title
stringlengths 7
246
| abstract
stringlengths 6
3.31k
|
|---|---|
The exploding gradient problem demystified - definition, prevalence,
impact, origin, tradeoffs, and solutions | Whereas it is believed that techniques such as Adam, batch normalization and,
more recently, SeLU nonlinearities "solve" the exploding gradient problem, we
show that this is not the case in general and that in a range of popular MLP
architectures, exploding gradients exist and that they limit the depth to which
networks can be effectively trained, both in theory and in practice. We explain
why exploding gradients occur and highlight the *collapsing domain problem*,
which can arise in architectures that avoid exploding gradients.
ResNets have significantly lower gradients and thus can circumvent the
exploding gradient problem, enabling the effective training of much deeper
networks. We show this is a direct consequence of the Pythagorean equation. By
noticing that *any neural network is a residual network*, we devise the
*residual trick*, which reveals that introducing skip connections simplifies
the network mathematically, and that this simplicity may be the major cause for
their success.
|
Invariant Synthesis for Incomplete Verification Engines | We propose a framework for synthesizing inductive invariants for incomplete
verification engines, which soundly reduce logical problems in undecidable
theories to decidable theories. Our framework is based on the counter-example
guided inductive synthesis principle (CEGIS) and allows verification engines to
communicate non-provability information to guide invariant synthesis. We show
precisely how the verification engine can compute such non-provability
information and how to build effective learning algorithms when invariants are
expressed as Boolean combinations of a fixed set of predicates. Moreover, we
evaluate our framework in two verification settings, one in which verification
engines need to handle quantified formulas and one in which verification
engines have to reason about heap properties expressed in an expressive but
undecidable separation logic. Our experiments show that our invariant synthesis
framework based on non-provability information can both effectively synthesize
inductive invariants and adequately strengthen contracts across a large suite
of programs.
|
Pre-training Attention Mechanisms | Recurrent neural networks with differentiable attention mechanisms have had
success in generative and classification tasks. We show that the classification
performance of such models can be enhanced by guiding a randomly initialized
model to attend to salient regions of the input in early training iterations.
We further show that, if explicit heuristics for guidance are unavailable, a
model that is pretrained on an unsupervised reconstruction task can discover
good attention policies without supervision. We demonstrate that increased
efficiency of the attention mechanism itself contributes to these performance
improvements. Based on these insights, we introduce bootstrapped glimpse
mimicking, a simple, theoretically task-general method of more effectively
training attention models. Our work draws inspiration from and parallels
results on human learning of attention.
|
BT-Nets: Simplifying Deep Neural Networks via Block Term Decomposition | Recently, deep neural networks (DNNs) have been regarded as the
state-of-the-art classification methods in a wide range of applications,
especially in image classification. Despite the success, the huge number of
parameters blocks its deployment to situations with light computing resources.
Researchers resort to the redundancy in the weights of DNNs and attempt to find
how fewer parameters can be chosen while preserving the accuracy at the same
time. Although several promising results have been shown along this research
line, most existing methods either fail to significantly compress a
well-trained deep network or require a heavy fine-tuning process for the
compressed network to regain the original performance. In this paper, we
propose the \textit{Block Term} networks (BT-nets) in which the commonly used
fully-connected layers (FC-layers) are replaced with block term layers
(BT-layers). In BT-layers, the inputs and the outputs are reshaped into two
low-dimensional high-order tensors, then block-term decomposition is applied as
tensor operators to connect them. We conduct extensive experiments on benchmark
datasets to demonstrate that BT-layers can achieve a very large compression
ratio on the number of parameters while preserving the representation power of
the original FC-layers as much as possible. Specifically, we can get a higher
performance while requiring fewer parameters compared with the tensor train
method.
|
Sockeye: A Toolkit for Neural Machine Translation | We describe Sockeye (version 1.12), an open-source sequence-to-sequence
toolkit for Neural Machine Translation (NMT). Sockeye is a production-ready
framework for training and applying models as well as an experimental platform
for researchers. Written in Python and built on MXNet, the toolkit offers
scalable training and inference for the three most prominent encoder-decoder
architectures: attentional recurrent neural networks, self-attentional
transformers, and fully convolutional networks. Sockeye also supports a wide
range of optimizers, normalization and regularization techniques, and inference
improvements from current NMT literature. Users can easily run standard
training recipes, explore different model settings, and incorporate new ideas.
In this paper, we highlight Sockeye's features and benchmark it against other
NMT toolkits on two language arcs from the 2017 Conference on Machine
Translation (WMT): English-German and Latvian-English. We report competitive
BLEU scores across all three architectures, including an overall best score for
Sockeye's transformer implementation. To facilitate further comparison, we
release all system outputs and training scripts used in our experiments. The
Sockeye toolkit is free software released under the Apache 2.0 license.
|
Lightweight Neural Networks | Most of the weights in a Lightweight Neural Network have a value of zero,
while the remaining ones are either +1 or -1. These universal approximators
require approximately 1.1 bits/weight of storage, posses a quick forward pass
and achieve classification accuracies similar to conventional continuous-weight
networks. Their training regimen focuses on error reduction initially, but
later emphasizes discretization of weights. They ignore insignificant inputs,
remove unnecessary weights, and drop unneeded hidden neurons. We have
successfully tested them on the MNIST, credit card fraud, and credit card
defaults data sets using networks having 2 to 16 hidden layers and up to 4.4
million weights.
|
Sentiment Predictability for Stocks | In this work, we present our findings and experiments for stock-market
prediction using various textual sentiment analysis tools, such as mood
analysis and event extraction, as well as prediction models, such as LSTMs and
specific convolutional architectures.
|
Deep Burst Denoising | Noise is an inherent issue of low-light image capture, one which is
exacerbated on mobile devices due to their narrow apertures and small sensors.
One strategy for mitigating noise in a low-light situation is to increase the
shutter time of the camera, thus allowing each photosite to integrate more
light and decrease noise variance. However, there are two downsides of long
exposures: (a) bright regions can exceed the sensor range, and (b) camera and
scene motion will result in blurred images. Another way of gathering more light
is to capture multiple short (thus noisy) frames in a "burst" and intelligently
integrate the content, thus avoiding the above downsides. In this paper, we use
the burst-capture strategy and implement the intelligent integration via a
recurrent fully convolutional deep neural net (CNN). We build our novel,
multiframe architecture to be a simple addition to any single frame denoising
model, and design to handle an arbitrary number of noisy input frames. We show
that it achieves state of the art denoising results on our burst dataset,
improving on the best published multi-frame techniques, such as VBM4D and
FlexISP. Finally, we explore other applications of image enhancement by
integrating content from multiple frames and demonstrate that our DNN
architecture generalizes well to image super-resolution.
|
Quantization and Training of Neural Networks for Efficient
Integer-Arithmetic-Only Inference | The rising popularity of intelligent mobile devices and the daunting
computational cost of deep learning-based models call for efficient and
accurate on-device inference schemes. We propose a quantization scheme that
allows inference to be carried out using integer-only arithmetic, which can be
implemented more efficiently than floating point inference on commonly
available integer-only hardware. We also co-design a training procedure to
preserve end-to-end model accuracy post quantization. As a result, the proposed
quantization scheme improves the tradeoff between accuracy and on-device
latency. The improvements are significant even on MobileNets, a model family
known for run-time efficiency, and are demonstrated in ImageNet classification
and COCO detection on popular CPUs.
|
On reproduction of On the regularization of Wasserstein GANs | This report has several purposes. First, our report is written to investigate
the reproducibility of the submitted paper On the regularization of Wasserstein
GANs (2018). Second, among the experiments performed in the submitted paper,
five aspects were emphasized and reproduced: learning speed, stability,
robustness against hyperparameter, estimating the Wasserstein distance, and
various sampling method. Finally, we identify which parts of the contribution
can be reproduced, and at what cost in terms of resources. All source code for
reproduction is open to the public.
|
Ray: A Distributed Framework for Emerging AI Applications | The next generation of AI applications will continuously interact with the
environment and learn from these interactions. These applications impose new
and demanding systems requirements, both in terms of performance and
flexibility. In this paper, we consider these requirements and present Ray---a
distributed system to address them. Ray implements a unified interface that can
express both task-parallel and actor-based computations, supported by a single
dynamic execution engine. To meet the performance requirements, Ray employs a
distributed scheduler and a distributed and fault-tolerant store to manage the
system's control state. In our experiments, we demonstrate scaling beyond 1.8
million tasks per second and better performance than existing specialized
systems for several challenging reinforcement learning applications.
|
Mitigating Asymmetric Nonlinear Weight Update Effects in Hardware Neural
Network based on Analog Resistive Synapse | Asymmetric nonlinear weight update is considered as one of the major
obstacles for realizing hardware neural networks based on analog resistive
synapses because it significantly compromises the online training capability.
This paper provides new solutions to this critical issue through
co-optimization with the hardware-applicable deep-learning algorithms. New
insights on engineering activation functions and a threshold weight update
scheme effectively suppress the undesirable training noise induced by
inaccurate weight update. We successfully trained a two-layer perceptron
network online and improved the classification accuracy of MNIST handwritten
digit dataset to 87.8/94.8% by using 6-bit/8-bit analog synapses, respectively,
with extremely high asymmetric nonlinearity.
|
Automatic Music Highlight Extraction using Convolutional Recurrent
Attention Networks | Music highlights are valuable contents for music services. Most methods
focused on low-level signal features. We propose a method for extracting
highlights using high-level features from convolutional recurrent attention
networks (CRAN). CRAN utilizes convolution and recurrent layers for sequential
learning with an attention mechanism. The attention allows CRAN to capture
significant snippets for distinguishing between genres, thus being used as a
high-level feature. CRAN was evaluated on over 32,000 popular tracks in Korea
for two months. Experimental results show our method outperforms three baseline
methods through quantitative and qualitative evaluations. Also, we analyze the
effects of attention and sequence information on performance.
|
NSML: A Machine Learning Platform That Enables You to Focus on Your
Models | Machine learning libraries such as TensorFlow and PyTorch simplify model
implementation. However, researchers are still required to perform a
non-trivial amount of manual tasks such as GPU allocation, training status
tracking, and comparison of models with different hyperparameter settings. We
propose a system to handle these tasks and help researchers focus on models. We
present the requirements of the system based on a collection of discussions
from an online study group comprising 25k members. These include automatic GPU
allocation, learning status visualization, handling model parameter snapshots
as well as hyperparameter modification during learning, and comparison of
performance metrics between models via a leaderboard. We describe the system
architecture that fulfills these requirements and present a proof-of-concept
implementation, NAVER Smart Machine Learning (NSML). We test the system and
confirm substantial efficiency improvements for model development.
|
Cyberattack Detection in Mobile Cloud Computing: A Deep Learning
Approach | With the rapid growth of mobile applications and cloud computing, mobile
cloud computing has attracted great interest from both academia and industry.
However, mobile cloud applications are facing security issues such as data
integrity, users' confidentiality, and service availability. A preventive
approach to such problems is to detect and isolate cyber threats before they
can cause serious impacts to the mobile cloud computing system. In this paper,
we propose a novel framework that leverages a deep learning approach to detect
cyberattacks in mobile cloud environment. Through experimental results, we show
that our proposed framework not only recognizes diverse cyberattacks, but also
achieves a high accuracy (up to 97.11%) in detecting the attacks. Furthermore,
we present the comparisons with current machine learning-based approaches to
demonstrate the effectiveness of our proposed solution.
|
A Machine Learning Framework for Resource Allocation Assisted by Cloud
Computing | Conventionally, the resource allocation is formulated as an optimization
problem and solved online with instantaneous scenario information. Since most
resource allocation problems are not convex, the optimal solutions are very
difficult to be obtained in real time. Lagrangian relaxation or greedy methods
are then often employed, which results in performance loss. Therefore, the
conventional methods of resource allocation are facing great challenges to meet
the ever-increasing QoS requirements of users with scarce radio resource.
Assisted by cloud computing, a huge amount of historical data on scenarios can
be collected for extracting similarities among scenarios using machine
learning. Moreover, optimal or near-optimal solutions of historical scenarios
can be searched offline and stored in advance. When the measured data of
current scenario arrives, the current scenario is compared with historical
scenarios to find the most similar one. Then, the optimal or near-optimal
solution in the most similar historical scenario is adopted to allocate the
radio resources for the current scenario. To facilitate the application of new
design philosophy, a machine learning framework is proposed for resource
allocation assisted by cloud computing. An example of beam allocation in
multi-user massive multiple-input-multiple-output (MIMO) systems shows that the
proposed machine-learning based resource allocation outperforms conventional
methods.
|
NDT: Neual Decision Tree Towards Fully Functioned Neural Graph | Though traditional algorithms could be embedded into neural architectures
with the proposed principle of \cite{xiao2017hungarian}, the variables that
only occur in the condition of branch could not be updated as a special case.
To tackle this issue, we multiply the conditioned branches with Dirac symbol
(i.e. $\mathbf{1}_{x>0}$), then approximate Dirac symbol with the continuous
functions (e.g. $1 - e^{-\alpha|x|}$). In this way, the gradients of
condition-specific variables could be worked out in the back-propagation
process, approximately, making a fully functioned neural graph. Within our
novel principle, we propose the neural decision tree \textbf{(NDT)}, which
takes simplified neural networks as decision function in each branch and
employs complex neural networks to generate the output in each leaf. Extensive
experiments verify our theoretical analysis and demonstrate the effectiveness
of our model.
|
An Artificial Neural Network Architecture Based on Context
Transformations in Cortical Minicolumns | Cortical minicolumns are considered a model of cortical organization. Their
function is still a source of research and not reflected properly in modern
architecture of nets in algorithms of Artificial Intelligence. We assume its
function and describe it in this article. Furthermore, we show how this
proposal allows to construct a new architecture, that is not based on
convolutional neural networks, test it on MNIST data and receive close to
Convolutional Neural Network accuracy. We also show that the proposed
architecture possesses an ability to train on a small quantity of samples. To
achieve these results, we enable the minicolumns to remember context
transformations.
|
Taming Wild High Dimensional Text Data with a Fuzzy Lash | The bag of words (BOW) represents a corpus in a matrix whose elements are the
frequency of words. However, each row in the matrix is a very high-dimensional
sparse vector. Dimension reduction (DR) is a popular method to address sparsity
and high-dimensionality issues. Among different strategies to develop DR
method, Unsupervised Feature Transformation (UFT) is a popular strategy to map
all words on a new basis to represent BOW. The recent increase of text data and
its challenges imply that DR area still needs new perspectives. Although a wide
range of methods based on the UFT strategy has been developed, the fuzzy
approach has not been considered for DR based on this strategy. This research
investigates the application of fuzzy clustering as a DR method based on the
UFT strategy to collapse BOW matrix to provide a lower-dimensional
representation of documents instead of the words in a corpus. The quantitative
evaluation shows that fuzzy clustering produces superior performance and
features to Principal Components Analysis (PCA) and Singular Value
Decomposition (SVD), two popular DR methods based on the UFT strategy.
|
StackInsights: Cognitive Learning for Hybrid Cloud Readiness | Hybrid cloud is an integrated cloud computing environment utilizing a mix of
public cloud, private cloud, and on-premise traditional IT infrastructures.
Workload awareness, defined as a detailed full range understanding of each
individual workload, is essential in implementing the hybrid cloud. While it is
critical to perform an accurate analysis to determine which workloads are
appropriate for on-premise deployment versus which workloads can be migrated to
a cloud off-premise, the assessment is mainly performed by rule or policy based
approaches. In this paper, we introduce StackInsights, a novel cognitive system
to automatically analyze and predict the cloud readiness of workloads for an
enterprise. Our system harnesses the critical metrics across the entire stack:
1) infrastructure metrics, 2) data relevance metrics, and 3) application
taxonomy, to identify workloads that have characteristics of a) low sensitivity
with respect to business security, criticality and compliance, and b) low
response time requirements and access patterns. Since the capture of the data
relevance metrics involves an intrusive and in-depth scanning of the content of
storage objects, a machine learning model is applied to perform the business
relevance classification by learning from the meta level metrics harnessed
across stack. In contrast to traditional methods, StackInsights significantly
reduces the total time for hybrid cloud readiness assessment by orders of
magnitude.
|
Avoiding Synchronization in First-Order Methods for Sparse Convex
Optimization | Parallel computing has played an important role in speeding up convex
optimization methods for big data analytics and large-scale machine learning
(ML). However, the scalability of these optimization methods is inhibited by
the cost of communicating and synchronizing processors in a parallel setting.
Iterative ML methods are particularly sensitive to communication cost since
they often require communication every iteration. In this work, we extend
well-known techniques from Communication-Avoiding Krylov subspace methods to
first-order, block coordinate descent methods for Support Vector Machines and
Proximal Least-Squares problems. Our Synchronization-Avoiding (SA) variants
reduce the latency cost by a tunable factor of $s$ at the expense of a factor
of $s$ increase in flops and bandwidth costs. We show that the SA-variants are
numerically stable and can attain large speedups of up to $5.1\times$ on a Cray
XC30 supercomputer.
|
Wasserstein Distributionally Robust Optimization and Variation
Regularization | Wasserstein distributionally robust optimization (DRO) has recently achieved
empirical success for various applications in operations research and machine
learning, owing partly to its regularization effect. Although connection
between Wasserstein DRO and regularization has been established in several
settings, existing results often require restrictive assumptions, such as
smoothness or convexity, that are not satisfied for many problems. In this
paper, we develop a general theory on the variation regularization effect of
the Wasserstein DRO - a new form of regularization that generalizes
total-variation regularization, Lipschitz regularization and gradient
regularization. Our results cover possibly non-convex and non-smooth losses and
losses on non-Euclidean spaces. Examples include multi-item newsvendor,
portfolio selection, linear prediction, neural networks, manifold learning, and
intensity estimation for Poisson processes, etc. As an application of our
theory of variation regularization, we derive new generalization guarantees for
adversarial robust learning.
|
Using Deep learning methods for generation of a personalized list of
shuffled songs | The shuffle mode, where songs are played in a randomized order that is
decided upon for all tracks at once, is widely found and known to exist in
music player systems. There are only few music enthusiasts who use this mode
since it either is too random to suit their mood or it keeps on repeating the
same list every time. In this paper, we propose to build a convolutional deep
belief network(CDBN) that is trained to perform genre recognition based on
audio features retrieved from the records of the Million Song Dataset. The
learned parameters shall be used to initialize a multi-layer perceptron which
takes extracted features of user's playlist as input alongside the metadata to
classify to various categories. These categories will be shuffled
retrospectively based on the metadata to autonomously provide with a list that
is efficacious in playing songs that are desired by humans in normal
conditions.
|
"Zero-Shot" Super-Resolution using Deep Internal Learning | Deep Learning has led to a dramatic leap in Super-Resolution (SR) performance
in the past few years. However, being supervised, these SR methods are
restricted to specific training data, where the acquisition of the
low-resolution (LR) images from their high-resolution (HR) counterparts is
predetermined (e.g., bicubic downscaling), without any distracting artifacts
(e.g., sensor noise, image compression, non-ideal PSF, etc). Real LR images,
however, rarely obey these restrictions, resulting in poor SR results by SotA
(State of the Art) methods. In this paper we introduce "Zero-Shot" SR, which
exploits the power of Deep Learning, but does not rely on prior training. We
exploit the internal recurrence of information inside a single image, and train
a small image-specific CNN at test time, on examples extracted solely from the
input image itself. As such, it can adapt itself to different settings per
image. This allows to perform SR of real old photos, noisy images, biological
data, and other images where the acquisition process is unknown or non-ideal.
On such images, our method outperforms SotA CNN-based SR methods, as well as
previous unsupervised SR methods. To the best of our knowledge, this is the
first unsupervised CNN-based SR method.
|
Efficient B-mode Ultrasound Image Reconstruction from Sub-sampled RF
Data using Deep Learning | In portable, three dimensional, and ultra-fast ultrasound imaging systems,
there is an increasing demand for the reconstruction of high quality images
from a limited number of radio-frequency (RF) measurements due to receiver (Rx)
or transmit (Xmit) event sub-sampling. However, due to the presence of side
lobe artifacts from RF sub-sampling, the standard beamformer often produces
blurry images with less contrast, which are unsuitable for diagnostic purposes.
Existing compressed sensing approaches often require either hardware changes or
computationally expensive algorithms, but their quality improvements are
limited. To address this problem, here we propose a novel deep learning
approach that directly interpolates the missing RF data by utilizing redundancy
in the Rx-Xmit plane. Our extensive experimental results using sub-sampled RF
data from a multi-line acquisition B-mode system confirm that the proposed
method can effectively reduce the data rate without sacrificing image quality.
|
Integral Equations and Machine Learning | As both light transport simulation and reinforcement learning are ruled by
the same Fredholm integral equation of the second kind, reinforcement learning
techniques may be used for photorealistic image synthesis: Efficiency may be
dramatically improved by guiding light transport paths by an approximate
solution of the integral equation that is learned during rendering. In the
light of the recent advances in reinforcement learning for playing games, we
investigate the representation of an approximate solution of an integral
equation by artificial neural networks and derive a loss function for that
purpose. The resulting Monte Carlo and quasi-Monte Carlo methods train neural
networks with standard information instead of linear information and naturally
are able to generate an arbitrary number of training samples. The methods are
demonstrated for applications in light transport simulation.
|
Super-sparse Learning in Similarity Spaces | In several applications, input samples are more naturally represented in
terms of similarities between each other, rather than in terms of feature
vectors. In these settings, machine-learning algorithms can become very
computationally demanding, as they may require matching the test samples
against a very large set of reference prototypes. To mitigate this issue,
different approaches have been developed to reduce the number of required
reference prototypes. Current reduction approaches select a small subset of
representative prototypes in the space induced by the similarity measure, and
then separately train the classification function on the reduced subset.
However, decoupling these two steps may not allow reducing the number of
prototypes effectively without compromising accuracy. We overcome this
limitation by jointly learning the classification function along with an
optimal set of virtual prototypes, whose number can be either fixed a priori or
optimized according to application-specific criteria. Creating a super-sparse
set of virtual prototypes provides much sparser solutions, drastically reducing
complexity at test time, at the expense of a slightly increased complexity
during training. A much smaller set of prototypes also results in
easier-to-interpret decisions. We empirically show that our approach can reduce
up to ten times the complexity of Support Vector Machines, LASSO and ridge
regression at test time, without almost affecting their classification
accuracy.
|
Dynamic Boltzmann Machines for Second Order Moments and Generalized
Gaussian Distributions | Dynamic Boltzmann Machine (DyBM) has been shown highly efficient to predict
time-series data. Gaussian DyBM is a DyBM that assumes the predicted data is
generated by a Gaussian distribution whose first-order moment (mean)
dynamically changes over time but its second-order moment (variance) is fixed.
However, in many financial applications, the assumption is quite limiting in
two aspects. First, even when the data follows a Gaussian distribution, its
variance may change over time. Such variance is also related to important
temporal economic indicators such as the market volatility. Second, financial
time-series data often requires learning datasets generated by the generalized
Gaussian distribution with an additional shape parameter that is important to
approximate heavy-tailed distributions. Addressing those aspects, we show how
to extend DyBM that results in significant performance improvement in
predicting financial time-series data.
|
TensorFlow-Serving: Flexible, High-Performance ML Serving | We describe TensorFlow-Serving, a system to serve machine learning models
inside Google which is also available in the cloud and via open-source. It is
extremely flexible in terms of the types of ML platforms it supports, and ways
to integrate with systems that convey new models and updated versions from
training to serving. At the same time, the core code paths around model lookup
and inference have been carefully optimized to avoid performance pitfalls
observed in naive implementations. Google uses it in many production
deployments, including a multi-tenant model hosting service called TFS^2.
|
Generating and designing DNA with deep generative models | We propose generative neural network methods to generate DNA sequences and
tune them to have desired properties. We present three approaches: creating
synthetic DNA sequences using a generative adversarial network; a DNA-based
variant of the activation maximization ("deep dream") design method; and a
joint procedure which combines these two approaches together. We show that
these tools capture important structures of the data and, when applied to
designing probes for protein binding microarrays, allow us to generate new
sequences whose properties are estimated to be superior to those found in the
training data. We believe that these results open the door for applying deep
generative models to advance genomics research.
|
Deep Neural Networks as 0-1 Mixed Integer Linear Programs: A Feasibility
Study | Deep Neural Networks (DNNs) are very popular these days, and are the subject
of a very intense investigation. A DNN is made by layers of internal units (or
neurons), each of which computes an affine combination of the output of the
units in the previous layer, applies a nonlinear operator, and outputs the
corresponding value (also known as activation). A commonly-used nonlinear
operator is the so-called rectified linear unit (ReLU), whose output is just
the maximum between its input value and zero. In this (and other similar cases
like max pooling, where the max operation involves more than one input value),
one can model the DNN as a 0-1 Mixed Integer Linear Program (0-1 MILP) where
the continuous variables correspond to the output values of each unit, and a
binary variable is associated with each ReLU to model its yes/no nature. In
this paper we discuss the peculiarity of this kind of 0-1 MILP models, and
describe an effective bound-tightening technique intended to ease its solution.
We also present possible applications of the 0-1 MILP model arising in feature
visualization and in the construction of adversarial examples. Preliminary
computational results are reported, aimed at investigating (on small DNNs) the
computational performance of a state-of-the-art MILP solver when applied to a
known test case, namely, hand-written digit recognition.
|
Structured Optimal Transport | Optimal Transport has recently gained interest in machine learning for
applications ranging from domain adaptation, sentence similarities to deep
learning. Yet, its ability to capture frequently occurring structure beyond the
"ground metric" is limited. In this work, we develop a nonlinear generalization
of (discrete) optimal transport that is able to reflect much additional
structure. We demonstrate how to leverage the geometry of this new model for
fast algorithms, and explore connections and properties. Illustrative
experiments highlight the benefit of the induced structured couplings for tasks
in domain adaptation and natural language processing.
|
Predicting Individual Physiologically Acceptable States for Discharge
from a Pediatric Intensive Care Unit | Objective: Predict patient-specific vitals deemed medically acceptable for
discharge from a pediatric intensive care unit (ICU). Design: The means of each
patient's hr, sbp and dbp measurements between their medical and physical
discharge from the ICU were computed as a proxy for their physiologically
acceptable state space (PASS) for successful ICU discharge. These individual
PASS values were compared via root mean squared error (rMSE) to population
age-normal vitals, a polynomial regression through the PASS values of a
Pediatric ICU (PICU) population and predictions from two recurrent neural
network models designed to predict personalized PASS within the first twelve
hours following ICU admission. Setting: PICU at Children's Hospital Los Angeles
(CHLA). Patients: 6,899 PICU episodes (5,464 patients) collected between 2009
and 2016. Interventions: None. Measurements: Each episode data contained 375
variables representing vitals, labs, interventions, and drugs. They also
included a time indicator for PICU medical discharge and physical discharge.
Main Results: The rMSEs between individual PASS values and population
age-normals (hr: 25.9 bpm, sbp: 13.4 mmHg, dbp: 13.0 mmHg) were larger than the
rMSEs corresponding to the polynomial regression (hr: 19.1 bpm, sbp: 12.3 mmHg,
dbp: 10.8 mmHg). The rMSEs from the best performing RNN model were the lowest
(hr: 16.4 bpm; sbp: 9.9 mmHg, dbp: 9.0 mmHg). Conclusion: PICU patients are a
unique subset of the general population, and general age-normal vitals may not
be suitable as target values indicating physiologic stability at discharge.
Age-normal vitals that were specifically derived from the medical-to-physical
discharge window of ICU patients may be more appropriate targets for
'acceptable' physiologic state for critical care patients. Going beyond simple
age bins, an RNN model can provide more personalized target values.
|
Visual Explanations from Hadamard Product in Multimodal Deep Networks | The visual explanation of learned representation of models helps to
understand the fundamentals of learning. The attentional models of previous
works used to visualize the attended regions over an image or text using their
learned weights to confirm their intended mechanism. Kim et al. (2016) show
that the Hadamard product in multimodal deep networks, which is well-known for
the joint function of visual question answering tasks, implicitly performs an
attentional mechanism for visual inputs. In this work, we extend their work to
show that the Hadamard product in multimodal deep networks performs not only
for visual inputs but also for textual inputs simultaneously using the proposed
gradient-based visualization technique. The attentional effect of Hadamard
product is visualized for both visual and textual inputs by analyzing the two
inputs and an output of the Hadamard product with the proposed method and
compared with learned attentional weights of a visual question answering model.
|
Misspecified Nonconvex Statistical Optimization for Phase Retrieval | Existing nonconvex statistical optimization theory and methods crucially rely
on the correct specification of the underlying "true" statistical models. To
address this issue, we take a first step towards taming model misspecification
by studying the high-dimensional sparse phase retrieval problem with
misspecified link functions. In particular, we propose a simple variant of the
thresholded Wirtinger flow algorithm that, given a proper initialization,
linearly converges to an estimator with optimal statistical accuracy for a
broad family of unknown link functions. We further provide extensive numerical
experiments to support our theoretical findings.
|
A Survey on Multi-View Clustering | With advances in information acquisition technologies, multi-view data become
ubiquitous. Multi-view learning has thus become more and more popular in
machine learning and data mining fields. Multi-view unsupervised or
semi-supervised learning, such as co-training, co-regularization has gained
considerable attention. Although recently, multi-view clustering (MVC) methods
have been developed rapidly, there has not been a survey to summarize and
analyze the current progress. Therefore, this paper reviews the common
strategies for combining multiple views of data and based on this summary we
propose a novel taxonomy of the MVC approaches. We further discuss the
relationships between MVC and multi-view representation, ensemble clustering,
multi-task clustering, multi-view supervised and semi-supervised learning.
Several representative real-world applications are elaborated. To promote
future development of MVC, we envision several open problems that may require
further investigation and thorough examination.
|
Deep Neural Generative Model of Functional MRI Images for Psychiatric
Disorder Diagnosis | Accurate diagnosis of psychiatric disorders plays a critical role in
improving the quality of life for patients and potentially supports the
development of new treatments. Many studies have been conducted on machine
learning techniques that seek brain imaging data for specific biomarkers of
disorders. These studies have encountered the following dilemma: A direct
classification overfits to a small number of high-dimensional samples but
unsupervised feature-extraction has the risk of extracting a signal of no
interest. In addition, such studies often provided only diagnoses for patients
without presenting the reasons for these diagnoses. This study proposed a deep
neural generative model of resting-state functional magnetic resonance imaging
(fMRI) data. The proposed model is conditioned by the assumption of the
subject's state and estimates the posterior probability of the subject's state
given the imaging data, using Bayes' rule. This study applied the proposed
model to diagnose schizophrenia and bipolar disorders. Diagnostic accuracy was
improved by a large margin over competitive approaches, namely classifications
of functional connectivity, discriminative/generative models of region-wise
signals, and those with unsupervised feature-extractors. The proposed model
visualizes brain regions largely related to the disorders, thus motivating
further biological investigation.
|
A New Data-Driven Sparse-Learning Approach to Study Chemical Reaction
Networks | Chemical kinetic mechanisms can be represented by sets of elementary
reactions that are easily translated into mathematical terms using
physicochemical relationships. The schematic representation of reactions
captures the interactions between reacting species and products. Determining
the minimal chemical interactions underlying the dynamic behavior of systems is
a major task. In this paper, we introduce a novel approach for the
identification of the influential reactions in chemical reaction networks for
combustion applications, using a data-driven sparse-learning technique. The
proposed approach identifies a set of influential reactions using species
concentrations and reaction rates, with minimal computational cost without
requiring additional data or simulations. The new approach is applied to
analyze the combustion chemistry of H2 and C3H8 in a constant-volume
homogeneous reactor. The influential reactions identified by the
sparse-learning method are consistent with the current kinetics knowledge of
chemical mechanisms. Additionally, we show that a reduced version of the parent
mechanism can be generated as a combination of the influential reactions
identified at different times and conditions and that for both H2 and C3H8 this
reduced mechanism performs closely to the parent mechanism as a function of
ignition delay over a wide range of conditions. Our results demonstrate the
potential of the sparse-learning approach as an effective and efficient tool
for mechanism analysis and mechanism reduction.
|
A Bridge Between Hyperparameter Optimization and Learning-to-learn | We consider a class of a nested optimization problems involving inner and
outer objectives. We observe that by taking into explicit account the
optimization dynamics for the inner objective it is possible to derive a
general framework that unifies gradient-based hyperparameter optimization and
meta-learning (or learning-to-learn). Depending on the specific setting, the
variables of the outer objective take either the meaning of hyperparameters in
a supervised learning problem or parameters of a meta-learner. We show that
some recently proposed methods in the latter setting can be instantiated in our
framework and tackled with the same gradient-based algorithms. Finally, we
discuss possible design patterns for learning-to-learn and present encouraging
preliminary experiments for few-shot learning.
|
Visual Explanation by Interpretation: Improving Visual Feedback
Capabilities of Deep Neural Networks | Interpretation and explanation of deep models is critical towards wide
adoption of systems that rely on them. In this paper, we propose a novel scheme
for both interpretation as well as explanation in which, given a pretrained
model, we automatically identify internal features relevant for the set of
classes considered by the model, without relying on additional annotations. We
interpret the model through average visualizations of this reduced set of
features. Then, at test time, we explain the network prediction by accompanying
the predicted class label with supporting visualizations derived from the
identified features. In addition, we propose a method to address the artifacts
introduced by stridded operations in deconvNet-based visualizations. Moreover,
we introduce an8Flower, a dataset specifically designed for objective
quantitative evaluation of methods for visual explanation.Experiments on the
MNIST,ILSVRC12,Fashion144k and an8Flower datasets show that our method produces
detailed explanations with good coverage of relevant features of the classes of
interest
|
Language and Noise Transfer in Speech Enhancement Generative Adversarial
Network | Speech enhancement deep learning systems usually require large amounts of
training data to operate in broad conditions or real applications. This makes
the adaptability of those systems into new, low resource environments an
important topic. In this work, we present the results of adapting a speech
enhancement generative adversarial network by finetuning the generator with
small amounts of data. We investigate the minimum requirements to obtain a
stable behavior in terms of several objective metrics in two very different
languages: Catalan and Korean. We also study the variability of test
performance to unseen noise as a function of the amount of different types of
noise available for training. Results show that adapting a pre-trained English
model with 10 min of data already achieves a comparable performance to having
two orders of magnitude more data. They also demonstrate the relative stability
in test performance with respect to the number of training noise types.
|
Squeezed Convolutional Variational AutoEncoder for Unsupervised Anomaly
Detection in Edge Device Industrial Internet of Things | In this paper, we propose Squeezed Convolutional Variational AutoEncoder
(SCVAE) for anomaly detection in time series data for Edge Computing in
Industrial Internet of Things (IIoT). The proposed model is applied to labeled
time series data from UCI datasets for exact performance evaluation, and
applied to real world data for indirect model performance comparison. In
addition, by comparing the models before and after applying Fire Modules from
SqueezeNet, we show that model size and inference times are reduced while
similar levels of performance is maintained.
|
On the Effectiveness of Least Squares Generative Adversarial Networks | Unsupervised learning with generative adversarial networks (GANs) has proven
to be hugely successful. Regular GANs hypothesize the discriminator as a
classifier with the sigmoid cross entropy loss function. However, we found that
this loss function may lead to the vanishing gradients problem during the
learning process. To overcome such a problem, we propose in this paper the
Least Squares Generative Adversarial Networks (LSGANs) which adopt the least
squares loss for both the discriminator and the generator. We show that
minimizing the objective function of LSGAN yields minimizing the Pearson
$\chi^2$ divergence. We also show that the derived objective function that
yields minimizing the Pearson $\chi^2$ divergence performs better than the
classical one of using least squares for classification. There are two benefits
of LSGANs over regular GANs. First, LSGANs are able to generate higher quality
images than regular GANs. Second, LSGANs perform more stably during the
learning process. For evaluating the image quality, we conduct both qualitative
and quantitative experiments, and the experimental results show that LSGANs can
generate higher quality images than regular GANs. Furthermore, we evaluate the
stability of LSGANs in two groups. One is to compare between LSGANs and regular
GANs without gradient penalty. We conduct three experiments, including Gaussian
mixture distribution, difficult architectures, and a newly proposed method ---
datasets with small variability, to illustrate the stability of LSGANs. The
other one is to compare between LSGANs with gradient penalty (LSGANs-GP) and
WGANs with gradient penalty (WGANs-GP). The experimental results show that
LSGANs-GP succeed in training for all the difficult architectures used in
WGANs-GP, including 101-layer ResNet.
|
Learning to Write Stylized Chinese Characters by Reading a Handful of
Examples | Automatically writing stylized Chinese characters is an attractive yet
challenging task due to its wide applicabilities. In this paper, we propose a
novel framework named Style-Aware Variational Auto-Encoder (SA-VAE) to flexibly
generate Chinese characters. Specifically, we propose to capture the different
characteristics of a Chinese character by disentangling the latent features
into content-related and style-related components. Considering of the complex
shapes and structures, we incorporate the structure information as prior
knowledge into our framework to guide the generation. Our framework shows a
powerful one-shot/low-shot generalization ability by inferring the style
component given a character with unseen style. To the best of our knowledge,
this is the first attempt to learn to write new-style Chinese characters by
observing only one or a few examples. Extensive experiments demonstrate its
effectiveness in generating different stylized Chinese characters by fusing the
feature vectors corresponding to different contents and styles, which is of
significant importance in real-world applications.
|
A Shapelet Transform for Multivariate Time Series Classification | Shapelets are phase independent subsequences designed for time series
classification. We propose three adaptations to the Shapelet Transform (ST) to
capture multivariate features in multivariate time series classification. We
create a unified set of data to benchmark our work on, and compare with three
other algorithms. We demonstrate that multivariate shapelets are not
significantly worse than other state-of-the-art algorithms.
|
Dynamic Weight Alignment for Temporal Convolutional Neural Networks | In this paper, we propose a method of improving temporal Convolutional Neural
Networks (CNN) by determining the optimal alignment of weights and inputs using
dynamic programming. Conventional CNN convolutions linearly match the shared
weights to a window of the input. However, it is possible that there exists a
more optimal alignment of weights. Thus, we propose the use of Dynamic Time
Warping (DTW) to dynamically align the weights to the input of the
convolutional layer. Specifically, the dynamic alignment overcomes issues such
as temporal distortion by finding the minimal distance matching of the weights
and the inputs under constraints. We demonstrate the effectiveness of the
proposed architecture on the Unipen online handwritten digit and character
datasets, the UCI Spoken Arabic Digit dataset, and the UCI Activities of Daily
Life dataset.
|
Nonparametric Inference for Auto-Encoding Variational Bayes | We would like to learn latent representations that are low-dimensional and
highly interpretable. A model that has these characteristics is the Gaussian
Process Latent Variable Model. The benefits and negative of the GP-LVM are
complementary to the Variational Autoencoder, the former provides interpretable
low-dimensional latent representations while the latter is able to handle large
amounts of data and can use non-Gaussian likelihoods. Our inspiration for this
paper is to marry these two approaches and reap the benefits of both. In order
to do so we will introduce a novel approximate inference scheme inspired by the
GP-LVM and the VAE. We show experimentally that the approximation allows the
capacity of the generative bottle-neck (Z) of the VAE to be arbitrarily large
without losing a highly interpretable representation, allowing reconstruction
quality to be unlimited by Z at the same time as a low-dimensional space can be
used to perform ancestral sampling from as well as a means to reason about the
embedded data.
|
Size-Independent Sample Complexity of Neural Networks | We study the sample complexity of learning neural networks, by providing new
bounds on their Rademacher complexity assuming norm constraints on the
parameter matrix of each layer. Compared to previous work, these complexity
bounds have improved dependence on the network depth, and under some additional
assumptions, are fully independent of the network size (both depth and width).
These results are derived using some novel techniques, which may be of
independent interest.
|
The Power of Interpolation: Understanding the Effectiveness of SGD in
Modern Over-parametrized Learning | In this paper we aim to formally explain the phenomenon of fast convergence
of SGD observed in modern machine learning. The key observation is that most
modern learning architectures are over-parametrized and are trained to
interpolate the data by driving the empirical loss (classification and
regression) close to zero. While it is still unclear why these interpolated
solutions perform well on test data, we show that these regimes allow for fast
convergence of SGD, comparable in number of iterations to full gradient
descent.
For convex loss functions we obtain an exponential convergence bound for {\it
mini-batch} SGD parallel to that for full gradient descent. We show that there
is a critical batch size $m^*$ such that: (a) SGD iteration with mini-batch
size $m\leq m^*$ is nearly equivalent to $m$ iterations of mini-batch size $1$
(\emph{linear scaling regime}). (b) SGD iteration with mini-batch $m> m^*$ is
nearly equivalent to a full gradient descent iteration (\emph{saturation
regime}).
Moreover, for the quadratic loss, we derive explicit expressions for the
optimal mini-batch and step size and explicitly characterize the two regimes
above. The critical mini-batch size can be viewed as the limit for effective
mini-batch parallelization. It is also nearly independent of the data size,
implying $O(n)$ acceleration over GD per unit of computation. We give
experimental evidence on real data which closely follows our theoretical
analyses.
Finally, we show how our results fit in the recent developments in training
deep neural networks and discuss connections to adaptive rates for SGD and
variance reduction.
|
Deep Neuroevolution: Genetic Algorithms Are a Competitive Alternative
for Training Deep Neural Networks for Reinforcement Learning | Deep artificial neural networks (DNNs) are typically trained via
gradient-based learning algorithms, namely backpropagation. Evolution
strategies (ES) can rival backprop-based algorithms such as Q-learning and
policy gradients on challenging deep reinforcement learning (RL) problems.
However, ES can be considered a gradient-based algorithm because it performs
stochastic gradient descent via an operation similar to a finite-difference
approximation of the gradient. That raises the question of whether
non-gradient-based evolutionary algorithms can work at DNN scales. Here we
demonstrate they can: we evolve the weights of a DNN with a simple,
gradient-free, population-based genetic algorithm (GA) and it performs well on
hard deep RL problems, including Atari and humanoid locomotion. The Deep GA
successfully evolves networks with over four million free parameters, the
largest neural networks ever evolved with a traditional evolutionary algorithm.
These results (1) expand our sense of the scale at which GAs can operate, (2)
suggest intriguingly that in some cases following the gradient is not the best
choice for optimizing performance, and (3) make immediately available the
multitude of neuroevolution techniques that improve performance. We demonstrate
the latter by showing that combining DNNs with novelty search, which encourages
exploration on tasks with deceptive or sparse reward functions, can solve a
high-dimensional problem on which reward-maximizing algorithms (e.g.\ DQN, A3C,
ES, and the GA) fail. Additionally, the Deep GA is faster than ES, A3C, and DQN
(it can train Atari in ${\raise.17ex\hbox{$\scriptstyle\sim$}}$4 hours on one
desktop or ${\raise.17ex\hbox{$\scriptstyle\sim$}}$1 hour distributed on 720
cores), and enables a state-of-the-art, up to 10,000-fold compact encoding
technique.
|
Parallel Complexity of Forward and Backward Propagation | We show that the forward and backward propagation can be formulated as a
solution of lower and upper triangular systems of equations. For standard
feedforward (FNNs) and recurrent neural networks (RNNs) the triangular systems
are always block bi-diagonal, while for a general computation graph (directed
acyclic graph) they can have a more complex triangular sparsity pattern. We
discuss direct and iterative parallel algorithms that can be used for their
solution and interpreted as different ways of performing model parallelism.
Also, we show that for FNNs and RNNs with $k$ layers and $\tau$ time steps the
backward propagation can be performed in parallel in O($\log k$) and O($\log k
\log \tau$) steps, respectively. Finally, we outline the generalization of this
technique using Jacobians that potentially allows us to handle arbitrary
layers.
|
Third-order Smoothness Helps: Even Faster Stochastic Optimization
Algorithms for Finding Local Minima | We propose stochastic optimization algorithms that can find local minima
faster than existing algorithms for nonconvex optimization problems, by
exploiting the third-order smoothness to escape non-degenerate saddle points
more efficiently. More specifically, the proposed algorithm only needs
$\tilde{O}(\epsilon^{-10/3})$ stochastic gradient evaluations to converge to an
approximate local minimum $\mathbf{x}$, which satisfies $\|\nabla
f(\mathbf{x})\|_2\leq\epsilon$ and $\lambda_{\min}(\nabla^2 f(\mathbf{x}))\geq
-\sqrt{\epsilon}$ in the general stochastic optimization setting, where
$\tilde{O}(\cdot)$ hides logarithm polynomial terms and constants. This
improves upon the $\tilde{O}(\epsilon^{-7/2})$ gradient complexity achieved by
the state-of-the-art stochastic local minima finding algorithms by a factor of
$\tilde{O}(\epsilon^{-1/6})$. For nonconvex finite-sum optimization, our
algorithm also outperforms the best known algorithms in a certain regime.
|
When Not to Classify: Anomaly Detection of Attacks (ADA) on DNN
Classifiers at Test Time | A significant threat to the recent, wide deployment of machine learning-based
systems, including deep neural networks (DNNs), is adversarial learning
attacks. We analyze possible test-time evasion-attack mechanisms and show that,
in some important cases, when the image has been attacked, correctly
classifying it has no utility: i) when the image to be attacked is (even
arbitrarily) selected from the attacker's cache; ii) when the sole recipient of
the classifier's decision is the attacker. Moreover, in some application
domains and scenarios it is highly actionable to detect the attack irrespective
of correctly classifying in the face of it (with classification still performed
if no attack is detected). We hypothesize that, even if human-imperceptible,
adversarial perturbations are machine-detectable. We propose a purely
unsupervised anomaly detector (AD) that, unlike previous works: i) models the
joint density of a deep layer using highly suitable null hypothesis density
models (matched in particular to the non- negative support for RELU layers);
ii) exploits multiple DNN layers; iii) leverages a "source" and "destination"
class concept, source class uncertainty, the class confusion matrix, and DNN
weight information in constructing a novel decision statistic grounded in the
Kullback-Leibler divergence. Tested on MNIST and CIFAR-10 image databases under
three prominent attack strategies, our approach outperforms previous detection
methods, achieving strong ROC AUC detection accuracy on two attacks and better
accuracy than recently reported for a variety of methods on the strongest (CW)
attack. We also evaluate a fully white box attack on our system. Finally, we
evaluate other important performance measures, such as classification accuracy,
versus detection rate and attack strength.
|
Objects that Sound | In this paper our objectives are, first, networks that can embed audio and
visual inputs into a common space that is suitable for cross-modal retrieval;
and second, a network that can localize the object that sounds in an image,
given the audio signal. We achieve both these objectives by training from
unlabelled video using only audio-visual correspondence (AVC) as the objective
function. This is a form of cross-modal self-supervision from video.
To this end, we design new network architectures that can be trained for
cross-modal retrieval and localizing the sound source in an image, by using the
AVC task. We make the following contributions: (i) show that audio and visual
embeddings can be learnt that enable both within-mode (e.g. audio-to-audio) and
between-mode retrieval; (ii) explore various architectures for the AVC task,
including those for the visual stream that ingest a single image, or multiple
images, or a single image and multi-frame optical flow; (iii) show that the
semantic object that sounds within an image can be localized (using only the
sound, no motion or flow information); and (iv) give a cautionary tale on how
to avoid undesirable shortcuts in the data preparation.
|
MEBoost: Mixing Estimators with Boosting for Imbalanced Data
Classification | Class imbalance problem has been a challenging research problem in the fields
of machine learning and data mining as most real life datasets are imbalanced.
Several existing machine learning algorithms try to maximize the accuracy
classification by correctly identifying majority class samples while ignoring
the minority class. However, the concept of the minority class instances
usually represents a higher interest than the majority class. Recently, several
cost sensitive methods, ensemble models and sampling techniques have been used
in literature in order to classify imbalance datasets. In this paper, we
propose MEBoost, a new boosting algorithm for imbalanced datasets. MEBoost
mixes two different weak learners with boosting to improve the performance on
imbalanced datasets. MEBoost is an alternative to the existing techniques such
as SMOTEBoost, RUSBoost, Adaboost, etc. The performance of MEBoost has been
evaluated on 12 benchmark imbalanced datasets with state of the art ensemble
methods like SMOTEBoost, RUSBoost, Easy Ensemble, EUSBoost, DataBoost.
Experimental results show significant improvement over the other methods and it
can be concluded that MEBoost is an effective and promising algorithm to deal
with imbalance datasets. The python version of the code is available here:
https://github.com/farshidrayhanuiu/
|
Synthesizing Novel Pairs of Image and Text | Generating novel pairs of image and text is a problem that combines computer
vision and natural language processing. In this paper, we present strategies
for generating novel image and caption pairs based on existing captioning
datasets. The model takes advantage of recent advances in generative
adversarial networks and sequence-to-sequence modeling. We make generalizations
to generate paired samples from multiple domains. Furthermore, we study cycles
-- generating from image to text then back to image and vise versa, as well as
its connection with autoencoders.
|
Accurate Inference for Adaptive Linear Models | Estimators computed from adaptively collected data do not behave like their
non-adaptive brethren. Rather, the sequential dependence of the collection
policy can lead to severe distributional biases that persist even in the
infinite data limit. We develop a general method -- $\mathbf{W}$-decorrelation
-- for transforming the bias of adaptive linear regression estimators into
variance. The method uses only coarse-grained information about the data
collection policy and does not need access to propensity scores or exact
knowledge of the policy. We bound the finite-sample bias and variance of the
$\mathbf{W}$-estimator and develop asymptotically correct confidence intervals
based on a novel martingale central limit theorem. We then demonstrate the
empirical benefits of the generic $\mathbf{W}$-decorrelation procedure in two
different adaptive data settings: the multi-armed bandit and the autoregressive
time series.
|
HotFlip: White-Box Adversarial Examples for Text Classification | We propose an efficient method to generate white-box adversarial examples to
trick a character-level neural classifier. We find that only a few
manipulations are needed to greatly decrease the accuracy. Our method relies on
an atomic flip operation, which swaps one token for another, based on the
gradients of the one-hot input vectors. Due to efficiency of our method, we can
perform adversarial training which makes the model more robust to attacks at
test time. With the use of a few semantics-preserving constraints, we
demonstrate that HotFlip can be adapted to attack a word-level classifier as
well.
|
Two-dimensional Anti-jamming Mobile Communication Based on Reinforcement
Learning | By using smart radio devices, a jammer can dynamically change its jamming
policy based on opposing security mechanisms; it can even induce the mobile
device to enter a specific communication mode and then launch the jamming
policy accordingly. On the other hand, mobile devices can exploit spread
spectrum and user mobility to address both jamming and interference. In this
paper, a two-dimensional anti-jamming mobile communication scheme is proposed
in which a mobile device leaves a heavily jammed/interfered-with frequency or
area. It is shown that, by applying reinforcement learning techniques, a mobile
device can achieve an optimal communication policy without the need to know the
jamming and interference model and the radio channel model in a dynamic game
framework. More specifically, a hotbooting deep Q-network based two-dimensional
mobile communication scheme is proposed that exploits experiences in similar
scenarios to reduce the exploration time at the beginning of the game, and
applies deep convolutional neural network and macro-action techniques to
accelerate the learning speed in dynamic situations. Several real-world
scenarios are simulated to evaluate the proposed method. These simulation
results show that our proposed scheme can improve both the
signal-to-interference-plus-noise ratio of the signals and the utility of the
mobile devices against cooperative jamming compared with benchmark schemes.
|
End-to-end weakly-supervised semantic alignment | We tackle the task of semantic alignment where the goal is to compute dense
semantic correspondence aligning two images depicting objects of the same
category. This is a challenging task due to large intra-class variation,
changes in viewpoint and background clutter. We present the following three
principal contributions. First, we develop a convolutional neural network
architecture for semantic alignment that is trainable in an end-to-end manner
from weak image-level supervision in the form of matching image pairs. The
outcome is that parameters are learnt from rich appearance variation present in
different but semantically related images without the need for tedious manual
annotation of correspondences at training time. Second, the main component of
this architecture is a differentiable soft inlier scoring module, inspired by
the RANSAC inlier scoring procedure, that computes the quality of the alignment
based on only geometrically consistent correspondences thereby reducing the
effect of background clutter. Third, we demonstrate that the proposed approach
achieves state-of-the-art performance on multiple standard benchmarks for
semantic alignment.
|
Pattern recognition techniques for Boson Sampling validation | The difficulty of validating large-scale quantum devices, such as Boson
Samplers, poses a major challenge for any research program that aims to show
quantum advantages over classical hardware. To address this problem, we propose
a novel data-driven approach wherein models are trained to identify common
pathologies using unsupervised machine learning methods. We illustrate this
idea by training a classifier that exploits K-means clustering to distinguish
between Boson Samplers that use indistinguishable photons from those that do
not. We train the model on numerical simulations of small-scale Boson Samplers
and then validate the pattern recognition technique on larger numerical
simulations as well as on photonic chips in both traditional Boson Sampling and
scattershot experiments. The effectiveness of such method relies on
particle-type-dependent internal correlations present in the output
distributions. This approach performs substantially better on the test data
than previous methods and underscores the ability to further generalize its
operation beyond the scope of the examples that it was trained on.
|
Safe Policy Improvement with Baseline Bootstrapping | This paper considers Safe Policy Improvement (SPI) in Batch Reinforcement
Learning (Batch RL): from a fixed dataset and without direct access to the true
environment, train a policy that is guaranteed to perform at least as well as
the baseline policy used to collect the data. Our approach, called SPI with
Baseline Bootstrapping (SPIBB), is inspired by the knows-what-it-knows
paradigm: it bootstraps the trained policy with the baseline when the
uncertainty is high. Our first algorithm, $\Pi_b$-SPIBB, comes with SPI
theoretical guarantees. We also implement a variant, $\Pi_{\leq b}$-SPIBB, that
is even more efficient in practice. We apply our algorithms to a motivational
stochastic gridworld domain and further demonstrate on randomly generated MDPs
the superiority of SPIBB with respect to existing algorithms, not only in
safety but also in mean performance. Finally, we implement a model-free version
of SPIBB and show its benefits on a navigation task with deep RL implementation
called SPIBB-DQN, which is, to the best of our knowledge, the first RL
algorithm relying on a neural network representation able to train efficiently
and reliably from batch data, without any interaction with the environment.
|
Bilinear residual Neural Network for the identification and forecasting
of dynamical systems | Due to the increasing availability of large-scale observation and simulation
datasets, data-driven representations arise as efficient and relevant
computation representations of dynamical systems for a wide range of
applications, where model-driven models based on ordinary differential equation
remain the state-of-the-art approaches. In this work, we investigate neural
networks (NN) as physically-sound data-driven representations of such systems.
Reinterpreting Runge-Kutta methods as graphical models, we consider a residual
NN architecture and introduce bilinear layers to embed non-linearities which
are intrinsic features of dynamical systems. From numerical experiments for
classic dynamical systems, we demonstrate the relevance of the proposed
NN-based architecture both in terms of forecasting performance and model
identification.
|
Privacy-Preserving Adversarial Networks | We propose a data-driven framework for optimizing privacy-preserving data
release mechanisms to attain the information-theoretically optimal tradeoff
between minimizing distortion of useful data and concealing specific sensitive
information. Our approach employs adversarially-trained neural networks to
implement randomized mechanisms and to perform a variational approximation of
mutual information privacy. We validate our Privacy-Preserving Adversarial
Networks (PPAN) framework via proof-of-concept experiments on discrete and
continuous synthetic data, as well as the MNIST handwritten digits dataset. For
synthetic data, our model-agnostic PPAN approach achieves tradeoff points very
close to the optimal tradeoffs that are analytically-derived from model
knowledge. In experiments with the MNIST data, we visually demonstrate a
learned tradeoff between minimizing the pixel-level distortion versus
concealing the written digit.
|
Snake: a Stochastic Proximal Gradient Algorithm for Regularized Problems
over Large Graphs | A regularized optimization problem over a large unstructured graph is
studied, where the regularization term is tied to the graph geometry. Typical
regularization examples include the total variation and the Laplacian
regularizations over the graph. When applying the proximal gradient algorithm
to solve this problem, there exist quite affordable methods to implement the
proximity operator (backward step) in the special case where the graph is a
simple path without loops. In this paper, an algorithm, referred to as "Snake",
is proposed to solve such regularized problems over general graphs, by taking
benefit of these fast methods. The algorithm consists in properly selecting
random simple paths in the graph and performing the proximal gradient algorithm
over these simple paths. This algorithm is an instance of a new general
stochastic proximal gradient algorithm, whose convergence is proven.
Applications to trend filtering and graph inpainting are provided among others.
Numerical experiments are conducted over large graphs.
|
Development and evaluation of a deep learning model for protein-ligand
binding affinity prediction | Structure based ligand discovery is one of the most successful approaches for
augmenting the drug discovery process. Currently, there is a notable shift
towards machine learning (ML) methodologies to aid such procedures. Deep
learning has recently gained considerable attention as it allows the model to
"learn" to extract features that are relevant for the task at hand. We have
developed a novel deep neural network estimating the binding affinity of
ligand-receptor complexes. The complex is represented with a 3D grid, and the
model utilizes a 3D convolution to produce a feature map of this
representation, treating the atoms of both proteins and ligands in the same
manner. Our network was tested on the CASF "scoring power" benchmark and Astex
Diverse Set and outperformed classical scoring functions. The model, together
with usage instructions and examples, is available as a git repository at
http://gitlab.com/cheminfIBB/pafnucy
|
On Data-Dependent Random Features for Improved Generalization in
Supervised Learning | The randomized-feature approach has been successfully employed in large-scale
kernel approximation and supervised learning. The distribution from which the
random features are drawn impacts the number of features required to
efficiently perform a learning task. Recently, it has been shown that employing
data-dependent randomization improves the performance in terms of the required
number of random features. In this paper, we are concerned with the
randomized-feature approach in supervised learning for good generalizability.
We propose the Energy-based Exploration of Random Features (EERF) algorithm
based on a data-dependent score function that explores the set of possible
features and exploits the promising regions. We prove that the proposed score
function with high probability recovers the spectrum of the best fit within the
model class. Our empirical results on several benchmark datasets further verify
that our method requires smaller number of random features to achieve a certain
generalization error compared to the state-of-the-art while introducing
negligible pre-processing overhead. EERF can be implemented in a few lines of
code and requires no additional tuning parameters.
|
Exploring High-Dimensional Structure via Axis-Aligned Decomposition of
Linear Projections | Two-dimensional embeddings remain the dominant approach to visualize high
dimensional data. The choice of embeddings ranges from highly non-linear ones,
which can capture complex relationships but are difficult to interpret
quantitatively, to axis-aligned projections, which are easy to interpret but
are limited to bivariate relationships. Linear project can be considered as a
compromise between complexity and interpretability, as they allow explicit axes
labels, yet provide significantly more degrees of freedom compared to
axis-aligned projections. Nevertheless, interpreting the axes directions, which
are linear combinations often with many non-trivial components, remains
difficult. To address this problem we introduce a structure aware decomposition
of (multiple) linear projections into sparse sets of axis aligned projections,
which jointly capture all information of the original linear ones. In
particular, we use tools from Dempster-Shafer theory to formally define how
relevant a given axis aligned project is to explain the neighborhood relations
displayed in some linear projection. Furthermore, we introduce a new approach
to discover a diverse set of high quality linear projections and show that in
practice the information of $k$ linear projections is often jointly encoded in
$\sim k$ axis aligned plots. We have integrated these ideas into an interactive
visualization system that allows users to jointly browse both linear
projections and their axis aligned representatives. Using a number of case
studies we show how the resulting plots lead to more intuitive visualizations
and new insight.
|
Adversarial Examples: Attacks and Defenses for Deep Learning | With rapid progress and significant successes in a wide spectrum of
applications, deep learning is being applied in many safety-critical
environments. However, deep neural networks have been recently found vulnerable
to well-designed input samples, called adversarial examples. Adversarial
examples are imperceptible to human but can easily fool deep neural networks in
the testing/deploying stage. The vulnerability to adversarial examples becomes
one of the major risks for applying deep neural networks in safety-critical
environments. Therefore, attacks and defenses on adversarial examples draw
great attention. In this paper, we review recent findings on adversarial
examples for deep neural networks, summarize the methods for generating
adversarial examples, and propose a taxonomy of these methods. Under the
taxonomy, applications for adversarial examples are investigated. We further
elaborate on countermeasures for adversarial examples and explore the
challenges and the potential solutions.
|
Query-Efficient Black-box Adversarial Examples (superceded) | Note that this paper is superceded by "Black-Box Adversarial Attacks with
Limited Queries and Information."
Current neural network-based image classifiers are susceptible to adversarial
examples, even in the black-box setting, where the attacker is limited to query
access without access to gradients. Previous methods --- substitute networks
and coordinate-based finite-difference methods --- are either unreliable or
query-inefficient, making these methods impractical for certain problems.
We introduce a new method for reliably generating adversarial examples under
more restricted, practical black-box threat models. First, we apply natural
evolution strategies to perform black-box attacks using two to three orders of
magnitude fewer queries than previous methods. Second, we introduce a new
algorithm to perform targeted adversarial attacks in the partial-information
setting, where the attacker only has access to a limited number of target
classes. Using these techniques, we successfully perform the first targeted
adversarial attack against a commercially deployed machine learning system, the
Google Cloud Vision API, in the partial information setting.
|
Approximate Profile Maximum Likelihood | We propose an efficient algorithm for approximate computation of the profile
maximum likelihood (PML), a variant of maximum likelihood maximizing the
probability of observing a sufficient statistic rather than the empirical
sample. The PML has appealing theoretical properties, but is difficult to
compute exactly. Inspired by observations gleaned from exactly solvable cases,
we look for an approximate PML solution, which, intuitively, clumps comparably
frequent symbols into one symbol. This amounts to lower-bounding a certain
matrix permanent by summing over a subgroup of the symmetric group rather than
the whole group during the computation. We extensively experiment with the
approximate solution, and find the empirical performance of our approach is
competitive and sometimes significantly better than state-of-the-art
performance for various estimation problems.
|
On Wasserstein Reinforcement Learning and the Fokker-Planck equation | Policy gradients methods often achieve better performance when the change in
policy is limited to a small Kullback-Leibler divergence. We derive policy
gradients where the change in policy is limited to a small Wasserstein distance
(or trust region). This is done in the discrete and continuous multi-armed
bandit settings with entropy regularisation. We show that in the small steps
limit with respect to the Wasserstein distance $W_2$, policy dynamics are
governed by the Fokker-Planck (heat) equation, following the
Jordan-Kinderlehrer-Otto result. This means that policies undergo diffusion and
advection, concentrating near actions with high reward. This helps elucidate
the nature of convergence in the probability matching setup, and provides
justification for empirical practices such as Gaussian policy priors and
additive gradient noise.
|
Calibrating Noise to Variance in Adaptive Data Analysis | Datasets are often used multiple times and each successive analysis may
depend on the outcome of previous analyses. Standard techniques for ensuring
generalization and statistical validity do not account for this adaptive
dependence. A recent line of work studies the challenges that arise from such
adaptive data reuse by considering the problem of answering a sequence of
"queries" about the data distribution where each query may depend arbitrarily
on answers to previous queries.
The strongest results obtained for this problem rely on differential privacy
-- a strong notion of algorithmic stability with the important property that it
"composes" well when data is reused. However the notion is rather strict, as it
requires stability under replacement of an arbitrary data element. The simplest
algorithm is to add Gaussian (or Laplace) noise to distort the empirical
answers. However, analysing this technique using differential privacy yields
suboptimal accuracy guarantees when the queries have low variance. Here we
propose a relaxed notion of stability that also composes adaptively. We
demonstrate that a simple and natural algorithm based on adding noise scaled to
the standard deviation of the query provides our notion of stability. This
implies an algorithm that can answer statistical queries about the dataset with
substantially improved accuracy guarantees for low-variance queries. The only
previous approach that provides such accuracy guarantees is based on a more
involved differentially private median-of-means algorithm and its analysis
exploits stronger "group" stability of the algorithm.
|
Discovery of Shifting Patterns in Sequence Classification | In this paper, we investigate the multi-variate sequence classification
problem from a multi-instance learning perspective. Real-world sequential data
commonly show discriminative patterns only at specific time periods. For
instance, we can identify a cropland during its growing season, but it looks
similar to a barren land after harvest or before planting. Besides, even within
the same class, the discriminative patterns can appear in different periods of
sequential data. Due to such property, these discriminative patterns are also
referred to as shifting patterns. The shifting patterns in sequential data
severely degrade the performance of traditional classification methods without
sufficient training data.
We propose a novel sequence classification method by automatically mining
shifting patterns from multi-variate sequence. The method employs a
multi-instance learning approach to detect shifting patterns while also
modeling temporal relationships within each multi-instance bag by an LSTM model
to further improve the classification performance. We extensively evaluate our
method on two real-world applications - cropland mapping and affective state
recognition. The experiments demonstrate the superiority of our proposed method
in sequence classification performance and in detecting discriminative shifting
patterns.
|
Fusing Multifaceted Transaction Data for User Modeling and Demographic
Prediction | Inferring user characteristics such as demographic attributes is of the
utmost importance in many user-centric applications. Demographic data is an
enabler of personalization, identity security, and other applications. Despite
that, this data is sensitive and often hard to obtain. Previous work has shown
that purchase history can be used for multi-task prediction of many demographic
fields such as gender and marital status. Here we present an embedding based
method to integrate multifaceted sequences of transaction data, together with
auxiliary relational tables, for better user modeling and demographic
prediction.
|
Hyperparameters Optimization in Deep Convolutional Neural Network /
Bayesian Approach with Gaussian Process Prior | Convolutional Neural Network is known as ConvNet have been extensively used
in many complex machine learning tasks. However, hyperparameters optimization
is one of a crucial step in developing ConvNet architectures, since the
accuracy and performance are reliant on the hyperparameters. This multilayered
architecture parameterized by a set of hyperparameters such as the number of
convolutional layers, number of fully connected dense layers & neurons, the
probability of dropout implementation, learning rate. Hence the searching the
hyperparameter over the hyperparameter space are highly difficult to build such
complex hierarchical architecture. Many methods have been proposed over the
decade to explore the hyperparameter space and find the optimum set of
hyperparameter values. Reportedly, Gird search and Random search are said to be
inefficient and extremely expensive, due to a large number of hyperparameters
of the architecture. Hence, Sequential model-based Bayesian Optimization is a
promising alternative technique to address the extreme of the unknown cost
function. The recent study on Bayesian Optimization by Snoek in nine
convolutional network parameters is achieved the lowerest error report in the
CIFAR-10 benchmark. This article is intended to provide the overview of the
mathematical concept behind the Bayesian Optimization over a Gaussian prior.
|
Linear Time Clustering for High Dimensional Mixtures of Gaussian Clouds | Clustering mixtures of Gaussian distributions is a fundamental and
challenging problem that is ubiquitous in various high-dimensional data
processing tasks. While state-of-the-art work on learning Gaussian mixture
models has focused primarily on improving separation bounds and their
generalization to arbitrary classes of mixture models, less emphasis has been
paid to practical computational efficiency of the proposed solutions. In this
paper, we propose a novel and highly efficient clustering algorithm for $n$
points drawn from a mixture of two arbitrary Gaussian distributions in
$\mathbb{R}^p$. The algorithm involves performing random 1-dimensional
projections until a direction is found that yields a user-specified clustering
error $e$. For a 1-dimensional separation parameter $\gamma$ satisfying
$\gamma=Q^{-1}(e)$, the expected number of such projections is shown to be
bounded by $o(\ln p)$, when $\gamma$ satisfies $\gamma\leq
c\sqrt{\ln{\ln{p}}}$, with $c$ as the separability parameter of the two
Gaussians in $\mathbb{R}^p$. Consequently, the expected overall running time of
the algorithm is linear in $n$ and quasi-linear in $p$ at $o(\ln{p})O(np)$, and
the sample complexity is independent of $p$. This result stands in contrast to
prior works which provide polynomial, with at-best quadratic, running time in
$p$ and $n$. We show that our bound on the expected number of 1-dimensional
projections extends to the case of three or more Gaussian components, and we
present a generalization of our results to mixture distributions beyond the
Gaussian model.
|
Probabilistic Learning of Torque Controllers from Kinematic and Force
Constraints | When learning skills from demonstrations, one is often required to think in
advance about the appropriate task representation (usually in either
operational or configuration space). We here propose a probabilistic approach
for simultaneously learning and synthesizing torque control commands which take
into account task space, joint space and force constraints. We treat the
problem by considering different torque controllers acting on the robot, whose
relevance is learned probabilistically from demonstrations. This information is
used to combine the controllers by exploiting the properties of Gaussian
distributions, generating new torque commands that satisfy the important
features of the task. We validate the approach in two experimental scenarios
using 7-DoF torquecontrolled manipulators, with tasks that require the
consideration of different controllers to be properly executed.
|
Block-diagonal Hessian-free Optimization for Training Neural Networks | Second-order methods for neural network optimization have several advantages
over methods based on first-order gradient descent, including better scaling to
large mini-batch sizes and fewer updates needed for convergence. But they are
rarely applied to deep learning in practice because of high computational cost
and the need for model-dependent algorithmic variations. We introduce a variant
of the Hessian-free method that leverages a block-diagonal approximation of the
generalized Gauss-Newton matrix. Our method computes the curvature
approximation matrix only for pairs of parameters from the same layer or block
of the neural network and performs conjugate gradient updates independently for
each block. Experiments on deep autoencoders, deep convolutional networks, and
multilayer LSTMs demonstrate better convergence and generalization compared to
the original Hessian-free approach and the Adam method.
|
A Flexible Approach to Automated RNN Architecture Generation | The process of designing neural architectures requires expert knowledge and
extensive trial and error. While automated architecture search may simplify
these requirements, the recurrent neural network (RNN) architectures generated
by existing methods are limited in both flexibility and components. We propose
a domain-specific language (DSL) for use in automated architecture search which
can produce novel RNNs of arbitrary depth and width. The DSL is flexible enough
to define standard architectures such as the Gated Recurrent Unit and Long
Short Term Memory and allows the introduction of non-standard RNN components
such as trigonometric curves and layer normalization. Using two different
candidate generation techniques, random search with a ranking function and
reinforcement learning, we explore the novel architectures produced by the RNN
DSL for language modeling and machine translation domains. The resulting
architectures do not follow human intuition yet perform well on their targeted
tasks, suggesting the space of usable RNN architectures is far larger than
previously assumed.
|
Adversarial Structured Prediction for Multivariate Measures | Many predicted structured objects (e.g., sequences, matchings, trees) are
evaluated using the F-score, alignment error rate (AER), or other multivariate
performance measures. Since inductively optimizing these measures using
training data is typically computationally difficult, empirical risk
minimization of surrogate losses is employed, using, e.g., the hinge loss for
(structured) support vector machines. These approximations often introduce a
mismatch between the learner's objective and the desired application
performance, leading to inconsistency. We take a different approach:
adversarially approximate training data while optimizing the exact F-score or
AER. Structured predictions under this formulation result from solving zero-sum
games between a predictor seeking the best performance and an adversary seeking
the worst while required to (approximately) match certain structured properties
of the training data. We explore this approach for word alignment (AER
evaluation) and named entity recognition (F-score evaluation) with linear-chain
constraints.
|
Finding Competitive Network Architectures Within a Day Using UCT | The design of neural network architectures for a new data set is a laborious
task which requires human deep learning expertise. In order to make deep
learning available for a broader audience, automated methods for finding a
neural network architecture are vital. Recently proposed methods can already
achieve human expert level performances. However, these methods have run times
of months or even years of GPU computing time, ignoring hardware constraints as
faced by many researchers and companies. We propose the use of Monte Carlo
planning in combination with two different UCT (upper confidence bound applied
to trees) derivations to search for network architectures. We adapt the UCT
algorithm to the needs of network architecture search by proposing two ways of
sharing information between different branches of the search tree. In an
empirical study we are able to demonstrate that this method is able to find
competitive networks for MNIST, SVHN and CIFAR-10 in just a single GPU day.
Extending the search time to five GPU days, we are able to outperform human
architectures and our competitors which consider the same types of layers.
|
ADINE: An Adaptive Momentum Method for Stochastic Gradient Descent | Two major momentum-based techniques that have achieved tremendous success in
optimization are Polyak's heavy ball method and Nesterov's accelerated
gradient. A crucial step in all momentum-based methods is the choice of the
momentum parameter $m$ which is always suggested to be set to less than $1$.
Although the choice of $m < 1$ is justified only under very strong theoretical
assumptions, it works well in practice even when the assumptions do not
necessarily hold. In this paper, we propose a new momentum based method
$\textit{ADINE}$, which relaxes the constraint of $m < 1$ and allows the
learning algorithm to use adaptive higher momentum. We motivate our hypothesis
on $m$ by experimentally verifying that a higher momentum ($\ge 1$) can help
escape saddles much faster. Using this motivation, we propose our method
$\textit{ADINE}$ that helps weigh the previous updates more (by setting the
momentum parameter $> 1$), evaluate our proposed algorithm on deep neural
networks and show that $\textit{ADINE}$ helps the learning algorithm to
converge much faster without compromising on the generalization error.
|
In silico generation of novel, drug-like chemical matter using the LSTM
neural network | The exploration of novel chemical spaces is one of the most important tasks
of cheminformatics when supporting the drug discovery process. Properly
designed and trained deep neural networks can provide a viable alternative to
brute-force de novo approaches or various other machine-learning techniques for
generating novel drug-like molecules. In this article we present a method to
generate molecules using a long short-term memory (LSTM) neural network and
provide an analysis of the results, including a virtual screening test. Using
the network one million drug-like molecules were generated in 2 hours. The
molecules are novel, diverse (contain numerous novel chemotypes), have good
physicochemical properties and have good synthetic accessibility, even though
these qualities were not specific constraints. Although novel, their structural
features and functional groups remain closely within the drug-like space
defined by the bioactive molecules from ChEMBL. Virtual screening using the
profile QSAR approach confirms that the potential of these novel molecules to
show bioactivity is comparable to the ChEMBL set from which they were derived.
The molecule generator written in Python used in this study is available on
request.
|
Fast kNN mode seeking clustering applied to active learning | A significantly faster algorithm is presented for the original kNN mode
seeking procedure. It has the advantages over the well-known mean shift
algorithm that it is feasible in high-dimensional vector spaces and results in
uniquely, well defined modes. Moreover, without any additional computational
effort it may yield a multi-scale hierarchy of clusterings. The time complexity
is just O(n^1.5). resulting computing times range from seconds for 10^4 objects
to minutes for 10^5 objects and to less than an hour for 10^6 objects. The
space complexity is just O(n). The procedure is well suited for finding large
sets of small clusters and is thereby a candidate to analyze thousands of
clusters in millions of objects.
The kNN mode seeking procedure can be used for active learning by assigning
the clusters to the class of the modal objects of the clusters. Its feasibility
is shown by some examples with up to 1.5 million handwritten digits. The
obtained classification results based on the clusterings are compared with
those obtained by the nearest neighbor rule and the support vector classifier
based on the same labeled objects for training. It can be concluded that using
the clustering structure for classification can be significantly better than
using the trained classifiers. A drawback of using the clustering for
classification, however, is that no classifier is obtained that may be used for
out-of-sample objects.
|
Differentially Private Distributed Learning for Language Modeling Tasks | One of the big challenges in machine learning applications is that training
data can be different from the real-world data faced by the algorithm. In
language modeling, users' language (e.g. in private messaging) could change in
a year and be completely different from what we observe in publicly available
data. At the same time, public data can be used for obtaining general knowledge
(i.e. general model of English). We study approaches to distributed fine-tuning
of a general model on user private data with the additional requirements of
maintaining the quality on the general data and minimization of communication
costs. We propose a novel technique that significantly improves prediction
quality on users' language compared to a general model and outperforms gradient
compression methods in terms of communication efficiency. The proposed
procedure is fast and leads to an almost 70% perplexity reduction and 8.7
percentage point improvement in keystroke saving rate on informal English
texts. We also show that the range of tasks our approach is applicable to is
not limited by language modeling only. Finally, we propose an experimental
framework for evaluating differential privacy of distributed training of
language models and show that our approach has good privacy guarantees.
|
A Distributed Frank-Wolfe Framework for Learning Low-Rank Matrices with
the Trace Norm | We consider the problem of learning a high-dimensional but low-rank matrix
from a large-scale dataset distributed over several machines, where
low-rankness is enforced by a convex trace norm constraint. We propose
DFW-Trace, a distributed Frank-Wolfe algorithm which leverages the low-rank
structure of its updates to achieve efficiency in time, memory and
communication usage. The step at the heart of DFW-Trace is solved approximately
using a distributed version of the power method. We provide a theoretical
analysis of the convergence of DFW-Trace, showing that we can ensure sublinear
convergence in expectation to an optimal solution with few power iterations per
epoch. We implement DFW-Trace in the Apache Spark distributed programming
framework and validate the usefulness of our approach on synthetic and real
data, including the ImageNet dataset with high-dimensional features extracted
from a deep neural network.
|
Use of Deep Learning in Modern Recommendation System: A Summary of
Recent Works | With the exponential increase in the amount of digital information over the
internet, online shops, online music, video and image libraries, search engines
and recommendation system have become the most convenient ways to find relevant
information within a short time. In the recent times, deep learning's advances
have gained significant attention in the field of speech recognition, image
processing and natural language processing. Meanwhile, several recent studies
have shown the utility of deep learning in the area of recommendation systems
and information retrieval as well. In this short review, we cover the recent
advances made in the field of recommendation using various variants of deep
learning technology. We organize the review in three parts: Collaborative
system, Content based system and Hybrid system. The review also discusses the
contribution of deep learning integrated recommendation systems into several
application domains. The review concludes by discussion of the impact of deep
learning in recommendation system in various domain and whether deep learning
has shown any significant improvement over the conventional systems for
recommendation. Finally, we also provide future directions of research which
are possible based on the current state of use of deep learning in
recommendation systems.
|
Differentially Private Federated Learning: A Client Level Perspective | Federated learning is a recent advance in privacy protection. In this
context, a trusted curator aggregates parameters optimized in decentralized
fashion by multiple clients. The resulting model is then distributed back to
all clients, ultimately converging to a joint representative model without
explicitly having to share the data. However, the protocol is vulnerable to
differential attacks, which could originate from any party contributing during
federated optimization. In such an attack, a client's contribution during
training and information about their data set is revealed through analyzing the
distributed model. We tackle this problem and propose an algorithm for client
sided differential privacy preserving federated optimization. The aim is to
hide clients' contributions during training, balancing the trade-off between
privacy loss and model performance. Empirical studies suggest that given a
sufficiently large number of participating clients, our proposed procedure can
maintain client-level differential privacy at only a minor cost in model
performance.
|
Experimental Phase Estimation Enhanced By Machine Learning | Phase estimation protocols provide a fundamental benchmark for the field of
quantum metrology. The latter represents one of the most relevant applications
of quantum theory, potentially enabling the capability of measuring unknown
physical parameters with improved precision over classical strategies. Within
this context, most theoretical and experimental studies have focused on
determining the fundamental bounds and how to achieve them in the asymptotic
regime where a large number of resources is employed. However, in most
applications it is necessary to achieve optimal precisions by performing only a
limited number of measurements. To this end, machine learning techniques can be
applied as a powerful optimization tool. Here, we implement experimentally
single-photon adaptive phase estimation protocols enhanced by machine learning,
showing the capability of reaching optimal precision after a small number of
trials. In particular, we introduce a new approach for Bayesian estimation that
exhibit best performances for very low number of photons N. Furthermore, we
study the resilience to noise of the tested methods, showing that the optimized
Bayesian approach is very robust in the presence of imperfections. Application
of this methodology can be envisaged in the more general multiparameter case,
that represents a paradigmatic scenario for several tasks including imaging or
Hamiltonian learning.
|
Riemann-Theta Boltzmann Machine | A general Boltzmann machine with continuous visible and discrete integer
valued hidden states is introduced. Under mild assumptions about the connection
matrices, the probability density function of the visible units can be solved
for analytically, yielding a novel parametric density function involving a
ratio of Riemann-Theta functions. The conditional expectation of a hidden state
for given visible states can also be calculated analytically, yielding a
derivative of the logarithmic Riemann-Theta function. The conditional
expectation can be used as activation function in a feedforward neural network,
thereby increasing the modelling capacity of the network. Both the Boltzmann
machine and the derived feedforward neural network can be successfully trained
via standard gradient- and non-gradient-based optimization techniques.
|
Improving Generalization Performance by Switching from Adam to SGD | Despite superior training outcomes, adaptive optimization methods such as
Adam, Adagrad or RMSprop have been found to generalize poorly compared to
Stochastic gradient descent (SGD). These methods tend to perform well in the
initial portion of training but are outperformed by SGD at later stages of
training. We investigate a hybrid strategy that begins training with an
adaptive method and switches to SGD when appropriate. Concretely, we propose
SWATS, a simple strategy which switches from Adam to SGD when a triggering
condition is satisfied. The condition we propose relates to the projection of
Adam steps on the gradient subspace. By design, the monitoring process for this
condition adds very little overhead and does not increase the number of
hyperparameters in the optimizer. We report experiments on several standard
benchmarks such as: ResNet, SENet, DenseNet and PyramidNet for the CIFAR-10 and
CIFAR-100 data sets, ResNet on the tiny-ImageNet data set and language modeling
with recurrent networks on the PTB and WT2 data sets. The results show that our
strategy is capable of closing the generalization gap between SGD and Adam on a
majority of the tasks.
|
Deep Learning with Lung Segmentation and Bone Shadow Exclusion
Techniques for Chest X-Ray Analysis of Lung Cancer | The recent progress of computing, machine learning, and especially deep
learning, for image recognition brings a meaningful effect for automatic
detection of various diseases from chest X-ray images (CXRs). Here efficiency
of lung segmentation and bone shadow exclusion techniques is demonstrated for
analysis of 2D CXRs by deep learning approach to help radiologists identify
suspicious lesions and nodules in lung cancer patients. Training and validation
was performed on the original JSRT dataset (dataset #01), BSE-JSRT dataset,
i.e. the same JSRT dataset, but without clavicle and rib shadows (dataset #02),
original JSRT dataset after segmentation (dataset #03), and BSE-JSRT dataset
after segmentation (dataset #04). The results demonstrate the high efficiency
and usefulness of the considered pre-processing techniques in the simplified
configuration even. The pre-processed dataset without bones (dataset #02)
demonstrates the much better accuracy and loss results in comparison to the
other pre-processed datasets after lung segmentation (datasets #02 and #03).
|
Image Segmentation to Distinguish Between Overlapping Human Chromosomes | In medicine, visualizing chromosomes is important for medical diagnostics,
drug development, and biomedical research. Unfortunately, chromosomes often
overlap and it is necessary to identify and distinguish between the overlapping
chromosomes. A segmentation solution that is fast and automated will enable
scaling of cost effective medicine and biomedical research. We apply neural
network-based image segmentation to the problem of distinguishing between
partially overlapping DNA chromosomes. A convolutional neural network is
customized for this problem. The results achieved intersection over union (IOU)
scores of 94.7% for the overlapping region and 88-94% on the non-overlapping
chromosome regions.
|
Sim2Real View Invariant Visual Servoing by Recurrent Control | Humans are remarkably proficient at controlling their limbs and tools from a
wide range of viewpoints and angles, even in the presence of optical
distortions. In robotics, this ability is referred to as visual servoing:
moving a tool or end-point to a desired location using primarily visual
feedback. In this paper, we study how viewpoint-invariant visual servoing
skills can be learned automatically in a robotic manipulation scenario. To this
end, we train a deep recurrent controller that can automatically determine
which actions move the end-point of a robotic arm to a desired object. The
problem that must be solved by this controller is fundamentally ambiguous:
under severe variation in viewpoint, it may be impossible to determine the
actions in a single feedforward operation. Instead, our visual servoing system
must use its memory of past movements to understand how the actions affect the
robot motion from the current viewpoint, correcting mistakes and gradually
moving closer to the target. This ability is in stark contrast to most visual
servoing methods, which either assume known dynamics or require a calibration
phase. We show how we can learn this recurrent controller using simulated data
and a reinforcement learning objective. We then describe how the resulting
model can be transferred to a real-world robot by disentangling perception from
control and only adapting the visual layers. The adapted model can servo to
previously unseen objects from novel viewpoints on a real-world Kuka IIWA
robotic arm. For supplementary videos, see:
https://fsadeghi.github.io/Sim2RealViewInvariantServo
|
Deep Unsupervised Clustering Using Mixture of Autoencoders | Unsupervised clustering is one of the most fundamental challenges in machine
learning. A popular hypothesis is that data are generated from a union of
low-dimensional nonlinear manifolds; thus an approach to clustering is
identifying and separating these manifolds. In this paper, we present a novel
approach to solve this problem by using a mixture of autoencoders. Our model
consists of two parts: 1) a collection of autoencoders where each autoencoder
learns the underlying manifold of a group of similar objects, and 2) a mixture
assignment neural network, which takes the concatenated latent vectors from the
autoencoders as input and infers the distribution over clusters. By jointly
optimizing the two parts, we simultaneously assign data to clusters and learn
the underlying manifolds of each cluster.
|
Towards a Deep Improviser: a prototype deep learning post-tonal free
music generator | Two modest-sized symbolic corpora of post-tonal and post-metric keyboard
music have been constructed, one algorithmic, the other improvised. Deep
learning models of each have been trained and largely optimised. Our purpose is
to obtain a model with sufficient generalisation capacity that in response to a
small quantity of separate fresh input seed material, it can generate outputs
that are distinctive, rather than recreative of the learned corpora or the seed
material. This objective has been first assessed statistically, and as judged
by k-sample Anderson-Darling and Cramer tests, has been achieved. Music has
been generated using the approach, and informal judgements place it roughly on
a par with algorithmic and composed music in related forms. Future work will
aim to enhance the model such that it can be evaluated in relation to
expression, meaning and utility in real-time performance.
|
Wolf in Sheep's Clothing - The Downscaling Attack Against Deep Learning
Applications | This paper considers security risks buried in the data processing pipeline in
common deep learning applications. Deep learning models usually assume a fixed
scale for their training and input data. To allow deep learning applications to
handle a wide range of input data, popular frameworks, such as Caffe,
TensorFlow, and Torch, all provide data scaling functions to resize input to
the dimensions used by deep learning models. Image scaling algorithms are
intended to preserve the visual features of an image after scaling. However,
common image scaling algorithms are not designed to handle human crafted
images. Attackers can make the scaling outputs look dramatically different from
the corresponding input images.
This paper presents a downscaling attack that targets the data scaling
process in deep learning applications. By carefully crafting input data that
mismatches with the dimension used by deep learning models, attackers can
create deceiving effects. A deep learning application effectively consumes data
that are not the same as those presented to users. The visual inconsistency
enables practical evasion and data poisoning attacks to deep learning
applications. This paper presents proof-of-concept attack samples to popular
deep-learning-based image classification applications. To address the
downscaling attacks, the paper also suggests multiple potential mitigation
strategies.
|
Multi-dimensional Graph Fourier Transform | Many signals on Cartesian product graphs appear in the real world, such as
digital images, sensor observation time series, and movie ratings on Netflix.
These signals are "multi-dimensional" and have directional characteristics
along each factor graph. However, the existing graph Fourier transform does not
distinguish these directions, and assigns 1-D spectra to signals on product
graphs. Further, these spectra are often multi-valued at some frequencies. Our
main result is a multi-dimensional graph Fourier transform that solves such
problems associated with the conventional GFT. Using algebraic properties of
Cartesian products, the proposed transform rearranges 1-D spectra obtained by
the conventional GFT into the multi-dimensional frequency domain, of which each
dimension represents a directional frequency along each factor graph. Thus, the
multi-dimensional graph Fourier transform enables directional frequency
analysis, in addition to frequency analysis with the conventional GFT.
Moreover, this rearrangement resolves the multi-valuedness of spectra in some
cases. The multi-dimensional graph Fourier transform is a foundation of novel
filterings and stationarities that utilize dimensional information of graph
signals, which are also discussed in this study. The proposed methods are
applicable to a wide variety of data that can be regarded as signals on
Cartesian product graphs. This study also notes that multivariate graph signals
can be regarded as 2-D univariate graph signals. This correspondence provides
natural definitions of the multivariate graph Fourier transform and the
multivariate stationarity based on their 2-D univariate versions.
|
Indoor Sound Source Localization with Probabilistic Neural Network | It is known that adverse environments such as high reverberation and low
signal-to-noise ratio (SNR) pose a great challenge to indoor sound source
localization. To address this challenge, in this paper, we propose a sound
source localization algorithm based on probabilistic neural network, namely
Generalized cross correlation Classification Algorithm (GCA). Experimental
results for adverse environments with high reverberation time T60 up to 600ms
and low SNR such as -10dB show that, the average azimuth angle error and
elevation angle error by GCA are only 4.6 degrees and 3.1 degrees respectively.
Compared with three recently published algorithms, GCA has increased the
success rate on direction of arrival estimation significantly with good
robustness to environmental changes. These results show that the proposed GCA
can localize accurately and robustly for diverse indoor applications where the
site acoustic features can be studied prior to the localization stage.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.