title
stringlengths 7
246
| abstract
stringlengths 6
3.31k
|
|---|---|
Weight Initialization of Deep Neural Networks(DNNs) using Data
Statistics | Deep neural networks (DNNs) form the backbone of almost every
state-of-the-art technique in the fields such as computer vision, speech
processing, and text analysis. The recent advances in computational technology
have made the use of DNNs more practical. Despite the overwhelming performances
by DNN and the advances in computational technology, it is seen that very few
researchers try to train their models from the scratch. Training of DNNs still
remains a difficult and tedious job. The main challenges that researchers face
during training of DNNs are the vanishing/exploding gradient problem and the
highly non-convex nature of the objective function which has up to million
variables. The approaches suggested in He and Xavier solve the vanishing
gradient problem by providing a sophisticated initialization technique. These
approaches have been quite effective and have achieved good results on standard
datasets, but these same approaches do not work very well on more practical
datasets. We think the reason for this is not making use of data statistics for
initializing the network weights. Optimizing such a high dimensional loss
function requires careful initialization of network weights. In this work, we
propose a data dependent initialization and analyze its performance against the
standard initialization techniques such as He and Xavier. We performed our
experiments on some practical datasets and the results show our algorithm's
superior classification accuracy.
|
Certifying Some Distributional Robustness with Principled Adversarial
Training | Neural networks are vulnerable to adversarial examples and researchers have
proposed many heuristic attack and defense mechanisms. We address this problem
through the principled lens of distributionally robust optimization, which
guarantees performance under adversarial input perturbations. By considering a
Lagrangian penalty formulation of perturbing the underlying data distribution
in a Wasserstein ball, we provide a training procedure that augments model
parameter updates with worst-case perturbations of training data. For smooth
losses, our procedure provably achieves moderate levels of robustness with
little computational or statistical cost relative to empirical risk
minimization. Furthermore, our statistical guarantees allow us to efficiently
certify robustness for the population loss. For imperceptible perturbations,
our method matches or outperforms heuristic approaches.
|
Regularization approaches for support vector machines with applications
to biomedical data | The support vector machine (SVM) is a widely used machine learning tool for
classification based on statistical learning theory. Given a set of training
data, the SVM finds a hyperplane that separates two different classes of data
points by the largest distance. While the standard form of SVM uses L2-norm
regularization, other regularization approaches are particularly attractive for
biomedical datasets where, for example, sparsity and interpretability of the
classifier's coefficient values are highly desired features. Therefore, in this
paper we consider different types of regularization approaches for SVMs, and
explore them in both synthetic and real biomedical datasets.
|
Variational Continual Learning | This paper develops variational continual learning (VCL), a simple but
general framework for continual learning that fuses online variational
inference (VI) and recent advances in Monte Carlo VI for neural networks. The
framework can successfully train both deep discriminative models and deep
generative models in complex continual learning settings where existing tasks
evolve over time and entirely new tasks emerge. Experimental results show that
VCL outperforms state-of-the-art continual learning methods on a variety of
tasks, avoiding catastrophic forgetting in a fully automatic way.
|
Dimensionality reduction methods for molecular simulations | Molecular simulations produce very high-dimensional data-sets with millions
of data points. As analysis methods are often unable to cope with so many
dimensions, it is common to use dimensionality reduction and clustering methods
to reach a reduced representation of the data. Yet these methods often fail to
capture the most important features necessary for the construction of a Markov
model. Here we demonstrate the results of various dimensionality reduction
methods on two simulation data-sets, one of protein folding and another of
protein-ligand binding. The methods tested include a k-means clustering
variant, a non-linear auto encoder, principal component analysis and tICA. The
dimension-reduced data is then used to estimate the implied timescales of the
slowest process by a Markov state model analysis to assess the quality of the
projection. The projected dimensions learned from the data are visualized to
demonstrate which conformations the various methods choose to represent the
molecular process.
|
On the Consistency of Quick Shift | Quick Shift is a popular mode-seeking and clustering algorithm. We present
finite sample statistical consistency guarantees for Quick Shift on mode and
cluster recovery under mild distributional assumptions. We then apply our
results to construct a consistent modal regression algorithm.
|
Discrepancy-Based Algorithms for Non-Stationary Rested Bandits | We study the multi-armed bandit problem where the rewards are realizations of
general non-stationary stochastic processes, a setting that generalizes many
existing lines of work and analyses. In particular, we present a theoretical
analysis and derive regret guarantees for rested bandits in which the reward
distribution of each arm changes only when we pull that arm. Remarkably, our
regret bounds are logarithmic in the number of rounds under several natural
conditions. We introduce a new algorithm based on classical UCB ideas combined
with the notion of weighted discrepancy, a useful tool for measuring the
non-stationarity of a stochastic process. We show that the notion of
discrepancy can be used to design very general algorithms and a unified
framework for the analysis of multi-armed rested bandit problems with
non-stationary rewards. In particular, we show that we can recover the regret
guarantees of many specific instances of bandit problems with non-stationary
rewards that have been studied in the literature. We also provide experiments
demonstrating that our algorithms can enjoy a significant improvement in
practice compared to standard benchmarks.
|
Regularization for Deep Learning: A Taxonomy | Regularization is one of the crucial ingredients of deep learning, yet the
term regularization has various definitions, and regularization methods are
often studied separately from each other. In our work we present a systematic,
unifying taxonomy to categorize existing methods. We distinguish methods that
affect data, network architectures, error terms, regularization terms, and
optimization procedures. We do not provide all details about the listed
methods; instead, we present an overview of how the methods can be sorted into
meaningful categories and sub-categories. This helps revealing links and
fundamental similarities between them. Finally, we include practical
recommendations both for users and for developers of new regularization
methods.
|
Kernel Graph Convolutional Neural Networks | Graph kernels have been successfully applied to many graph classification
problems. Typically, a kernel is first designed, and then an SVM classifier is
trained based on the features defined implicitly by this kernel. This two-stage
approach decouples data representation from learning, which is suboptimal. On
the other hand, Convolutional Neural Networks (CNNs) have the capability to
learn their own features directly from the raw data during training.
Unfortunately, they cannot handle irregular data such as graphs. We address
this challenge by using graph kernels to embed meaningful local neighborhoods
of the graphs in a continuous vector space. A set of filters is then convolved
with these patches, pooled, and the output is then passed to a feedforward
network. With limited parameter tuning, our approach outperforms strong
baselines on 7 out of 10 benchmark datasets.
|
Training Probabilistic Spiking Neural Networks with First-to-spike
Decoding | Third-generation neural networks, or Spiking Neural Networks (SNNs), aim at
harnessing the energy efficiency of spike-domain processing by building on
computing elements that operate on, and exchange, spikes. In this paper, the
problem of training a two-layer SNN is studied for the purpose of
classification, under a Generalized Linear Model (GLM) probabilistic neural
model that was previously considered within the computational neuroscience
literature. Conventional classification rules for SNNs operate offline based on
the number of output spikes at each output neuron. In contrast, a novel
training method is proposed here for a first-to-spike decoding rule, whereby
the SNN can perform an early classification decision once spike firing is
detected at an output neuron. Numerical results bring insights into the optimal
parameter selection for the GLM neuron and on the accuracy-complexity trade-off
performance of conventional and first-to-spike decoding.
|
Contextual Regression: An Accurate and Conveniently Interpretable
Nonlinear Model for Mining Discovery from Scientific Data | Machine learning algorithms such as linear regression, SVM and neural network
have played an increasingly important role in the process of scientific
discovery. However, none of them is both interpretable and accurate on
nonlinear datasets. Here we present contextual regression, a method that joins
these two desirable properties together using a hybrid architecture of neural
network embedding and dot product layer. We demonstrate its high prediction
accuracy and sensitivity through the task of predictive feature selection on a
simulated dataset and the application of predicting open chromatin sites in the
human genome. On the simulated data, our method achieved high fidelity recovery
of feature contributions under random noise levels up to 200%. On the open
chromatin dataset, the application of our method not only outperformed the
state of the art method in terms of accuracy, but also unveiled two previously
unfound open chromatin related histone marks. Our method can fill the blank of
accurate and interpretable nonlinear modeling in scientific data mining tasks.
|
Attacking the Madry Defense Model with $L_1$-based Adversarial Examples | The Madry Lab recently hosted a competition designed to test the robustness
of their adversarially trained MNIST model. Attacks were constrained to perturb
each pixel of the input image by a scaled maximal $L_\infty$ distortion
$\epsilon$ = 0.3. This discourages the use of attacks which are not optimized
on the $L_\infty$ distortion metric. Our experimental results demonstrate that
by relaxing the $L_\infty$ constraint of the competition, the elastic-net
attack to deep neural networks (EAD) can generate transferable adversarial
examples which, despite their high average $L_\infty$ distortion, have minimal
visual distortion. These results call into question the use of $L_\infty$ as a
sole measure for visual distortion, and further demonstrate the power of EAD at
generating robust adversarial examples.
|
Linearly convergent stochastic heavy ball method for minimizing
generalization error | In this work we establish the first linear convergence result for the
stochastic heavy ball method. The method performs SGD steps with a fixed
stepsize, amended by a heavy ball momentum term. In the analysis, we focus on
minimizing the expected loss and not on finite-sum minimization, which is
typically a much harder problem. While in the analysis we constrain ourselves
to quadratic loss, the overall objective is not necessarily strongly convex.
|
Implicit Causal Models for Genome-wide Association Studies | Progress in probabilistic generative models has accelerated, developing
richer models with neural architectures, implicit densities, and with scalable
algorithms for their Bayesian inference. However, there has been limited
progress in models that capture causal relationships, for example, how
individual genetic factors cause major human diseases. In this work, we focus
on two challenges in particular: How do we build richer causal models, which
can capture highly nonlinear relationships and interactions between multiple
causes? How do we adjust for latent confounders, which are variables
influencing both cause and effect and which prevent learning of causal
relationships? To address these challenges, we synthesize ideas from causality
and modern probabilistic modeling. For the first, we describe implicit causal
models, a class of causal models that leverages neural architectures with an
implicit density. For the second, we describe an implicit causal model that
adjusts for confounders by sharing strength across examples. In experiments, we
scale Bayesian inference on up to a billion genetic measurements. We achieve
state of the art accuracy for identifying causal factors: we significantly
outperform existing genetics methods by an absolute difference of 15-45.3%.
|
Predicting Head Movement in Panoramic Video: A Deep Reinforcement
Learning Approach | Panoramic video provides immersive and interactive experience by enabling
humans to control the field of view (FoV) through head movement (HM). Thus, HM
plays a key role in modeling human attention on panoramic video. This paper
establishes a database collecting subjects' HM in panoramic video sequences.
From this database, we find that the HM data are highly consistent across
subjects. Furthermore, we find that deep reinforcement learning (DRL) can be
applied to predict HM positions, via maximizing the reward of imitating human
HM scanpaths through the agent's actions. Based on our findings, we propose a
DRL-based HM prediction (DHP) approach with offline and online versions, called
offline-DHP and online-DHP. In offline-DHP, multiple DRL workflows are run to
determine potential HM positions at each panoramic frame. Then, a heat map of
the potential HM positions, named the HM map, is generated as the output of
offline-DHP. In online-DHP, the next HM position of one subject is estimated
given the currently observed HM position, which is achieved by developing a DRL
algorithm upon the learned offline-DHP model. Finally, the experiments validate
that our approach is effective in both offline and online prediction of HM
positions for panoramic video, and that the learned offline-DHP model can
improve the performance of online-DHP.
|
PixelDefend: Leveraging Generative Models to Understand and Defend
against Adversarial Examples | Adversarial perturbations of normal images are usually imperceptible to
humans, but they can seriously confuse state-of-the-art machine learning
models. What makes them so special in the eyes of image classifiers? In this
paper, we show empirically that adversarial examples mainly lie in the low
probability regions of the training distribution, regardless of attack types
and targeted models. Using statistical hypothesis testing, we find that modern
neural density models are surprisingly good at detecting imperceptible image
perturbations. Based on this discovery, we devised PixelDefend, a new approach
that purifies a maliciously perturbed image by moving it back towards the
distribution seen in the training data. The purified image is then run through
an unmodified classifier, making our method agnostic to both the classifier and
the attacking method. As a result, PixelDefend can be used to protect already
deployed models and be combined with other model-specific defenses. Experiments
show that our method greatly improves resilience across a wide variety of
state-of-the-art attacking methods, increasing accuracy on the strongest attack
from 63% to 84% for Fashion MNIST and from 32% to 70% for CIFAR-10.
|
Communication-Avoiding Optimization Methods for Distributed
Massive-Scale Sparse Inverse Covariance Estimation | Across a variety of scientific disciplines, sparse inverse covariance
estimation is a popular tool for capturing the underlying dependency
relationships in multivariate data. Unfortunately, most estimators are not
scalable enough to handle the sizes of modern high-dimensional data sets (often
on the order of terabytes), and assume Gaussian samples. To address these
deficiencies, we introduce HP-CONCORD, a highly scalable optimization method
for estimating a sparse inverse covariance matrix based on a regularized
pseudolikelihood framework, without assuming Gaussianity. Our parallel proximal
gradient method uses a novel communication-avoiding linear algebra algorithm
and runs across a multi-node cluster with up to 1k nodes (24k cores), achieving
parallel scalability on problems with up to ~819 billion parameters (1.28
million dimensions); even on a single node, HP-CONCORD demonstrates
scalability, outperforming a state-of-the-art method. We also use HP-CONCORD to
estimate the underlying dependency structure of the brain from fMRI data, and
use the result to identify functional regions automatically. The results show
good agreement with a clustering from the neuroscience literature.
|
Riemannian Optimization via Frank-Wolfe Methods | We study projection-free methods for constrained Riemannian optimization. In
particular, we propose the Riemannian Frank-Wolfe (RFW) method. We analyze
non-asymptotic convergence rates of RFW to an optimum for (geodesically) convex
problems, and to a critical point for nonconvex objectives. We also present a
practical setting under which RFW can attain a linear convergence rate. As a
concrete example, we specialize RFW to the manifold of positive definite
matrices and apply it to two tasks: (i) computing the matrix geometric mean
(Riemannian centroid); and (ii) computing the Bures-Wasserstein barycenter.
Both tasks involve geodesically convex interval constraints, for which we show
that the Riemannian "linear" oracle required by RFW admits a closed-form
solution; this result may be of independent interest. We further specialize RFW
to the special orthogonal group and show that here too, the Riemannian "linear"
oracle can be solved in closed form. Here, we describe an application to the
synchronization of data matrices (Procrustes problem). We complement our
theoretical results with an empirical comparison of RFW against
state-of-the-art Riemannian optimization methods and observe that RFW performs
competitively on the task of computing Riemannian centroids.
|
Tensorizing Generative Adversarial Nets | Generative Adversarial Network (GAN) and its variants exhibit
state-of-the-art performance in the class of generative models. To capture
higher-dimensional distributions, the common learning procedure requires high
computational complexity and a large number of parameters. The problem of
employing such massive framework arises when deploying it on a platform with
limited computational power such as mobile phones. In this paper, we present a
new generative adversarial framework by representing each layer as a tensor
structure connected by multilinear operations, aiming to reduce the number of
model parameters by a large factor while preserving the generative performance
and sample quality. To learn the model, we employ an efficient algorithm which
alternatively optimizes both discriminator and generator. Experimental outcomes
demonstrate that our model can achieve high compression rate for model
parameters up to $35$ times when compared to the original GAN for MNIST
dataset.
|
Sequence-to-Sequence ASR Optimization via Reinforcement Learning | Despite the success of sequence-to-sequence approaches in automatic speech
recognition (ASR) systems, the models still suffer from several problems,
mainly due to the mismatch between the training and inference conditions. In
the sequence-to-sequence architecture, the model is trained to predict the
grapheme of the current time-step given the input of speech signal and the
ground-truth grapheme history of the previous time-steps. However, it remains
unclear how well the model approximates real-world speech during inference.
Thus, generating the whole transcription from scratch based on previous
predictions is complicated and errors can propagate over time. Furthermore, the
model is optimized to maximize the likelihood of training data instead of error
rate evaluation metrics that actually quantify recognition quality. This paper
presents an alternative strategy for training sequence-to-sequence ASR models
by adopting the idea of reinforcement learning (RL). Unlike the standard
training scheme with maximum likelihood estimation, our proposed approach
utilizes the policy gradient algorithm. We can (1) sample the whole
transcription based on the model's prediction in the training process and (2)
directly optimize the model with negative Levenshtein distance as the reward.
Experimental results demonstrate that we significantly improved the performance
compared to a model trained only with maximum likelihood estimation.
|
Transfer Learning to Learn with Multitask Neural Model Search | Deep learning models require extensive architecture design exploration and
hyperparameter optimization to perform well on a given task. The exploration of
the model design space is often made by a human expert, and optimized using a
combination of grid search and search heuristics over a large space of possible
choices. Neural Architecture Search (NAS) is a Reinforcement Learning approach
that has been proposed to automate architecture design. NAS has been
successfully applied to generate Neural Networks that rival the best
human-designed architectures. However, NAS requires sampling, constructing, and
training hundreds to thousands of models to achieve well-performing
architectures. This procedure needs to be executed from scratch for each new
task. The application of NAS to a wide set of tasks currently lacks a way to
transfer generalizable knowledge across tasks. In this paper, we present the
Multitask Neural Model Search (MNMS) controller. Our goal is to learn a
generalizable framework that can condition model construction on successful
model searches for previously seen tasks, thus significantly speeding up the
search for new tasks. We demonstrate that MNMS can conduct an automated
architecture search for multiple tasks simultaneously while still learning
well-performing, specialized models for each task. We then show that
pre-trained MNMS controllers can transfer learning to new tasks. By leveraging
knowledge from previous searches, we find that pre-trained MNMS models start
from a better location in the search space and reduce search time on unseen
tasks, while still discovering models that outperform published human-designed
models.
|
Generative Adversarial Source Separation | Generative source separation methods such as non-negative matrix
factorization (NMF) or auto-encoders, rely on the assumption of an output
probability density. Generative Adversarial Networks (GANs) can learn data
distributions without needing a parametric assumption on the output density. We
show on a speech source separation experiment that, a multi-layer perceptron
trained with a Wasserstein-GAN formulation outperforms NMF, auto-encoders
trained with maximum likelihood, and variational auto-encoders in terms of
source to distortion ratio.
|
Stochastic variance reduced multiplicative update for nonnegative matrix
factorization | Nonnegative matrix factorization (NMF), a dimensionality reduction and factor
analysis method, is a special case in which factor matrices have low-rank
nonnegative constraints. Considering the stochastic learning in NMF, we
specifically address the multiplicative update (MU) rule, which is the most
popular, but which has slow convergence property. This present paper introduces
on the stochastic MU rule a variance-reduced technique of stochastic gradient.
Numerical comparisons suggest that our proposed algorithms robustly outperform
state-of-the-art algorithms across different synthetic and real-world datasets.
|
How deep learning works --The geometry of deep learning | Why and how that deep learning works well on different tasks remains a
mystery from a theoretical perspective. In this paper we draw a geometric
picture of the deep learning system by finding its analogies with two existing
geometric structures, the geometry of quantum computations and the geometry of
the diffeomorphic template matching. In this framework, we give the geometric
structures of different deep learning systems including convolutional neural
networks, residual networks, recursive neural networks, recurrent neural
networks and the equilibrium prapagation framework. We can also analysis the
relationship between the geometrical structures and their performance of
different networks in an algorithmic level so that the geometric framework may
guide the design of the structures and algorithms of deep learning systems.
|
Understanding GANs: the LQG Setting | Generative Adversarial Networks (GANs) have become a popular method to learn
a probability model from data. In this paper, we aim to provide an
understanding of some of the basic issues surrounding GANs including their
formulation, generalization and stability on a simple benchmark where the data
has a high-dimensional Gaussian distribution. Even in this simple benchmark,
the GAN problem has not been well-understood as we observe that existing
state-of-the-art GAN architectures may fail to learn a proper generative
distribution owing to (1) stability issues (i.e., convergence to bad local
solutions or not converging at all), (2) approximation issues (i.e., having
improper global GAN optimizers caused by inappropriate GAN's loss functions),
and (3) generalizability issues (i.e., requiring large number of samples for
training). In this setup, we propose a GAN architecture which recovers the
maximum-likelihood solution and demonstrates fast generalization. Moreover, we
analyze global stability of different computational approaches for the proposed
GAN optimization and highlight their pros and cons. Finally, we outline an
extension of our model-based approach to design GANs in more complex setups
than the considered Gaussian benchmark.
|
Rough extreme learning machine: a new classification method based on
uncertainty measure | Extreme learning machine (ELM) is a new single hidden layer feedback neural
network. The weights of the input layer and the biases of neurons in hidden
layer are randomly generated, the weights of the output layer can be
analytically determined. ELM has been achieved good results for a large number
of classification tasks. In this paper, a new extreme learning machine called
rough extreme learning machine (RELM) was proposed. RELM uses rough set to
divide data into upper approximation set and lower approximation set, and the
two approximation sets are utilized to train upper approximation neurons and
lower approximation neurons. In addition, an attribute reduction is executed in
this algorithm to remove redundant attributes. The experimental results showed,
comparing with the comparison algorithms, RELM can get a better accuracy and
repeatability in most cases, RELM can not only maintain the advantages of fast
speed, but also effectively cope with the classification task for
high-dimensional data.
|
Unifying Value Iteration, Advantage Learning, and Dynamic Policy
Programming | Approximate dynamic programming algorithms, such as approximate value
iteration, have been successfully applied to many complex reinforcement
learning tasks, and a better approximate dynamic programming algorithm is
expected to further extend the applicability of reinforcement learning to
various tasks. In this paper we propose a new, robust dynamic programming
algorithm that unifies value iteration, advantage learning, and dynamic policy
programming. We call it generalized value iteration (GVI) and its approximated
version, approximate GVI (AGVI). We show AGVI's performance guarantee, which
includes performance guarantees for existing algorithms, as special cases. We
discuss theoretical weaknesses of existing algorithms, and explain the
advantages of AGVI. Numerical experiments in a simple environment support
theoretical arguments, and suggest that AGVI is a promising alternative to
previous algorithms.
|
Fast Linear Model for Knowledge Graph Embeddings | This paper shows that a simple baseline based on a Bag-of-Words (BoW)
representation learns surprisingly good knowledge graph embeddings. By casting
knowledge base completion and question answering as supervised classification
problems, we observe that modeling co-occurences of entities and relations
leads to state-of-the-art performance with a training time of a few minutes
using the open sourced library fastText.
|
Graph Attention Networks | We present graph attention networks (GATs), novel neural network
architectures that operate on graph-structured data, leveraging masked
self-attentional layers to address the shortcomings of prior methods based on
graph convolutions or their approximations. By stacking layers in which nodes
are able to attend over their neighborhoods' features, we enable (implicitly)
specifying different weights to different nodes in a neighborhood, without
requiring any kind of costly matrix operation (such as inversion) or depending
on knowing the graph structure upfront. In this way, we address several key
challenges of spectral-based graph neural networks simultaneously, and make our
model readily applicable to inductive as well as transductive problems. Our GAT
models have achieved or matched state-of-the-art results across four
established transductive and inductive graph benchmarks: the Cora, Citeseer and
Pubmed citation network datasets, as well as a protein-protein interaction
dataset (wherein test graphs remain unseen during training).
|
Optimization Landscape and Expressivity of Deep CNNs | We analyze the loss landscape and expressiveness of practical deep
convolutional neural networks (CNNs) with shared weights and max pooling
layers. We show that such CNNs produce linearly independent features at a
"wide" layer which has more neurons than the number of training samples. This
condition holds e.g. for the VGG network. Furthermore, we provide for such wide
CNNs necessary and sufficient conditions for global minima with zero training
error. For the case where the wide layer is followed by a fully connected layer
we show that almost every critical point of the empirical loss is a global
minimum with zero training error. Our analysis suggests that both depth and
width are very important in deep learning. While depth brings more
representational power and allows the network to learn high level features,
width smoothes the optimization landscape of the loss function in the sense
that a sufficiently wide network has a well-behaved loss surface with almost no
bad local minima.
|
A Supervised STDP-based Training Algorithm for Living Neural Networks | Neural networks have shown great potential in many applications like speech
recognition, drug discovery, image classification, and object detection. Neural
network models are inspired by biological neural networks, but they are
optimized to perform machine learning tasks on digital computers. The proposed
work explores the possibilities of using living neural networks in vitro as
basic computational elements for machine learning applications. A new
supervised STDP-based learning algorithm is proposed in this work, which
considers neuron engineering constrains. A 74.7% accuracy is achieved on the
MNIST benchmark for handwritten digit recognition.
|
Artificial Intelligence as Structural Estimation: Economic
Interpretations of Deep Blue, Bonanza, and AlphaGo | Artificial intelligence (AI) has achieved superhuman performance in a growing
number of tasks, but understanding and explaining AI remain challenging. This
paper clarifies the connections between machine-learning algorithms to develop
AIs and the econometrics of dynamic structural models through the case studies
of three famous game AIs. Chess-playing Deep Blue is a calibrated value
function, whereas shogi-playing Bonanza is an estimated value function via
Rust's (1987) nested fixed-point method. AlphaGo's "supervised-learning policy
network" is a deep neural network implementation of Hotz and Miller's (1993)
conditional choice probability estimation; its "reinforcement-learning value
network" is equivalent to Hotz, Miller, Sanders, and Smith's (1994) conditional
choice simulation method. Relaxing these AIs' implicit econometric assumptions
would improve their structural interpretability.
|
Denoising random forests | This paper proposes a novel type of random forests called a denoising random
forests that are robust against noises contained in test samples. Such
noise-corrupted samples cause serious damage to the estimation performances of
random forests, since unexpected child nodes are often selected and the leaf
nodes that the input sample reaches are sometimes far from those for a clean
sample. Our main idea for tackling this problem originates from a binary
indicator vector that encodes a traversal path of a sample in the forest. Our
proposed method effectively employs this vector by introducing denoising
autoencoders into random forests. A denoising autoencoder can be trained with
indicator vectors produced from clean and noisy input samples, and non-leaf
nodes where incorrect decisions are made can be identified by comparing the
input and output of the trained denoising autoencoder. Multiple traversal paths
with respect to the nodes with incorrect decisions caused by the noises can
then be considered for the estimation.
|
Stochastic gradient descent performs variational inference, converges to
limit cycles for deep networks | Stochastic gradient descent (SGD) is widely believed to perform implicit
regularization when used to train deep neural networks, but the precise manner
in which this occurs has thus far been elusive. We prove that SGD minimizes an
average potential over the posterior distribution of weights along with an
entropic regularization term. This potential is however not the original loss
function in general. So SGD does perform variational inference, but for a
different loss than the one used to compute the gradients. Even more
surprisingly, SGD does not even converge in the classical sense: we show that
the most likely trajectories of SGD for deep networks do not behave like
Brownian motion around critical points. Instead, they resemble closed loops
with deterministic components. We prove that such "out-of-equilibrium" behavior
is a consequence of highly non-isotropic gradient noise in SGD; the covariance
matrix of mini-batch gradients for deep networks has a rank as small as 1% of
its dimension. We provide extensive empirical validation of these claims,
proven in the appendix.
|
Unsupervised Neural Machine Translation | In spite of the recent success of neural machine translation (NMT) in
standard benchmarks, the lack of large parallel corpora poses a major practical
problem for many language pairs. There have been several proposals to alleviate
this issue with, for instance, triangulation and semi-supervised learning
techniques, but they still require a strong cross-lingual signal. In this work,
we completely remove the need of parallel data and propose a novel method to
train an NMT system in a completely unsupervised manner, relying on nothing but
monolingual corpora. Our model builds upon the recent work on unsupervised
embedding mappings, and consists of a slightly modified attentional
encoder-decoder model that can be trained on monolingual corpora alone using a
combination of denoising and backtranslation. Despite the simplicity of the
approach, our system obtains 15.56 and 10.21 BLEU points in WMT 2014
French-to-English and German-to-English translation. The model can also profit
from small parallel corpora, and attains 21.81 and 15.24 points when combined
with 100,000 parallel sentences, respectively. Our implementation is released
as an open source project.
|
A Connection between Feed-Forward Neural Networks and Probabilistic
Graphical Models | Two of the most popular modelling paradigms in computer vision are
feed-forward neural networks (FFNs) and probabilistic graphical models (GMs).
Various connections between the two have been studied in recent works, such as
e.g. expressing mean-field based inference in a GM as an FFN. This paper
establishes a new connection between FFNs and GMs. Our key observation is that
any FFN implements a certain approximation of a corresponding Bayesian network
(BN). We characterize various benefits of having this connection. In
particular, it results in a new learning algorithm for BNs. We validate the
proposed methods for a classification problem on CIFAR-10 dataset and for
binary image segmentation on Weizmann Horse dataset. We show that statistically
learned BNs improve performance, having at the same time essentially better
generalization capability, than their FFN counterparts.
|
Convergence Rates of Latent Topic Models Under Relaxed Identifiability
Conditions | In this paper we study the frequentist convergence rate for the Latent
Dirichlet Allocation (Blei et al., 2003) topic models. We show that the maximum
likelihood estimator converges to one of the finitely many equivalent
parameters in Wasserstein's distance metric at a rate of $n^{-1/4}$ without
assuming separability or non-degeneracy of the underlying topics and/or the
existence of more than three words per document, thus generalizing the previous
works of Anandkumar et al. (2012, 2014) from an information-theoretical
perspective. We also show that the $n^{-1/4}$ convergence rate is optimal in
the worst case.
|
Eigenoption Discovery through the Deep Successor Representation | Options in reinforcement learning allow agents to hierarchically decompose a
task into subtasks, having the potential to speed up learning and planning.
However, autonomously learning effective sets of options is still a major
challenge in the field. In this paper we focus on the recently introduced idea
of using representation learning methods to guide the option discovery process.
Specifically, we look at eigenoptions, options obtained from representations
that encode diffusive information flow in the environment. We extend the
existing algorithms for eigenoption discovery to settings with stochastic
transitions and in which handcrafted features are not available. We propose an
algorithm that discovers eigenoptions while learning non-linear state
representations from raw pixels. It exploits recent successes in the deep
reinforcement learning literature and the equivalence between proto-value
functions and the successor representation. We use traditional tabular domains
to provide intuition about our approach and Atari 2600 games to demonstrate its
potential.
|
Customized Routing Optimization Based on Gradient Boost Regressor Model | In this paper, we discussed limitation of current
electronic-design-automoation (EDA) tool and proposed a machine learning
framework to overcome the limitations and achieve better design quality. We
explored how to efficiently extract relevant features and leverage gradient
boost regressor (GBR) model to predict underestimated risky net (URN).
Customized routing optimizations are applied to the URNs and results show clear
timing improvement and trend to converge toward timing closure.
|
Predicting Floor-Level for 911 Calls with Neural Networks and Smartphone
Sensor Data | In cities with tall buildings, emergency responders need an accurate floor
level location to find 911 callers quickly. We introduce a system to estimate a
victim's floor level via their mobile device's sensor data in a two-step
process. First, we train a neural network to determine when a smartphone enters
or exits a building via GPS signal changes. Second, we use a barometer equipped
smartphone to measure the change in barometric pressure from the entrance of
the building to the victim's indoor location. Unlike impractical previous
approaches, our system is the first that does not require the use of beacons,
prior knowledge of the building infrastructure, or knowledge of user behavior.
We demonstrate real-world feasibility through 63 experiments across five
different tall buildings throughout New York City where our system predicted
the correct floor level with 100% accuracy.
|
Onsets and Frames: Dual-Objective Piano Transcription | We advance the state of the art in polyphonic piano music transcription by
using a deep convolutional and recurrent neural network which is trained to
jointly predict onsets and frames. Our model predicts pitch onset events and
then uses those predictions to condition framewise pitch predictions. During
inference, we restrict the predictions from the framewise detector by not
allowing a new note to start unless the onset detector also agrees that an
onset for that pitch is present in the frame. We focus on improving onsets and
offsets together instead of either in isolation as we believe this correlates
better with human musical perception. Our approach results in over a 100%
relative improvement in note F1 score (with offsets) on the MAPS dataset.
Furthermore, we extend the model to predict relative velocities of normalized
audio which results in more natural-sounding transcriptions.
|
CrescendoNet: A Simple Deep Convolutional Neural Network with Ensemble
Behavior | We introduce a new deep convolutional neural network, CrescendoNet, by
stacking simple building blocks without residual connections. Each Crescendo
block contains independent convolution paths with increased depths. The numbers
of convolution layers and parameters are only increased linearly in Crescendo
blocks. In experiments, CrescendoNet with only 15 layers outperforms almost all
networks without residual connections on benchmark datasets, CIFAR10, CIFAR100,
and SVHN. Given sufficient amount of data as in SVHN dataset, CrescendoNet with
15 layers and 4.1M parameters can match the performance of DenseNet-BC with 250
layers and 15.3M parameters. CrescendoNet provides a new way to construct high
performance deep convolutional neural networks without residual connections.
Moreover, through investigating the behavior and performance of subnetworks in
CrescendoNet, we note that the high performance of CrescendoNet may come from
its implicit ensemble behavior, which differs from the FractalNet that is also
a deep convolutional neural network without residual connections. Furthermore,
the independence between paths in CrescendoNet allows us to introduce a new
path-wise training procedure, which can reduce the memory needed for training.
|
Action-depedent Control Variates for Policy Optimization via Stein's
Identity | Policy gradient methods have achieved remarkable successes in solving
challenging reinforcement learning problems. However, it still often suffers
from the large variance issue on policy gradient estimation, which leads to
poor sample efficiency during training. In this work, we propose a control
variate method to effectively reduce variance for policy gradient methods.
Motivated by the Stein's identity, our method extends the previous control
variate methods used in REINFORCE and advantage actor-critic by introducing
more general action-dependent baseline functions. Empirical studies show that
our method significantly improves the sample efficiency of the state-of-the-art
policy gradient approaches.
|
Critical Points of Neural Networks: Analytical Forms and Landscape
Properties | Due to the success of deep learning to solving a variety of challenging
machine learning tasks, there is a rising interest in understanding loss
functions for training neural networks from a theoretical aspect. Particularly,
the properties of critical points and the landscape around them are of
importance to determine the convergence performance of optimization algorithms.
In this paper, we provide full (necessary and sufficient) characterization of
the analytical forms for the critical points (as well as global minimizers) of
the square loss functions for various neural networks. We show that the
analytical forms of the critical points characterize the values of the
corresponding loss functions as well as the necessary and sufficient conditions
to achieve global minimum. Furthermore, we exploit the analytical forms of the
critical points to characterize the landscape properties for the loss functions
of these neural networks. One particular conclusion is that: The loss function
of linear networks has no spurious local minimum, while the loss function of
one-hidden-layer nonlinear networks with ReLU activation function does have
local minimum that is not global minimum.
|
How Algorithmic Confounding in Recommendation Systems Increases
Homogeneity and Decreases Utility | Recommendation systems are ubiquitous and impact many domains; they have the
potential to influence product consumption, individuals' perceptions of the
world, and life-altering decisions. These systems are often evaluated or
trained with data from users already exposed to algorithmic recommendations;
this creates a pernicious feedback loop. Using simulations, we demonstrate how
using data confounded in this way homogenizes user behavior without increasing
utility.
|
Fast and Scalable Learning of Sparse Changes in High-Dimensional
Gaussian Graphical Model Structure | We focus on the problem of estimating the change in the dependency structures
of two $p$-dimensional Gaussian Graphical models (GGMs). Previous studies for
sparse change estimation in GGMs involve expensive and difficult non-smooth
optimization. We propose a novel method, DIFFEE for estimating DIFFerential
networks via an Elementary Estimator under a high-dimensional situation. DIFFEE
is solved through a faster and closed form solution that enables it to work in
large-scale settings. We conduct a rigorous statistical analysis showing that
surprisingly DIFFEE achieves the same asymptotic convergence rates as the
state-of-the-art estimators that are much more difficult to compute. Our
experimental results on multiple synthetic datasets and one real-world data
about brain connectivity show strong performance improvements over baselines,
as well as significant computational benefits.
|
Prototype Matching Networks for Large-Scale Multi-label Genomic Sequence
Classification | One of the fundamental tasks in understanding genomics is the problem of
predicting Transcription Factor Binding Sites (TFBSs). With more than hundreds
of Transcription Factors (TFs) as labels, genomic-sequence based TFBS
prediction is a challenging multi-label classification task. There are two
major biological mechanisms for TF binding: (1) sequence-specific binding
patterns on genomes known as "motifs" and (2) interactions among TFs known as
co-binding effects. In this paper, we propose a novel deep architecture, the
Prototype Matching Network (PMN) to mimic the TF binding mechanisms. Our PMN
model automatically extracts prototypes ("motif"-like features) for each TF
through a novel prototype-matching loss. Borrowing ideas from few-shot matching
models, we use the notion of support set of prototypes and an LSTM to learn how
TFs interact and bind to genomic sequences. On a reference TFBS dataset with
$2.1$ $million$ genomic sequences, PMN significantly outperforms baselines and
validates our design choices empirically. To our knowledge, this is the first
deep learning architecture that introduces prototype learning and considers
TF-TF interactions for large-scale TFBS prediction. Not only is the proposed
architecture accurate, but it also models the underlying biology.
|
Time-lagged autoencoders: Deep learning of slow collective variables for
molecular kinetics | Inspired by the success of deep learning techniques in the physical and
chemical sciences, we apply a modification of an autoencoder type deep neural
network to the task of dimension reduction of molecular dynamics data. We can
show that our time-lagged autoencoder reliably finds low-dimensional embeddings
for high-dimensional feature spaces which capture the slow dynamics of the
underlying stochastic processes - beyond the capabilities of linear dimension
reduction techniques.
|
Theoretical properties of the global optimizer of two layer neural
network | In this paper, we study the problem of optimizing a two-layer artificial
neural network that best fits a training dataset. We look at this problem in
the setting where the number of parameters is greater than the number of
sampled points. We show that for a wide class of differentiable activation
functions (this class involves "almost" all functions which are not piecewise
linear), we have that first-order optimal solutions satisfy global optimality
provided the hidden layer is non-singular. Our results are easily extended to
hidden layers given by a flat matrix from that of a square matrix. Results are
applicable even if network has more than one hidden layer provided all hidden
layers satisfy non-singularity, all activations are from the given "good" class
of differentiable functions and optimization is only with respect to the last
hidden layer. We also study the smoothness properties of the objective function
and show that it is actually Lipschitz smooth, i.e., its gradients do not
change sharply. We use smoothness properties to guarantee asymptotic
convergence of O(1/number of iterations) to a first-order optimal solution. We
also show that our algorithm will maintain non-singularity of hidden layer for
any finite number of iterations.
|
Learning Robust Rewards with Adversarial Inverse Reinforcement Learning | Reinforcement learning provides a powerful and general framework for decision
making and control, but its application in practice is often hindered by the
need for extensive feature and reward engineering. Deep reinforcement learning
methods can remove the need for explicit engineering of policy or value
features, but still require a manually specified reward function. Inverse
reinforcement learning holds the promise of automatic reward acquisition, but
has proven exceptionally difficult to apply to large, high-dimensional problems
with unknown dynamics. In this work, we propose adverserial inverse
reinforcement learning (AIRL), a practical and scalable inverse reinforcement
learning algorithm based on an adversarial reward learning formulation. We
demonstrate that AIRL is able to recover reward functions that are robust to
changes in dynamics, enabling us to learn policies even under significant
variation in the environment seen during training. Our experiments show that
AIRL greatly outperforms prior methods in these transfer settings.
|
Approximation Algorithms for $\ell_0$-Low Rank Approximation | We study the $\ell_0$-Low Rank Approximation Problem, where the goal is,
given an $m \times n$ matrix $A$, to output a rank-$k$ matrix $A'$ for which
$\|A'-A\|_0$ is minimized. Here, for a matrix $B$, $\|B\|_0$ denotes the number
of its non-zero entries. This NP-hard variant of low rank approximation is
natural for problems with no underlying metric, and its goal is to minimize the
number of disagreeing data positions. We provide approximation algorithms which
significantly improve the running time and approximation factor of previous
work. For $k > 1$, we show how to find, in poly$(mn)$ time for every $k$, a
rank $O(k \log(n/k))$ matrix $A'$ for which $\|A'-A\|_0 \leq O(k^2 \log(n/k))
\mathrm{OPT}$. To the best of our knowledge, this is the first algorithm with
provable guarantees for the $\ell_0$-Low Rank Approximation Problem for $k >
1$, even for bicriteria algorithms. For the well-studied case when $k = 1$, we
give a $(2+\epsilon)$-approximation in {\it sublinear time}, which is
impossible for other variants of low rank approximation such as for the
Frobenius norm. We strengthen this for the well-studied case of binary matrices
to obtain a $(1+O(\psi))$-approximation in sublinear time, where $\psi =
\mathrm{OPT}/\lVert A\rVert_0$. For small $\psi$, our approximation factor is
$1+o(1)$.
|
Empirical analysis of non-linear activation functions for Deep Neural
Networks in classification tasks | We provide an overview of several non-linear activation functions in a neural
network architecture that have proven successful in many machine learning
applications. We conduct an empirical analysis on the effectiveness of using
these function on the MNIST classification task, with the aim of clarifying
which functions produce the best results overall. Based on this first set of
results, we examine the effects of building deeper architectures with an
increasing number of hidden layers. We also survey the impact of using, on the
same task, different initialisation schemes for the weights of our neural
network. Using these sets of experiments as a base, we conclude by providing a
optimal neural network architecture that yields impressive results in accuracy
on the MNIST classification task.
|
Adversarial Advantage Actor-Critic Model for Task-Completion Dialogue
Policy Learning | This paper presents a new method --- adversarial advantage actor-critic
(Adversarial A2C), which significantly improves the efficiency of dialogue
policy learning in task-completion dialogue systems. Inspired by generative
adversarial networks (GAN), we train a discriminator to differentiate
responses/actions generated by dialogue agents from responses/actions by
experts. Then, we incorporate the discriminator as another critic into the
advantage actor-critic (A2C) framework, to encourage the dialogue agent to
explore state-action within the regions where the agent takes actions similar
to those of the experts. Experimental results in a movie-ticket booking domain
show that the proposed Adversarial A2C can accelerate policy exploration
efficiently.
|
Approximating Continuous Functions by ReLU Nets of Minimal Width | This article concerns the expressive power of depth in deep feed-forward
neural nets with ReLU activations. Specifically, we answer the following
question: for a fixed $d_{in}\geq 1,$ what is the minimal width $w$ so that
neural nets with ReLU activations, input dimension $d_{in}$, hidden layer
widths at most $w,$ and arbitrary depth can approximate any continuous,
real-valued function of $d_{in}$ variables arbitrarily well? It turns out that
this minimal width is exactly equal to $d_{in}+1.$ That is, if all the hidden
layer widths are bounded by $d_{in}$, then even in the infinite depth limit,
ReLU nets can only express a very limited class of functions, and, on the other
hand, any continuous function on the $d_{in}$-dimensional unit cube can be
approximated to arbitrary precision by ReLU nets in which all hidden layers
have width exactly $d_{in}+1.$ Our construction in fact shows that any
continuous function $f:[0,1]^{d_{in}}\to\mathbb R^{d_{out}}$ can be
approximated by a net of width $d_{in}+d_{out}$. We obtain quantitative depth
estimates for such an approximation in terms of the modulus of continuity of
$f$.
|
Algorithmic learning of probability distributions from random data in
the limit | We study the problem of identifying a probability distribution for some given
randomly sampled data in the limit, in the context of algorithmic learning
theory as proposed recently by Vinanyi and Chater. We show that there exists a
computable partial learner for the computable probability measures, while by
Bienvenu, Monin and Shen it is known that there is no computable learner for
the computable probability measures. Our main result is the characterization of
the oracles that compute explanatory learners for the computable (continuous)
probability measures as the high oracles. This provides an analogue of a
well-known result of Adleman and Blum in the context of learning computable
probability distributions. We also discuss related learning notions such as
behaviorally correct learning and orther variations of explanatory learning, in
the context of learning probability distributions from data.
|
The Exact Solution to Rank-1 L1-norm TUCKER2 Decomposition | We study rank-1 {L1-norm-based TUCKER2} (L1-TUCKER2) decomposition of 3-way
tensors, treated as a collection of $N$ $D \times M$ matrices that are to be
jointly decomposed. Our contributions are as follows. i) We prove that the
problem is equivalent to combinatorial optimization over $N$ antipodal-binary
variables. ii) We derive the first two algorithms in the literature for its
exact solution. The first algorithm has cost exponential in $N$; the second one
has cost polynomial in $N$ (under a mild assumption). Our algorithms are
accompanied by formal complexity analysis. iii) We conduct numerical studies to
compare the performance of exact L1-TUCKER2 (proposed) with standard HOSVD,
HOOI, GLRAM, PCA, L1-PCA, and TPCA-L1. Our studies show that L1-TUCKER2
outperforms (in tensor approximation) all the above counterparts when the
processed data are outlier corrupted.
|
Deep Forward and Inverse Perceptual Models for Tracking and Prediction | We consider the problems of learning forward models that map state to
high-dimensional images and inverse models that map high-dimensional images to
state in robotics. Specifically, we present a perceptual model for generating
video frames from state with deep networks, and provide a framework for its use
in tracking and prediction tasks. We show that our proposed model greatly
outperforms standard deconvolutional methods and GANs for image generation,
producing clear, photo-realistic images. We also develop a convolutional neural
network model for state estimation and compare the result to an Extended Kalman
Filter to estimate robot trajectories. We validate all models on a real robotic
system.
|
Resolution and Relevance Trade-offs in Deep Learning | Deep learning has been successfully applied to various tasks, but its
underlying mechanism remains unclear. Neural networks associate similar inputs
in the visible layer to the same state of hidden variables in deep layers. The
fraction of inputs that are associated to the same state is a natural measure
of similarity and is simply related to the cost in bits required to represent
these inputs. The degeneracy of states with the same information cost provides
instead a natural measure of noise and is simply related the entropy of the
frequency of states, that we call relevance. Representations with minimal
noise, at a given level of similarity (resolution), are those that maximise the
relevance. A signature of such efficient representations is that frequency
distributions follow power laws. We show, in extensive numerical experiments,
that deep neural networks extract a hierarchy of efficient representations from
data, because they i) achieve low levels of noise (i.e. high relevance) and ii)
exhibit power law distributions. We also find that the layer that is most
efficient to reliably generate patterns of training data is the one for which
relevance and resolution are traded at the same price, which implies that
frequency distribution follows Zipf's law.
|
Generating Natural Adversarial Examples | Due to their complex nature, it is hard to characterize the ways in which
machine learning models can misbehave or be exploited when deployed. Recent
work on adversarial examples, i.e. inputs with minor perturbations that result
in substantially different model predictions, is helpful in evaluating the
robustness of these models by exposing the adversarial scenarios where they
fail. However, these malicious perturbations are often unnatural, not
semantically meaningful, and not applicable to complicated domains such as
language. In this paper, we propose a framework to generate natural and legible
adversarial examples that lie on the data manifold, by searching in semantic
space of dense and continuous data representation, utilizing the recent
advances in generative adversarial networks. We present generated adversaries
to demonstrate the potential of the proposed approach for black-box classifiers
for a wide range of applications such as image classification, textual
entailment, and machine translation. We include experiments to show that the
generated adversaries are natural, legible to humans, and useful in evaluating
and analyzing black-box classifiers.
|
Tensor Regression Meets Gaussian Processes | Low-rank tensor regression, a new model class that learns high-order
correlation from data, has recently received considerable attention. At the
same time, Gaussian processes (GP) are well-studied machine learning models for
structure learning. In this paper, we demonstrate interesting connections
between the two, especially for multi-way data analysis. We show that low-rank
tensor regression is essentially learning a multi-linear kernel in Gaussian
processes, and the low-rank assumption translates to the constrained Bayesian
inference problem. We prove the oracle inequality and derive the average case
learning curve for the equivalent GP model. Our finding implies that low-rank
tensor regression, though empirically successful, is highly dependent on the
eigenvalues of covariance functions as well as variable correlations.
|
ChainerMN: Scalable Distributed Deep Learning Framework | One of the keys for deep learning to have made a breakthrough in various
fields was to utilize high computing powers centering around GPUs. Enabling the
use of further computing abilities by distributed processing is essential not
only to make the deep learning bigger and faster but also to tackle unsolved
challenges. We present the design, implementation, and evaluation of ChainerMN,
the distributed deep learning framework we have developed. We demonstrate that
ChainerMN can scale the learning process of the ResNet-50 model to the ImageNet
dataset up to 128 GPUs with the parallel efficiency of 90%.
|
Semantic Interpolation in Implicit Models | In implicit models, one often interpolates between sampled points in latent
space. As we show in this paper, care needs to be taken to match-up the
distributional assumptions on code vectors with the geometry of the
interpolating paths. Otherwise, typical assumptions about the quality and
semantics of in-between points may not be justified. Based on our analysis we
propose to modify the prior code distribution to put significantly more
probability mass closer to the origin. As a result, linear interpolation paths
are not only shortest paths, but they are also guaranteed to pass through
high-density regions, irrespective of the dimensionality of the latent space.
Experiments on standard benchmark image datasets demonstrate clear visual
improvements in the quality of the generated samples and exhibit more
meaningful interpolation paths.
|
Flexible Prior Distributions for Deep Generative Models | We consider the problem of training generative models with deep neural
networks as generators, i.e. to map latent codes to data points. Whereas the
dominant paradigm combines simple priors over codes with complex deterministic
models, we argue that it might be advantageous to use more flexible code
distributions. We demonstrate how these distributions can be induced directly
from the data. The benefits include: more powerful generative models, better
modeling of latent structure and explicit control of the degree of
generalization.
|
Parametrizing filters of a CNN with a GAN | It is commonly agreed that the use of relevant invariances as a good
statistical bias is important in machine-learning. However, most approaches
that explicitly incorporate invariances into a model architecture only make use
of very simple transformations, such as translations and rotations. Hence,
there is a need for methods to model and extract richer transformations that
capture much higher-level invariances. To that end, we introduce a tool
allowing to parametrize the set of filters of a trained convolutional neural
network with the latent space of a generative adversarial network. We then show
that the method can capture highly non-linear invariances of the data by
visualizing their effect in the data space.
|
TreeQN and ATreeC: Differentiable Tree-Structured Models for Deep
Reinforcement Learning | Combining deep model-free reinforcement learning with on-line planning is a
promising approach to building on the successes of deep RL. On-line planning
with look-ahead trees has proven successful in environments where transition
models are known a priori. However, in complex environments where transition
models need to be learned from data, the deficiencies of learned models have
limited their utility for planning. To address these challenges, we propose
TreeQN, a differentiable, recursive, tree-structured model that serves as a
drop-in replacement for any value function network in deep RL with discrete
actions. TreeQN dynamically constructs a tree by recursively applying a
transition model in a learned abstract state space and then aggregating
predicted rewards and state-values using a tree backup to estimate Q-values. We
also propose ATreeC, an actor-critic variant that augments TreeQN with a
softmax layer to form a stochastic policy network. Both approaches are trained
end-to-end, such that the learned model is optimised for its actual use in the
tree. We show that TreeQN and ATreeC outperform n-step DQN and A2C on a
box-pushing task, as well as n-step DQN and value prediction networks (Oh et
al. 2017) on multiple Atari games. Furthermore, we present ablation studies
that demonstrate the effect of different auxiliary losses on learning
transition models.
|
Regret Minimization for Partially Observable Deep Reinforcement Learning | Deep reinforcement learning algorithms that estimate state and state-action
value functions have been shown to be effective in a variety of challenging
domains, including learning control strategies from raw image pixels. However,
algorithms that estimate state and state-action value functions typically
assume a fully observed state and must compensate for partial observations by
using finite length observation histories or recurrent networks. In this work,
we propose a new deep reinforcement learning algorithm based on counterfactual
regret minimization that iteratively updates an approximation to an
advantage-like function and is robust to partially observed state. We
demonstrate that this new algorithm can substantially outperform strong
baseline methods on several partially observed reinforcement learning tasks:
learning first-person 3D navigation in Doom and Minecraft, and acting in the
presence of partially observed objects in Doom and Pong.
|
SVSGAN: Singing Voice Separation via Generative Adversarial Network | Separating two sources from an audio mixture is an important task with many
applications. It is a challenging problem since only one signal channel is
available for analysis. In this paper, we propose a novel framework for singing
voice separation using the generative adversarial network (GAN) with a
time-frequency masking function. The mixture spectra is considered to be a
distribution and is mapped to the clean spectra which is also considered a
distribtution. The approximation of distributions between mixture spectra and
clean spectra is performed during the adversarial training process. In contrast
with current deep learning approaches for source separation, the parameters of
the proposed framework are first initialized in a supervised setting and then
optimized by the training procedure of GAN in an unsupervised setting.
Experimental results on three datasets (MIR-1K, iKala and DSD100) show that
performance can be improved by the proposed framework consisting of
conventional networks.
|
Physics-guided Neural Networks (PGNN): An Application in Lake
Temperature Modeling | This paper introduces a framework for combining scientific knowledge of
physics-based models with neural networks to advance scientific discovery. This
framework, termed physics-guided neural networks (PGNN), leverages the output
of physics-based model simulations along with observational features in a
hybrid modeling setup to generate predictions using a neural network
architecture. Further, this framework uses physics-based loss functions in the
learning objective of neural networks to ensure that the model predictions not
only show lower errors on the training set but are also scientifically
consistent with the known physics on the unlabeled set. We illustrate the
effectiveness of PGNN for the problem of lake temperature modeling, where
physical relationships between the temperature, density, and depth of water are
used to design a physics-based loss function. By using scientific knowledge to
guide the construction and learning of neural networks, we are able to show
that the proposed framework ensures better generalizability as well as
scientific consistency of results. All the code and datasets used in this study
have been made available on this link \url{https://github.com/arkadaw9/PGNN}.
|
Learning Neural Representations of Human Cognition across Many fMRI
Studies | Cognitive neuroscience is enjoying rapid increase in extensive public
brain-imaging datasets. It opens the door to large-scale statistical models.
Finding a unified perspective for all available data calls for scalable and
automated solutions to an old challenge: how to aggregate heterogeneous
information on brain function into a universal cognitive system that relates
mental operations/cognitive processes/psychological tasks to brain networks? We
cast this challenge in a machine-learning approach to predict conditions from
statistical brain maps across different studies. For this, we leverage
multi-task learning and multi-scale dimension reduction to learn
low-dimensional representations of brain images that carry cognitive
information and can be robustly associated with psychological stimuli. Our
multi-dataset classification model achieves the best prediction performance on
several large reference datasets, compared to models without cognitive-aware
low-dimension representations, it brings a substantial performance boost to the
analysis of small datasets, and can be introspected to identify universal
template cognitive concepts.
|
Statistical Speech Enhancement Based on Probabilistic Integration of
Variational Autoencoder and Non-Negative Matrix Factorization | This paper presents a statistical method of single-channel speech enhancement
that uses a variational autoencoder (VAE) as a prior distribution on clean
speech. A standard approach to speech enhancement is to train a deep neural
network (DNN) to take noisy speech as input and output clean speech. Although
this supervised approach requires a very large amount of pair data for
training, it is not robust against unknown environments. Another approach is to
use non-negative matrix factorization (NMF) based on basis spectra trained on
clean speech in advance and those adapted to noise on the fly. This
semi-supervised approach, however, causes considerable signal distortion in
enhanced speech due to the unrealistic assumption that speech spectrograms are
linear combinations of the basis spectra. Replacing the poor linear generative
model of clean speech in NMF with a VAE---a powerful nonlinear deep generative
model---trained on clean speech, we formulate a unified probabilistic
generative model of noisy speech. Given noisy speech as observed data, we can
sample clean speech from its posterior distribution. The proposed method
outperformed the conventional DNN-based method in unseen noisy environments.
|
Conditional Variance Penalties and Domain Shift Robustness | When training a deep neural network for image classification, one can broadly
distinguish between two types of latent features of images that will drive the
classification. We can divide latent features into (i) "core" or "conditionally
invariant" features $X^\text{core}$ whose distribution $X^\text{core}\vert Y$,
conditional on the class $Y$, does not change substantially across domains and
(ii) "style" features $X^{\text{style}}$ whose distribution $X^{\text{style}}
\vert Y$ can change substantially across domains. Examples for style features
include position, rotation, image quality or brightness but also more complex
ones like hair color, image quality or posture for images of persons. Our goal
is to minimize a loss that is robust under changes in the distribution of these
style features. In contrast to previous work, we assume that the domain itself
is not observed and hence a latent variable.
We do assume that we can sometimes observe a typically discrete identifier or
"$\mathrm{ID}$ variable". In some applications we know, for example, that two
images show the same person, and $\mathrm{ID}$ then refers to the identity of
the person. The proposed method requires only a small fraction of images to
have $\mathrm{ID}$ information. We group observations if they share the same
class and identifier $(Y,\mathrm{ID})=(y,\mathrm{id})$ and penalize the
conditional variance of the prediction or the loss if we condition on
$(Y,\mathrm{ID})$. Using a causal framework, this conditional variance
regularization (CoRe) is shown to protect asymptotically against shifts in the
distribution of the style variables. Empirically, we show that the CoRe penalty
improves predictive accuracy substantially in settings where domain changes
occur in terms of image quality, brightness and color while we also look at
more complex changes such as changes in movement and posture.
|
Multi-Resolution Fully Convolutional Neural Networks for Monaural Audio
Source Separation | In deep neural networks with convolutional layers, each layer typically has
fixed-size/single-resolution receptive field (RF). Convolutional layers with a
large RF capture global information from the input features, while layers with
small RF size capture local details with high resolution from the input
features. In this work, we introduce novel deep multi-resolution fully
convolutional neural networks (MR-FCNN), where each layer has different RF
sizes to extract multi-resolution features that capture the global and local
details information from its input features. The proposed MR-FCNN is applied to
separate a target audio source from a mixture of many audio sources.
Experimental results show that using MR-FCNN improves the performance compared
to feedforward deep neural networks (DNNs) and single resolution deep fully
convolutional neural networks (FCNNs) on the audio source separation problem.
|
A Convergent Algorithm for Bi-orthogonal Nonnegative Matrix
Tri-Factorization | A convergent algorithm for nonnegative matrix factorization with
orthogonality constraints imposed on both factors is proposed in this paper.
This factorization concept was first introduced by Ding et al. with intent to
further improve clustering capability of NMF. However, as the original
algorithm was developed based on multiplicative update rules, the convergence
of the algorithm cannot be guaranteed. In this paper, we utilize the technique
presented in our previous work to develop the algorithm and prove that it
converges to a stationary point inside the solution space.
|
Compact Multi-Class Boosted Trees | Gradient boosted decision trees are a popular machine learning technique, in
part because of their ability to give good accuracy with small models. We
describe two extensions to the standard tree boosting algorithm designed to
increase this advantage. The first improvement extends the boosting formalism
from scalar-valued trees to vector-valued trees. This allows individual trees
to be used as multiclass classifiers, rather than requiring one tree per class,
and drastically reduces the model size required for multiclass problems. We
also show that some other popular vector-valued gradient boosted trees
modifications fit into this formulation and can be easily obtained in our
implementation. The second extension, layer-by-layer boosting, takes smaller
steps in function space, which is empirically shown to lead to a faster
convergence and to a more compact ensemble. We have added both improvements to
the open-source TensorFlow Boosted trees (TFBT) package, and we demonstrate
their efficacy on a variety of multiclass datasets. We expect these extensions
will be of particular interest to boosted tree applications that require small
models, such as embedded devices, applications requiring fast inference, or
applications desiring more interpretable models.
|
TF Boosted Trees: A scalable TensorFlow based framework for gradient
boosting | TF Boosted Trees (TFBT) is a new open-sourced frame-work for the distributed
training of gradient boosted trees. It is based on TensorFlow, and its
distinguishing features include a novel architecture, automatic loss
differentiation, layer-by-layer boosting that results in smaller ensembles and
faster prediction, principled multi-class handling, and a number of
regularization techniques to prevent overfitting.
|
Deep Learning as a Mixed Convex-Combinatorial Optimization Problem | As neural networks grow deeper and wider, learning networks with
hard-threshold activations is becoming increasingly important, both for network
quantization, which can drastically reduce time and energy requirements, and
for creating large integrated systems of deep networks, which may have
non-differentiable components and must avoid vanishing and exploding gradients
for effective learning. However, since gradient descent is not applicable to
hard-threshold functions, it is not clear how to learn networks of them in a
principled way. We address this problem by observing that setting targets for
hard-threshold hidden units in order to minimize loss is a discrete
optimization problem, and can be solved as such. The discrete optimization goal
is to find a set of targets such that each unit, including the output, has a
linearly separable problem to solve. Given these targets, the network
decomposes into individual perceptrons, which can then be learned with standard
convex approaches. Based on this, we develop a recursive mini-batch algorithm
for learning deep hard-threshold networks that includes the popular but poorly
justified straight-through estimator as a special case. Empirically, we show
that our algorithm improves classification accuracy in a number of settings,
including for AlexNet and ResNet-18 on ImageNet, when compared to the
straight-through estimator.
|
Learning Depthwise Separable Graph Convolution from Data Manifold | Convolution Neural Network (CNN) has gained tremendous success in computer
vision tasks with its outstanding ability to capture the local latent features.
Recently, there has been an increasing interest in extending convolution
operations to the non-Euclidean geometry. Although various types of convolution
operations have been proposed for graphs or manifolds, their connections with
traditional convolution over grid-structured data are not well-understood. In
this paper, we show that depthwise separable convolution can be successfully
generalized for the unification of both graph-based and grid-based convolution
methods. Based on this insight we propose a novel Depthwise Separable Graph
Convolution (DSGC) approach which is compatible with the tradition convolution
network and subsumes existing convolution methods as special cases. It is
equipped with the combined strengths in model expressiveness, compatibility
(relatively small number of parameters), modularity and computational
efficiency in training. Extensive experiments show the outstanding performance
of DSGC in comparison with strong baselines on multi-domain benchmark datasets.
|
On Learning Mixtures of Well-Separated Gaussians | We consider the problem of efficiently learning mixtures of a large number of
spherical Gaussians, when the components of the mixture are well separated. In
the most basic form of this problem, we are given samples from a uniform
mixture of $k$ standard spherical Gaussians, and the goal is to estimate the
means up to accuracy $\delta$ using $poly(k,d, 1/\delta)$ samples.
In this work, we study the following question: what is the minimum separation
needed between the means for solving this task? The best known algorithm due to
Vempala and Wang [JCSS 2004] requires a separation of roughly
$\min\{k,d\}^{1/4}$. On the other hand, Moitra and Valiant [FOCS 2010] showed
that with separation $o(1)$, exponentially many samples are required. We
address the significant gap between these two bounds, by showing the following
results.
1. We show that with separation $o(\sqrt{\log k})$, super-polynomially many
samples are required. In fact, this holds even when the $k$ means of the
Gaussians are picked at random in $d=O(\log k)$ dimensions.
2. We show that with separation $\Omega(\sqrt{\log k})$, $poly(k,d,1/\delta)$
samples suffice. Note that the bound on the separation is independent of
$\delta$. This result is based on a new and efficient "accuracy boosting"
algorithm that takes as input coarse estimates of the true means and in time
$poly(k,d, 1/\delta)$ outputs estimates of the means up to arbitrary accuracy
$\delta$ assuming the separation between the means is $\Omega(\min\{\sqrt{\log
k},\sqrt{d}\})$ (independently of $\delta$).
We also present a computationally efficient algorithm in $d=O(1)$ dimensions
with only $\Omega(\sqrt{d})$ separation. These results together essentially
characterize the optimal order of separation between components that is needed
to learn a mixture of $k$ spherical Gaussians with polynomial samples.
|
Small Moving Window Calibration Models for Soft Sensing Processes with
Limited History | Five simple soft sensor methodologies with two update conditions were
compared on two experimentally-obtained datasets and one simulated dataset. The
soft sensors investigated were moving window partial least squares regression
(and a recursive variant), moving window random forest regression, the mean
moving window of $y$, and a novel random forest partial least squares
regression ensemble (RF-PLS), all of which can be used with small sample sizes
so that they can be rapidly placed online. It was found that, on two of the
datasets studied, small window sizes led to the lowest prediction errors for
all of the moving window methods studied. On the majority of datasets studied,
the RF-PLS calibration method offered the lowest one-step-ahead prediction
errors compared to those of the other methods, and it demonstrated greater
predictive stability at larger time delays than moving window PLS alone. It was
found that both the random forest and RF-PLS methods most adequately modeled
the datasets that did not feature purely monotonic increases in property
values, but that both methods performed more poorly than moving window PLS
models on one dataset with purely monotonic property values. Other data
dependent findings are presented and discussed.
|
Meta-Learning and Universality: Deep Representations and Gradient
Descent can Approximate any Learning Algorithm | Learning to learn is a powerful paradigm for enabling models to learn from
data more effectively and efficiently. A popular approach to meta-learning is
to train a recurrent model to read in a training dataset as input and output
the parameters of a learned model, or output predictions for new test inputs.
Alternatively, a more recent approach to meta-learning aims to acquire deep
representations that can be effectively fine-tuned, via standard gradient
descent, to new tasks. In this paper, we consider the meta-learning problem
from the perspective of universality, formalizing the notion of learning
algorithm approximation and comparing the expressive power of the
aforementioned recurrent models to the more recent approaches that embed
gradient descent into the meta-learner. In particular, we seek to answer the
following question: does deep representation combined with standard gradient
descent have sufficient capacity to approximate any learning algorithm? We find
that this is indeed true, and further find, in our experiments, that
gradient-based meta-learning consistently leads to learning strategies that
generalize more widely compared to those represented by recurrent models.
|
Gene Ontology (GO) Prediction using Machine Learning Methods | We applied machine learning to predict whether a gene is involved in axon
regeneration. We extracted 31 features from different databases and trained
five machine learning models. Our optimal model, a Random Forest Classifier
with 50 submodels, yielded a test score of 85.71%, which is 4.1% higher than
the baseline score. We concluded that our models have some predictive
capability. Similar methodology and features could be applied to predict other
Gene Ontology (GO) terms.
|
Log-DenseNet: How to Sparsify a DenseNet | Skip connections are increasingly utilized by deep neural networks to improve
accuracy and cost-efficiency. In particular, the recent DenseNet is efficient
in computation and parameters, and achieves state-of-the-art predictions by
directly connecting each feature layer to all previous ones. However,
DenseNet's extreme connectivity pattern may hinder its scalability to high
depths, and in applications like fully convolutional networks, full DenseNet
connections are prohibitively expensive. This work first experimentally shows
that one key advantage of skip connections is to have short distances among
feature layers during backpropagation. Specifically, using a fixed number of
skip connections, the connection patterns with shorter backpropagation distance
among layers have more accurate predictions. Following this insight, we propose
a connection template, Log-DenseNet, which, in comparison to DenseNet, only
slightly increases the backpropagation distances among layers from 1 to ($1 +
\log_2 L$), but uses only $L\log_2 L$ total connections instead of $O(L^2)$.
Hence, Log-DenseNets are easier than DenseNets to implement and to scale. We
demonstrate the effectiveness of our design principle by showing better
performance than DenseNets on tabula rasa semantic segmentation, and
competitive results on visual recognition.
|
Accelerate RNN-based Training with Importance Sampling | Importance sampling (IS) as an elegant and efficient variance reduction (VR)
technique for the acceleration of stochastic optimization problems has
attracted many researches recently. Unlike commonly adopted stochastic uniform
sampling in stochastic optimizations, IS-integrated algorithms sample training
data at each iteration with respect to a weighted sampling probability
distribution $P$, which is constructed according to the precomputed importance
factors. Previous experimental results show that IS has achieved remarkable
progresses in the acceleration of training convergence. Unfortunately, the
calculation of the sampling probability distribution $P$ causes a major
limitation of IS: it requires the input data to be well-structured, i.e., the
feature vector is properly defined. Consequently, recurrent neural networks
(RNN) as a popular learning algorithm is not able to enjoy the benefits of IS
due to the fact that its raw input data, i.e., the training sequences, are
often unstructured which makes calculation of $P$ impossible. In considering of
the the popularity of RNN-based learning applications and their relative long
training time, we are interested in accelerating them through IS. This paper
propose a novel Fast-Importance-Mining algorithm to calculate the importance
factor for unstructured data which makes the application of IS in RNN-based
applications possible. Our experimental evaluation on popular open-source
RNN-based learning applications validate the effectiveness of IS in improving
the convergence rate of RNNs.
|
Adversarial Semi-Supervised Audio Source Separation applied to Singing
Voice Extraction | The state of the art in music source separation employs neural networks
trained in a supervised fashion on multi-track databases to estimate the
sources from a given mixture. With only few datasets available, often extensive
data augmentation is used to combat overfitting. Mixing random tracks, however,
can even reduce separation performance as instruments in real music are
strongly correlated. The key concept in our approach is that source estimates
of an optimal separator should be indistinguishable from real source signals.
Based on this idea, we drive the separator towards outputs deemed as realistic
by discriminator networks that are trained to tell apart real from separator
samples. This way, we can also use unpaired source and mixture recordings
without the drawbacks of creating unrealistic music mixtures. Our framework is
widely applicable as it does not assume a specific network architecture or
number of sources. To our knowledge, this is the first adoption of adversarial
training for music source separation. In a prototype experiment for singing
voice separation, separation performance increases with our approach compared
to purely supervised training.
|
Medical Image Segmentation Based on Multi-Modal Convolutional Neural
Network: Study on Image Fusion Schemes | Image analysis using more than one modality (i.e. multi-modal) has been
increasingly applied in the field of biomedical imaging. One of the challenges
in performing the multimodal analysis is that there exist multiple schemes for
fusing the information from different modalities, where such schemes are
application-dependent and lack a unified framework to guide their designs. In
this work we firstly propose a conceptual architecture for the image fusion
schemes in supervised biomedical image analysis: fusing at the feature level,
fusing at the classifier level, and fusing at the decision-making level.
Further, motivated by the recent success in applying deep learning for natural
image analysis, we implement the three image fusion schemes above based on the
Convolutional Neural Network (CNN) with varied structures, and combined into a
single framework. The proposed image segmentation framework is capable of
analyzing the multi-modality images using different fusing schemes
simultaneously. The framework is applied to detect the presence of soft tissue
sarcoma from the combination of Magnetic Resonance Imaging (MRI), Computed
Tomography (CT) and Positron Emission Tomography (PET) images. It is found from
the results that while all the fusion schemes outperform the single-modality
schemes, fusing at the feature level can generally achieve the best performance
in terms of both accuracy and computational cost, but also suffers from the
decreased robustness in the presence of large errors in any image modalities.
|
Calibration for Stratified Classification Models | In classification problems, sampling bias between training data and testing
data is critical to the ranking performance of classification scores. Such bias
can be both unintentionally introduced by data collection and intentionally
introduced by the algorithm, such as under-sampling or weighting techniques
applied to imbalanced data. When such sampling bias exists, using the raw
classification score to rank observations in the testing data can lead to
suboptimal results. In this paper, I investigate the optimal calibration
strategy in general settings, and develop a practical solution for one specific
sampling bias case, where the sampling bias is introduced by stratified
sampling. The optimal solution is developed by analytically solving the problem
of optimizing the ROC curve. For practical data, I propose a ranking algorithm
for general classification models with stratified data. Numerical experiments
demonstrate that the proposed algorithm effectively addresses the stratified
sampling bias issue. Interestingly, the proposed method shows its potential
applicability in two other machine learning areas: unsupervised learning and
model ensembling, which can be future research topics.
|
Fraternal Dropout | Recurrent neural networks (RNNs) are important class of architectures among
neural networks useful for language modeling and sequential prediction.
However, optimizing RNNs is known to be harder compared to feed-forward neural
networks. A number of techniques have been proposed in literature to address
this problem. In this paper we propose a simple technique called fraternal
dropout that takes advantage of dropout to achieve this goal. Specifically, we
propose to train two identical copies of an RNN (that share parameters) with
different dropout masks while minimizing the difference between their
(pre-softmax) predictions. In this way our regularization encourages the
representations of RNNs to be invariant to dropout mask, thus being robust. We
show that our regularization term is upper bounded by the expectation-linear
dropout objective which has been shown to address the gap due to the difference
between the train and inference phases of dropout. We evaluate our model and
achieve state-of-the-art results in sequence modeling tasks on two benchmark
datasets - Penn Treebank and Wikitext-2. We also show that our approach leads
to performance improvement by a significant margin in image captioning
(Microsoft COCO) and semi-supervised (CIFAR-10) tasks.
|
Long-term Forecasting using Higher Order Tensor RNNs | We present Higher-Order Tensor RNN (HOT-RNN), a novel family of neural
sequence architectures for multivariate forecasting in environments with
nonlinear dynamics. Long-term forecasting in such systems is highly
challenging, since there exist long-term temporal dependencies, higher-order
correlations and sensitivity to error propagation. Our proposed recurrent
architecture addresses these issues by learning the nonlinear dynamics directly
using higher-order moments and higher-order state transition functions.
Furthermore, we decompose the higher-order structure using the tensor-train
decomposition to reduce the number of parameters while preserving the model
performance. We theoretically establish the approximation guarantees and the
variance bound for HOT-RNN for general sequence inputs. We also demonstrate 5%
~ 12% improvements for long-term prediction over general RNN and LSTM
architectures on a range of simulated environments with nonlinear dynamics, as
well on real-world time series data.
|
Separation of Water and Fat Magnetic Resonance Imaging Signals Using
Deep Learning with Convolutional Neural Networks | Purpose: A new method for magnetic resonance (MR) imaging water-fat
separation using a convolutional neural network (ConvNet) and deep learning
(DL) is presented. Feasibility of the method with complex and magnitude images
is demonstrated with a series of patient studies and accuracy of predicted
quantitative values is analyzed.
Methods: Water-fat separation of 1200 gradient-echo acquisitions from 90
imaging sessions (normal, acute and chronic myocardial infarction) was
performed using a conventional model based method with modeling of R2* and
off-resonance and a multi-peak fat spectrum. A U-Net convolutional neural
network for calculation of water-only, fat-only, R2* and off-resonance images
was trained with 900 gradient-echo Multiple and single-echo complex and
magnitude input data algorithms were studied and compared to conventional
extended echo modeling.
Results: The U-Net ConvNet was easily trained and provided water-fat
separation results visually comparable to conventional methods. Myocardial fat
deposition in chronic myocardial infarction and intramyocardial hemorrhage in
acute myocardial infarction were well visualized in the DL results. Predicted
values for R2*, off-resonance, water and fat signal intensities were well
correlated with conventional model based water fat separation (R2>=0.97,
p<0.001). DL images had a 14% higher signal-to-noise ratio (p<0.001) when
compared to the conventional method.
Conclusion: Deep learning utilizing ConvNets is a feasible method for MR
water-fat separationimaging with complex, magnitude and single echo image data.
A trained U-Net can be efficiently used for MR water-fat separation, providing
results comparable to conventional model based methods.
|
Beyond Shared Hierarchies: Deep Multitask Learning through Soft Layer
Ordering | Existing deep multitask learning (MTL) approaches align layers shared between
tasks in a parallel ordering. Such an organization significantly constricts the
types of shared structure that can be learned. The necessity of parallel
ordering for deep MTL is first tested by comparing it with permuted ordering of
shared layers. The results indicate that a flexible ordering can enable more
effective sharing, thus motivating the development of a soft ordering approach,
which learns how shared layers are applied in different ways for different
tasks. Deep MTL with soft ordering outperforms parallel ordering methods across
a series of domains. These results suggest that the power of deep MTL comes
from learning highly general building blocks that can be assembled to meet the
demands of each task.
|
Backpropagation through the Void: Optimizing control variates for
black-box gradient estimation | Gradient-based optimization is the foundation of deep learning and
reinforcement learning. Even when the mechanism being optimized is unknown or
not differentiable, optimization using high-variance or biased gradient
estimates is still often the best strategy. We introduce a general framework
for learning low-variance, unbiased gradient estimators for black-box functions
of random variables. Our method uses gradients of a neural network trained
jointly with model parameters or policies, and is applicable in both discrete
and continuous settings. We demonstrate this framework for training discrete
latent-variable models. We also give an unbiased, action-conditional extension
of the advantage actor-critic reinforcement learning algorithm.
|
Accelerated Sparse Subspace Clustering | State-of-the-art algorithms for sparse subspace clustering perform spectral
clustering on a similarity matrix typically obtained by representing each data
point as a sparse combination of other points using either basis pursuit (BP)
or orthogonal matching pursuit (OMP). BP-based methods are often prohibitive in
practice while the performance of OMP-based schemes are unsatisfactory,
especially in settings where data points are highly similar. In this paper, we
propose a novel algorithm that exploits an accelerated variant of orthogonal
least-squares to efficiently find the underlying subspaces. We show that under
certain conditions the proposed algorithm returns a subspace-preserving
solution. Simulation results illustrate that the proposed method compares
favorably with BP-based method in terms of running time while being
significantly more accurate than OMP-based schemes.
|
Pomegranate: fast and flexible probabilistic modeling in python | We present pomegranate, an open source machine learning package for
probabilistic modeling in Python. Probabilistic modeling encompasses a wide
range of methods that explicitly describe uncertainty using probability
distributions. Three widely used probabilistic models implemented in
pomegranate are general mixture models, hidden Markov models, and Bayesian
networks. A primary focus of pomegranate is to abstract away the complexities
of training models from their definition. This allows users to focus on
specifying the correct model for their application instead of being limited by
their understanding of the underlying algorithms. An aspect of this focus
involves the collection of additive sufficient statistics from data sets as a
strategy for training models. This approach trivially enables many useful
learning strategies, such as out-of-core learning, minibatch learning, and
semi-supervised learning, without requiring the user to consider how to
partition data or modify the algorithms to handle these tasks themselves.
pomegranate is written in Cython to speed up calculations and releases the
global interpreter lock to allow for built-in multithreaded parallelism, making
it competitive with---or outperform---other implementations of similar
algorithms. This paper presents an overview of the design choices in
pomegranate, and how they have enabled complex features to be supported by
simple code.
|
Training GANs with Optimism | We address the issue of limit cycling behavior in training Generative
Adversarial Networks and propose the use of Optimistic Mirror Decent (OMD) for
training Wasserstein GANs. Recent theoretical results have shown that
optimistic mirror decent (OMD) can enjoy faster regret rates in the context of
zero-sum games. WGANs is exactly a context of solving a zero-sum game with
simultaneous no-regret dynamics. Moreover, we show that optimistic mirror
decent addresses the limit cycling problem in training WGANs. We formally show
that in the case of bi-linear zero-sum games the last iterate of OMD dynamics
converges to an equilibrium, in contrast to GD dynamics which are bound to
cycle. We also portray the huge qualitative difference between GD and OMD
dynamics with toy examples, even when GD is modified with many adaptations
proposed in the recent literature, such as gradient penalty or momentum. We
apply OMD WGAN training to a bioinformatics problem of generating DNA
sequences. We observe that models trained with OMD achieve consistently smaller
KL divergence with respect to the true underlying distribution, than models
trained with GD variants. Finally, we introduce a new algorithm, Optimistic
Adam, which is an optimistic variant of Adam. We apply it to WGAN training on
CIFAR10 and observe improved performance in terms of inception score as
compared to Adam.
|
Sampling and Reconstruction of Graph Signals via Weak Submodularity and
Semidefinite Relaxation | We study the problem of sampling a bandlimited graph signal in the presence
of noise, where the objective is to select a node subset of prescribed
cardinality that minimizes the signal reconstruction mean squared error (MSE).
To that end, we formulate the task at hand as the minimization of MSE subject
to binary constraints, and approximate the resulting NP-hard problem via
semidefinite programming (SDP) relaxation. Moreover, we provide an alternative
formulation based on maximizing a monotone weak submodular function and propose
a randomized-greedy algorithm to find a sub-optimal subset. We then derive a
worst-case performance guarantee on the MSE returned by the randomized greedy
algorithm for general non-stationary graph signals. The efficacy of the
proposed methods is illustrated through numerical simulations on synthetic and
real-world graphs. Notably, the randomized greedy algorithm yields an
order-of-magnitude speedup over state-of-the-art greedy sampling schemes, while
incurring only a marginal MSE performance loss.
|
A multitask deep learning model for real-time deployment in embedded
systems | We propose an approach to Multitask Learning (MTL) to make deep learning
models faster and lighter for applications in which multiple tasks need to be
solved simultaneously, which is particularly useful in embedded, real-time
systems. We develop a multitask model for both Object Detection and Semantic
Segmentation and analyze the challenges that appear during its training. Our
multitask network is 1.6x faster, lighter and uses less memory than deploying
the single-task models in parallel. We conclude that MTL has the potential to
give superior performance in exchange of a more complex training process that
introduces challenges not present in single-task models.
|
Deep Neural Networks as Gaussian Processes | It has long been known that a single-layer fully-connected neural network
with an i.i.d. prior over its parameters is equivalent to a Gaussian process
(GP), in the limit of infinite network width. This correspondence enables exact
Bayesian inference for infinite width neural networks on regression tasks by
means of evaluating the corresponding GP. Recently, kernel functions which
mimic multi-layer random neural networks have been developed, but only outside
of a Bayesian framework. As such, previous work has not identified that these
kernels can be used as covariance functions for GPs and allow fully Bayesian
prediction with a deep neural network.
In this work, we derive the exact equivalence between infinitely wide deep
networks and GPs. We further develop a computationally efficient pipeline to
compute the covariance function for these GPs. We then use the resulting GPs to
perform Bayesian inference for wide deep neural networks on MNIST and CIFAR-10.
We observe that trained neural network accuracy approaches that of the
corresponding GP with increasing layer width, and that the GP uncertainty is
strongly correlated with trained network prediction error. We further find that
test performance increases as finite-width trained networks are made wider and
more similar to a GP, and thus that GP predictions typically outperform those
of finite-width networks. Finally we connect the performance of these GPs to
the recent theory of signal propagation in random neural networks.
|
Minimum Energy Quantized Neural Networks | This work targets the automated minimum-energy optimization of Quantized
Neural Networks (QNNs) - networks using low precision weights and activations.
These networks are trained from scratch at an arbitrary fixed point precision.
At iso-accuracy, QNNs using fewer bits require deeper and wider network
architectures than networks using higher precision operators, while they
require less complex arithmetic and less bits per weights. This fundamental
trade-off is analyzed and quantified to find the minimum energy QNN for any
benchmark and hence optimize energy-efficiency. To this end, the energy
consumption of inference is modeled for a generic hardware platform. This
allows drawing several conclusions across different benchmarks. First, energy
consumption varies orders of magnitude at iso-accuracy depending on the number
of bits used in the QNN. Second, in a typical system, BinaryNets or int4
implementations lead to the minimum energy solution, outperforming int8
networks up to 2-10x at iso-accuracy. All code used for QNN training is
available from https://github.com/BertMoons.
|
Stochastic Variational Inference for Bayesian Sparse Gaussian Process
Regression | This paper presents a novel variational inference framework for deriving a
family of Bayesian sparse Gaussian process regression (SGPR) models whose
approximations are variationally optimal with respect to the full-rank GPR
model enriched with various corresponding correlation structures of the
observation noises. Our variational Bayesian SGPR (VBSGPR) models jointly treat
both the distributions of the inducing variables and hyperparameters as
variational parameters, which enables the decomposability of the variational
lower bound that in turn can be exploited for stochastic optimization. Such a
stochastic optimization involves iteratively following the stochastic gradient
of the variational lower bound to improve its estimates of the optimal
variational distributions of the inducing variables and hyperparameters (and
hence the predictive distribution) of our VBSGPR models and is guaranteed to
achieve asymptotic convergence to them. We show that the stochastic gradient is
an unbiased estimator of the exact gradient and can be computed in constant
time per iteration, hence achieving scalability to big data. We empirically
evaluate the performance of our proposed framework on two real-world, massive
datasets.
|
Secure Classification With Augmented Features | With the evolution of data collection ways, it is possible to produce
abundant data described by multiple feature sets. Previous studies show that
including more features does not necessarily bring positive effect. How to
prevent the augmented features worsening classification performance is crucial
but rarely studied. In this paper, we study this challenging problem by
proposing a secure classification approach, whose accuracy is never degenerated
when exploiting augmented features. We propose two ways to achieve the security
of our method named as SEcure Classification (SEC). Firstly, to leverage
augmented features, we learn various types of classifiers and adapt them by
employing a specially designed robust loss. It provides various candidate
classifiers to meet the following assumption of security operation. Secondly,
we integrate all candidate classifiers by approximately maximizing the
performance improvement. Under a mild assumption, the integrated classifier has
theoretical security guarantee. Several new optimization methods have been
developed to accommodate the problems with proved convergence. Besides
evaluating SEC on 16 data sets, we also apply SEC in the application of
diagnostic classification of schizophrenia since it has vast application
potentiality. Experimental results demonstrate the effectiveness of SEC in both
tackling security problem and discriminating schizophrenic patients from
healthy controls.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.