of gradient descent is the next challenge we address. We
study this question through the geometry of the level sets, and we introduce an
algorithm to efficiently estimate the regularity of such sets on large-scale
networks. Our empirical results show that these level sets remain connected
throughout all the learning phase, suggesting a near convex behavior, but they
become exponentially more curvy as the energy level decays, in accordance to
what is observed in practice with very low curvature attractors.
| C. Daniel Freeman and Joan Bruna | null | 1611.0154 | null | null |
Classification with Ultrahigh-Dimensional Features | stat.ML cs.LG | Although much progress has been made in classification with high-dimensional
features \citep{Fan_Fan:2008, JGuo:2010, CaiSun:2014, PRXu:2014},
classification with ultrahigh-dimensional features, wherein the features much
outnumber the sample size, defies most existing work. This paper introduces a
novel and computationally feasible multivariate screening and classification
method for ultrahigh-dimensional data. Leveraging inter-feature correlations,
the proposed method enables detection of marginally weak and sparse signals and
recovery of the true informative feature set, and achieves asymptotic optimal
misclassification rates. We also show that the proposed procedure provides more
powerful discovery boundaries compared to those in \citet{CaiSun:2014} and
\citet{JJin:2009}. The performance of the proposed procedure is evaluated using
simulation studies and demonstrated via classification of patients with
different post-transplantation renal functional types.
| Yanming Li, Hyokyoung Hong, Jian Kang, Kevin He, Ji Zhu, Yi Li | null | 1611.01541 | null | null |
Automated Generation of Multilingual Clusters for the Evaluation of
Distributed Representations | cs.CL cs.LG | We propose a language-agnostic way of automatically generating sets of
semantically similar clusters of entities along with sets of "outlier"
elements, which may then be used to perform an intrinsic evaluation of word
embeddings in the outlier detection task. We used our methodology to create a
gold-standard dataset, which we call WikiSem500, and evaluated multiple
state-of-the-art embeddings. The results show a correlation between performance
on this dataset and performance on sentiment analysis.
| Philip Blair, Yuval Merhav, and Joel Barry | null | 1611.01547 | null | null |
Quasi-Recurrent Neural Networks | cs.NE cs.AI cs.CL cs.LG | Recurrent neural networks are a powerful tool for modeling sequential data,
but the dependence of each timestep's computation on the previous timestep's
output limits parallelism and makes RNNs unwieldy for very long sequences. We
introduce quasi-recurrent neural networks (QRNNs), an approach to neural
sequence modeling that alternates convolutional layers, which apply in parallel
across timesteps, and a minimalist recurrent pooling function that applies in
parallel across channels. Despite lacking trainable recurrent layers, stacked
QRNNs have better predictive accuracy than stacked LSTMs of the same hidden
size. Due to their increased parallelism, they are up to 16 times faster at
train and test time. Experiments on language modeling, sentiment
classification, and character-level neural machine translation demonstrate
these advantages and underline the viability of QRNNs as a basic building block
for a variety of sequence tasks.
| James Bradbury, Stephen Merity, Caiming Xiong, Richard Socher | null | 1611.01576 | null | null |
Neural Architecture Search with Reinforcement Learning | cs.LG cs.AI cs.NE | Neural networks are powerful and flexible models that work well for many
difficult learning tasks in image, speech and natural language understanding.
Despite their success, neural networks are still hard to design. In this paper,
we use a recurrent network to generate the model descriptions of neural
networks and train this RNN with reinforcement learning to maximize the
expected accuracy of the generated architectures on a validation set. On the
CIFAR-10 dataset, our method, starting from scratch, can design a novel network
architecture that rivals the best human-invented architecture in terms of test
set accuracy. Our CIFAR-10 model achieves a test error rate of 3.65, which is
0.09 percent better and 1.05x faster than the previous state-of-the-art model
that used a similar architectural scheme. On the Penn Treebank dataset, our
model can compose a novel recurrent cell that outperforms the widely-used LSTM
cell, and other state-of-the-art baselines. Our cell achieves a test set
perplexity of 62.4 on the Penn Treebank, which is 3.6 perplexity better than
the previous state-of-the-art model. The cell can also be transferred to the
character language modeling task on PTB and achieves a state-of-the-art
perplexity of 1.214.
| Barret Zoph and Quoc V. Le | null | 1611.01578 | null | null |
Class-prior Estimation for Learning from Positive and Unlabeled Data | cs.LG stat.ML | We consider the problem of estimating the class prior in an unlabeled
dataset. Under the assumption that an additional labeled dataset is available,
the class prior can be estimated by fitting a mixture of class-wise data
distributions to the unlabeled data distribution. However, in practice, such an
additional labeled dataset is often not available. In this paper, we show that,
with additional samples coming only from the positive class, the class prior of
the unlabeled dataset can be estimated correctly. Our key idea is to use
properly penalized divergences for model fitting to cancel the error caused by
the absence of negative samples. We further show that the use of the penalized
$L_1$-distance gives a computationally efficient algorithm with an analytic
solution. The consistency, stability, and estimation error are theoretically
analyzed. Finally, we experimentally demonstrate the usefulness of the proposed
method.
| Marthinus C. du Plessis, Gang Niu, and Masashi Sugiyama | 10.1007/s10994-016-5604-6 | 1611.01586 | null | null |
LipNet: End-to-End Sentence-level Lipreading | cs.LG cs.CL cs.CV | Lipreading is the task of decoding text from the movement of a speaker's
mouth. Traditional approaches separated the problem into two stages: designing
or learning visual features, and prediction. More recent deep lipreading
approaches are end-to-end trainable (Wand et al., 2016; Chung & Zisserman,
2016a). However, existing work on models trained end-to-end perform only word
classification, rather than sentence-level sequence prediction. Studies have
shown that human lipreading performance increases for longer words (Easton &
Basala, 1982), indicating the importance of features capturing temporal context
in an ambiguous communication channel. Motivated by this observation, we
present LipNet, a model that maps a variable-length sequence of video frames to
text, making use of spatiotemporal convolutions, a recurrent network, and the
connectionist temporal classification loss, trained entirely end-to-end. To the
best of our knowledge, LipNet is the first end-to-end sentence-level lipreading
model that simultaneously learns spatiotemporal visual features and a sequence
model. On the GRID corpus, LipNet achieves 95.2% accuracy in sentence-level,
overlapped speaker split task, outperforming experienced human lipreaders and
the previous 86.4% word-level state-of-the-art accuracy (Gergen et al., 2016).
| Yannis M. Assael, Brendan Shillingford, Shimon Whiteson, Nando de
Freitas | null | 1611.01599 | null | null |
Loss-aware Binarization of Deep Networks | cs.NE cs.LG | Deep neural network models, though very powerful and highly successful, are
computationally expensive in terms of space and time. Recently, there have been
a number of attempts on binarizing the network weights and activations. This
greatly reduces the network size, and replaces the underlying multiplications
to additions or even XNOR bit operations. However, existing binarization
schemes are based on simple matrix approximation and ignore the effect of
binarization on the loss. In this paper, we propose a proximal Newton algorithm
with diagonal Hessian approximation that directly minimizes the loss w.r.t. the
binarized weights. The underlying proximal step has an efficient closed-form
solution, and the second-order information can be efficiently obtained from the
second moments already computed by the Adam optimizer. Experiments on both
feedforward and recurrent networks show that the proposed loss-aware
binarization algorithm outperforms existing binarization schemes, and is also
more robust for wide and deep networks.
| Lu Hou, Quanming Yao, James T. Kwok | null | 1611.016 | null | null |
Learning to Play in a Day: Faster Deep Reinforcement Learning by
Optimality Tightening | cs.LG stat.ML | We propose a novel training algorithm for reinforcement learning which
combines the strength of deep Q-learning with a constrained optimization
approach to tighten optimality and encourage faster reward propagation. Our
novel technique makes deep reinforcement learning more practical by drastically
reducing the training time. We evaluate the performance of our approach on the
49 games of the challenging Arcade Learning Environment, and report significant
improvements in both training time and accuracy.
| Frank S. He and Yang Liu and Alexander G. Schwing and Jian Peng | null | 1611.01606 | null | null |
Combining policy gradient and Q-learning | cs.LG cs.AI math.OC stat.ML | Policy gradient is an efficient technique for improving a policy in a
reinforcement learning setting. However, vanilla online variants are on-policy
only and not able to take advantage of off-policy data. In this paper we
describe a new technique that combines policy gradient with off-policy
Q-learning, drawing experience from a replay buffer. This is motivated by
making a connection between the fixed points of the regularized policy gradient
algorithm and the Q-values. This connection allows us to estimate the Q-values
from the action preferences of the policy, to which we apply Q-learning
updates. We refer to the new technique as 'PGQL', for policy gradient and
Q-learning. We also establish an equivalency between action-value fitting
techniques and actor-critic algorithms, showing that regularized policy
gradient techniques can be interpreted as advantage function learning
algorithms. We conclude with some numerical examples that demonstrate improved
data efficiency and stability of PGQL. In particular, we tested PGQL on the
full suite of Atari games and achieved performance exceeding that of both
asynchronous advantage actor-critic (A3C) and Q-learning.
| Brendan O'Donoghue, Remi Munos, Koray Kavukcuoglu and Volodymyr Mnih | null | 1611.01626 | null | null |
Robustly representing uncertainty in deep neural networks through
sampling | cs.LG cs.CV cs.NE q-bio.NC | As deep neural networks (DNNs) are applied to increasingly challenging
problems, they will need to be able to represent their own uncertainty.
Modeling uncertainty is one of the key features of Bayesian methods. Using
Bernoulli dropout with sampling at prediction time has recently been proposed
as an efficient and well performing variational inference method for DNNs.
However, sampling from other multiplicative noise based variational
distributions has not been investigated in depth. We evaluated Bayesian DNNs
trained with Bernoulli or Gaussian multiplicative masking of either the units
(dropout) or the weights (dropconnect). We tested the calibration of the
probabilistic predictions of Bayesian convolutional neural networks (CNNs) on
MNIST and CIFAR-10. Sampling at prediction time increased the calibration of
the DNNs' probabalistic predictions. Sampling weights, whether Gaussian or
Bernoulli, led to more robust representation of uncertainty compared to
sampling of units. However, using either Gaussian or Bernoulli dropout led to
increased test set classification accuracy. Based on these findings we used
both Bernoulli dropout and Gaussian dropconnect concurrently, which we show
approximates the use of a spike-and-slab variational distribution without
increasing the number of learned parameters. We found that spike-and-slab
sampling had higher test set performance than Gaussian dropconnect and more
robustly represented its uncertainty compared to Bernoulli dropout.
| Patrick McClure, Nikolaus Kriegeskorte | null | 1611.01639 | null | null |
Twenty (simple) questions | cs.DM cs.DS cs.IT cs.LG math.CO math.IT | A basic combinatorial interpretation of Shannon's entropy function is via the
"20 questions" game. This cooperative game is played by two players, Alice and
Bob: Alice picks a distribution $\pi$ over the numbers $\{1,\ldots,n\}$, and
announces it to Bob. She then chooses a number $x$ according to $\pi$, and Bob
attempts to identify $x$ using as few Yes/No queries as possible, on average.
An optimal strategy for the "20 questions" game is given by a Huffman code
for $\pi$: Bob's questions reveal the codeword for $x$ bit by bit. This
strategy finds $x$ using fewer than $H(\pi)+1$ questions on average. However,
the questions asked by Bob could be arbitrary. In this paper, we investigate
the following question: Are there restricted sets of questions that match the
performance of Huffman codes, either exactly or approximately?
Our first main result shows that for every distribution $\pi$, Bob has a
strategy that uses only questions of the form "$x < c$?" and "$x = c$?", and
uncovers $x$ using at most $H(\pi)+1$ questions on average, matching the
performance of Huffman codes in this sense. We also give a natural set of
$O(rn^{1/r})$ questions that achieve a performance of at most $H(\pi)+r$, and
show that $\Omega(rn^{1/r})$ questions are required to achieve such a
guarantee.
Our second main result gives a set $\mathcal{Q}$ of $1.25^{n+o(n)}$ questions
such that for every distribution $\pi$, Bob can implement an optimal strategy
for $\pi$ using only questions from $\mathcal{Q}$. We also show that
$1.25^{n-o(n)}$ questions are needed, for infinitely many $n$. If we allow a
small slack of $r$ over the optimal strategy, then roughly $(rn)^{\Theta(1/r)}$
questions are necessary and sufficient.
| Yuval Dagan, Yuval Filmus, Ariel Gabizon, Shay Moran | null | 1611.01655 | null | null |
Generative Multi-Adversarial Networks | cs.LG cs.MA cs.NE | Generative adversarial networks (GANs) are a framework for producing a
generative model by way of a two-player minimax game. In this paper, we propose
the \emph{Generative Multi-Adversarial Network} (GMAN), a framework that
extends GANs to multiple discriminators. In previous work, the successful
training of GANs requires modifying the minimax objective to accelerate
training early on. In contrast, GMAN can be reliably trained with the original,
untampered objective. We explore a number of design perspectives with the
discriminator role ranging from formidable adversary to forgiving teacher.
Image generation tasks comparing the proposed framework to standard GANs
demonstrate GMAN produces higher quality samples in a fraction of the
iterations when measured by a pairwise GAM-type metric.
| Ishan Durugkar, Ian Gemp, Sridhar Mahadevan | null | 1611.01673 | null | null |
Comparing learning algorithms in neural network for diagnosing
cardiovascular disease | cs.LG cs.NE | Today data mining techniques are exploited in medical science for diagnosing,
overcoming and treating diseases. Neural network is one of the techniques which
are widely used for diagnosis in medical field. In this article efficiency of
nine algorithms, which are basis of neural network learning in diagnosing
cardiovascular diseases, will be assessed. Algorithms are assessed in terms of
accuracy, sensitivity, transparency, AROC and convergence rate by means of 10
fold cross validation. The results suggest that in training phase, Lonberg-M
algorithm has the best efficiency in terms of all metrics, algorithm OSS has
maximum accuracy in testing phase, algorithm SCG has the maximum transparency
and algorithm CGB has the maximum sensitivity.
| Mirmorsal Madani | null | 1611.01678 | null | null |
Oracle-Efficient Online Learning and Auction Design | cs.LG cs.DS cs.GT | We consider the design of computationally efficient online learning
algorithms in an adversarial setting in which the learner has access to an
offline optimization oracle. We present an algorithm called Generalized
Follow-the-Perturbed-Leader and provide conditions under which it is
oracle-efficient while achieving vanishing regret. Our results make significant
progress on an open problem raised by Hazan and Koren, who showed that
oracle-efficient algorithms do not exist in general and asked whether one can
identify properties under which oracle-efficient online learning may be
possible.
Our auction-design framework considers an auctioneer learning an optimal
auction for a sequence of adversarially selected valuations with the goal of
achieving revenue that is almost as good as the optimal auction in hindsight,
among a class of auctions. We give oracle-efficient learning results for: (1)
VCG auctions with bidder-specific reserves in single-parameter settings, (2)
envy-free item pricing in multi-item auctions, and (3) s-level auctions of
Morgenstern and Roughgarden for single-item settings. The last result leads to
an approximation of the overall optimal Myerson auction when bidders'
valuations are drawn according to a fast-mixing Markov process, extending prior
work that only gave such guarantees for the i.i.d. setting.
Finally, we derive various extensions, including: (1) oracle-efficient
algorithms for the contextual learning setting in which the learner has access
to side information (such as bidder demographics), (2) learning with
approximate oracles such as those based on Maximal-in-Range algorithms, and (3)
no-regret bidding in simultaneous auctions, resolving an open problem of
Daskalakis and Syrgkanis.
| Miroslav Dud\'ik, Nika Haghtalab, Haipeng Luo, Robert E. Schapire,
Vasilis Syrgkanis, Jennifer Wortman Vaughan | null | 1611.01688 | null | null |
TopicRNN: A Recurrent Neural Network with Long-Range Semantic Dependency | cs.CL cs.AI cs.LG stat.ML | In this paper, we propose TopicRNN, a recurrent neural network (RNN)-based
language model designed to directly capture the global semantic meaning
relating words in a document via latent topics. Because of their sequential
nature, RNNs are good at capturing the local structure of a word sequence -
both semantic and syntactic - but might face difficulty remembering long-range
dependencies. Intuitively, these long-range dependencies are of semantic
nature. In contrast, latent topic models are able to capture the global
underlying semantic structure of a document but do not account for word
ordering. The proposed TopicRNN model integrates the merits of RNNs and latent
topic models: it captures local (syntactic) dependencies using an RNN and
global (semantic) dependencies using latent topics. Unlike previous work on
contextual RNN language modeling, our model is learned end-to-end. Empirical
results on word prediction show that TopicRNN outperforms existing contextual
RNN baselines. In addition, TopicRNN can be used as an unsupervised feature
extractor for documents. We do this for sentiment analysis on the IMDB movie
review dataset and report an error rate of $6.28\%$. This is comparable to the
state-of-the-art $5.91\%$ resulting from a semi-supervised approach. Finally,
TopicRNN also yields sensible topics, making it a useful alternative to
document models such as latent Dirichlet allocation.
| Adji B. Dieng, Chong Wang, Jianfeng Gao, John Paisley | null | 1611.01702 | null | null |
Detecting Dependencies in Sparse, Multivariate Databases Using
Probabilistic Programming and Non-parametric Bayes | stat.ML cs.AI cs.LG | Datasets with hundreds of variables and many missing values are commonplace.
In this setting, it is both statistically and computationally challenging to
detect true predictive relationships between variables and also to suppress
false positives. This paper proposes an approach that combines probabilistic
programming, information theory, and non-parametric Bayes. It shows how to use
Bayesian non-parametric modeling to (i) build an ensemble of joint probability
models for all the variables; (ii) efficiently detect marginal independencies;
and (iii) estimate the conditional mutual information between arbitrary subsets
of variables, subject to a broad class of constraints. Users can access these
capabilities using BayesDB, a probabilistic programming platform for
probabilistic data analysis, by writing queries in a simple, SQL-like language.
This paper demonstrates empirically that the method can (i) detect
context-specific (in)dependencies on challenging synthetic problems and (ii)
yield improved sensitivity and specificity over baselines from statistics and
machine learning, on a real-world database of over 300 sparsely observed
indicators of macroeconomic development and public health.
| Feras Saad, Vikash Mansinghka | null | 1611.01708 | null | null |
Beyond Fine Tuning: A Modular Approach to Learning on Small Data | cs.LG cs.CL | In this paper we present a technique to train neural network models on small
amounts of data. Current methods for training neural networks on small amounts
of rich data typically rely on strategies such as fine-tuning a pre-trained
neural network or the use of domain-specific hand-engineered features. Here we
take the approach of treating network layers, or entire networks, as modules
and combine pre-trained modules with untrained modules, to learn the shift in
distributions between data sets. The central impact of using a modular approach
comes from adding new representations to a network, as opposed to replacing
representations via fine-tuning. Using this technique, we are able surpass
results using standard fine-tuning transfer learning approaches, and we are
also able to significantly increase performance over such approaches when using
smaller amounts of data.
| Ark Anderson, Kyle Shaffer, Artem Yankov, Court D. Corley, Nathan O.
Hodas | null | 1611.01714 | null | null |
Learning to Draw Samples: With Application to Amortized MLE for
Generative Adversarial Learning | stat.ML cs.LG | We propose a simple algorithm to train stochastic neural networks to draw
samples from given target distributions for probabilistic inference. Our method
is based on iteratively adjusting the neural network parameters so that the
output changes along a Stein variational gradient that maximumly decreases the
KL divergence with the target distribution. Our method works for any target
distribution specified by their unnormalized density function, and can train
any black-box architectures that are differentiable in terms of the parameters
we want to adapt. As an application of our method, we propose an amortized MLE
algorithm for training deep energy model, where a neural sampler is adaptively
trained to approximate the likelihood function. Our method mimics an
adversarial game between the deep energy model and the neural sampler, and
obtains realistic-looking images competitive with the state-of-the-art results.
| Dilin Wang, Qiang Liu | null | 1611.01722 | null | null |
Words or Characters? Fine-grained Gating for Reading Comprehension | cs.CL cs.LG | Previous work combines word-level and character-level representations using
concatenation or scalar weighting, which is suboptimal for high-level tasks
like reading comprehension. We present a fine-grained gating mechanism to
dynamically combine word-level and character-level representations based on
properties of the words. We also extend the idea of fine-grained gating to
modeling the interaction between questions and paragraphs for reading
comprehension. Experiments show that our approach can improve the performance
on reading comprehension tasks, achieving new state-of-the-art results on the
Children's Book Test dataset. To demonstrate the generality of our gating
mechanism, we also show improved results on a social media tag prediction task.
| Zhilin Yang, Bhuwan Dhingra, Ye Yuan, Junjie Hu, William W. Cohen,
Ruslan Salakhutdinov | null | 1611.01724 | null | null |
LSTM-Based System-Call Language Modeling and Robust Ensemble Method for
Designing Host-Based Intrusion Detection Systems | cs.CR cs.LG | In computer security, designing a robust intrusion detection system is one of
the most fundamental and important problems. In this paper, we propose a
system-call language-modeling approach for designing anomaly-based host
intrusion detection systems. To remedy the issue of high false-alarm rates
commonly arising in conventional methods, we employ a novel ensemble method
that blends multiple thresholding classifiers into a single one, making it
possible to accumulate 'highly normal' sequences. The proposed system-call
language model has various advantages leveraged by the fact that it can learn
the semantic meaning and interactions of each system call that existing methods
cannot effectively consider. Through diverse experiments on public benchmark
datasets, we demonstrate the validity and effectiveness of the proposed method.
Moreover, we show that our model possesses high portability, which is one of
the key aspects of realizing successful intrusion detection systems.
| Gyuwan Kim, Hayoon Yi, Jangho Lee, Yunheung Paek, Sungroh Yoon | null | 1611.01726 | null | null |
Learning a Static Analyzer from Data | cs.PL cs.LG | To be practically useful, modern static analyzers must precisely model the
effect of both, statements in the programming language as well as frameworks
used by the program under analysis. While important, manually addressing these
challenges is difficult for at least two reasons: (i) the effects on the
overall analysis can be non-trivial, and (ii) as the size and complexity of
modern libraries increase, so is the number of cases the analysis must handle.
In this paper we present a new, automated approach for creating static
analyzers: instead of manually providing the various inference rules of the
analyzer, the key idea is to learn these rules from a dataset of programs. Our
method consists of two ingredients: (i) a synthesis algorithm capable of
learning a candidate analyzer from a given dataset, and (ii) a counter-example
guided learning procedure which generates new programs beyond those in the
initial dataset, critical for discovering corner cases and ensuring the learned
analysis generalizes to unseen programs.
We implemented and instantiated our approach to the task of learning
JavaScript static analysis rules for a subset of points-to analysis and for
allocation sites analysis. These are challenging yet important problems that
have received significant research attention. We show that our approach is
effective: our system automatically discovered practical and useful inference
rules for many cases that are tricky to manually identify and are missed by
state-of-the-art, manually tuned analyzers.
| Pavol Bielik, Veselin Raychev, Martin Vechev | null | 1611.01752 | null | null |
Learning to Act by Predicting the Future | cs.LG cs.AI cs.CV | We present an approach to sensorimotor control in immersive environments. Our
approach utilizes a high-dimensional sensory stream and a lower-dimensional
measurement stream. The cotemporal structure of these streams provides a rich
supervisory signal, which enables training a sensorimotor control model by
interacting with the environment. The model is trained using supervised
learning techniques, but without extraneous supervision. It learns to act based
on raw sensory input from a complex three-dimensional environment. The
presented formulation enables learning without a fixed goal at training time,
and pursuing dynamically changing goals at test time. We conduct extensive
experiments in three-dimensional simulations based on the classical
first-person game Doom. The results demonstrate that the presented approach
outperforms sophisticated prior formulations, particularly on challenging
tasks. The results also show that trained models successfully generalize across
environments and goals. A model trained using the presented approach won the
Full Deathmatch track of the Visual Doom AI Competition, which was held in
previously unseen environments.
| Alexey Dosovitskiy and Vladlen Koltun | null | 1611.01779 | null | null |
Learning to superoptimize programs | cs.LG | Code super-optimization is the task of transforming any given program to a
more efficient version while preserving its input-output behaviour. In some
sense, it is similar to the paraphrase problem from natural language processing
where the intention is to change the syntax of an utterance without changing
its semantics. Code-optimization has been the subject of years of research that
has resulted in the development of rule-based transformation strategies that
are used by compilers. More recently, however, a class of stochastic search
based methods have been shown to outperform these strategies. This approach
involves repeated sampling of modifications to the program from a proposal
distribution, which are accepted or rejected based on whether they preserve
correctness, and the improvement they achieve. These methods, however, neither
learn from past behaviour nor do they try to leverage the semantics of the
program under consideration. Motivated by this observation, we present a novel
learning based approach for code super-optimization. Intuitively, our method
works by learning the proposal distribution using unbiased estimators of the
gradient of the expected improvement. Experiments on benchmarks comprising of
automatically generated as well as existing ("Hacker's Delight") programs show
that the proposed method is able to significantly outperform state of the art
approaches for code super-optimization.
| Rudy Bunel, Alban Desmaison, M. Pawan Kumar, Philip H.S. Torr and
Pushmeet Kohli | null | 1611.01787 | null | null |
Modular Multitask Reinforcement Learning with Policy Sketches | cs.LG cs.NE | We describe a framework for multitask deep reinforcement learning guided by
policy sketches. Sketches annotate tasks with sequences of named subtasks,
providing information about high-level structural relationships among tasks but
not how to implement them---specifically not providing the detailed guidance
used by much previous work on learning policy abstractions for RL (e.g.
intermediate rewards, subtask completion signals, or intrinsic motivations). To
learn from sketches, we present a model that associates every subtask with a
modular subpolicy, and jointly maximizes reward over full task-specific
policies by tying parameters across shared subpolicies. Optimization is
accomplished via a decoupled actor--critic training objective that facilitates
learning common behaviors from multiple dissimilar reward functions. We
evaluate the effectiveness of our approach in three environments featuring both
discrete and continuous control, and with sparse rewards that can be obtained
only after completing a number of high-level subgoals. Experiments show that
using our approach to learn policies guided by sketches gives better
performance than existing techniques for learning task-specific or shared
policies, while naturally inducing a library of interpretable primitive
behaviors that can be recombined to rapidly adapt to new tasks.
| Jacob Andreas and Dan Klein and Sergey Levine | null | 1611.01796 | null | null |
Generative Adversarial Networks as Variational Training of Energy Based
Models | cs.LG | In this paper, we study deep generative models for effective unsupervised
learning. We propose VGAN, which works by minimizing a variational lower bound
of the negative log likelihood (NLL) of an energy based model (EBM), where the
model density $p(\mathbf{x})$ is approximated by a variational distribution
$q(\mathbf{x})$ that is easy to sample from. The training of VGAN takes a two
step procedure: given $p(\mathbf{x})$, $q(\mathbf{x})$ is updated to maximize
the lower bound; $p(\mathbf{x})$ is then updated one step with samples drawn
from $q(\mathbf{x})$ to decrease the lower bound. VGAN is inspired by the
generative adversarial networks (GANs), where $p(\mathbf{x})$ corresponds to
the discriminator and $q(\mathbf{x})$ corresponds to the generator, but with
several notable differences. We hence name our model variational GANs (VGANs).
VGAN provides a practical solution to training deep EBMs in high dimensional
space, by eliminating the need of MCMC sampling. From this view, we are also
able to identify causes to the difficulty of training GANs and propose viable
solutions. \footnote{Experimental code is available at
https://github.com/Shuangfei/vgan}
| Shuangfei Zhai, Yu Cheng, Rogerio Feris, Zhongfei Zhang | null | 1611.01799 | null | null |
Entropy-SGD: Biasing Gradient Descent Into Wide Valleys | cs.LG stat.ML | This paper proposes a new optimization algorithm called Entropy-SGD for
training deep neural networks that is motivated by the local geometry of the
energy landscape. Local extrema with low generalization error have a large
proportion of almost-zero eigenvalues in the Hessian with very few positive or
negative eigenvalues. We leverage upon this observation to construct a
local-entropy-based objective function that favors well-generalizable solutions
lying in large flat regions of the energy landscape, while avoiding
poorly-generalizable solutions located in the sharp valleys. Conceptually, our
algorithm resembles two nested loops of SGD where we use Langevin dynamics in
the inner loop to compute the gradient of the local entropy before each update
of the weights. We show that the new objective has a smoother energy landscape
and show improved generalization over SGD using uniform stability, under
certain assumptions. Our experiments on convolutional and recurrent networks
demonstrate that Entropy-SGD compares favorably to state-of-the-art techniques
in terms of generalization error and training time.
| Pratik Chaudhari, Anna Choromanska, Stefano Soatto, Yann LeCun, Carlo
Baldassi, Christian Borgs, Jennifer Chayes, Levent Sagun, Riccardo Zecchina | null | 1611.01838 | null | null |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.