title
stringlengths 7
246
| abstract
stringlengths 6
3.31k
|
|---|---|
Plan, Attend, Generate: Planning for Sequence-to-Sequence Models | We investigate the integration of a planning mechanism into
sequence-to-sequence models using attention. We develop a model which can plan
ahead in the future when it computes its alignments between input and output
sequences, constructing a matrix of proposed future alignments and a commitment
vector that governs whether to follow or recompute the plan. This mechanism is
inspired by the recently proposed strategic attentive reader and writer (STRAW)
model for Reinforcement Learning. Our proposed model is end-to-end trainable
using primarily differentiable operations. We show that it outperforms a strong
baseline on character-level translation tasks from WMT'15, the algorithmic task
of finding Eulerian circuits of graphs, and question generation from the text.
Our analysis demonstrates that the model computes qualitatively intuitive
alignments, converges faster than the baselines, and achieves superior
performance with fewer parameters.
|
Implicit Regularization in Nonconvex Statistical Estimation: Gradient
Descent Converges Linearly for Phase Retrieval, Matrix Completion, and Blind
Deconvolution | Recent years have seen a flurry of activities in designing provably efficient
nonconvex procedures for solving statistical estimation problems. Due to the
highly nonconvex nature of the empirical loss, state-of-the-art procedures
often require proper regularization (e.g. trimming, regularized cost,
projection) in order to guarantee fast convergence. For vanilla procedures such
as gradient descent, however, prior theory either recommends highly
conservative learning rates to avoid overshooting, or completely lacks
performance guarantees.
This paper uncovers a striking phenomenon in nonconvex optimization: even in
the absence of explicit regularization, gradient descent enforces proper
regularization implicitly under various statistical models. In fact, gradient
descent follows a trajectory staying within a basin that enjoys nice geometry,
consisting of points incoherent with the sampling mechanism. This "implicit
regularization" feature allows gradient descent to proceed in a far more
aggressive fashion without overshooting, which in turn results in substantial
computational savings. Focusing on three fundamental statistical estimation
problems, i.e. phase retrieval, low-rank matrix completion, and blind
deconvolution, we establish that gradient descent achieves near-optimal
statistical and computational guarantees without explicit regularization. In
particular, by marrying statistical modeling with generic optimization theory,
we develop a general recipe for analyzing the trajectories of iterative
algorithms via a leave-one-out perturbation argument. As a byproduct, for noisy
matrix completion, we demonstrate that gradient descent achieves near-optimal
error control --- measured entrywise and by the spectral norm --- which might
be of independent interest.
|
ISLAND: In-Silico Prediction of Proteins Binding Affinity Using Sequence
Descriptors | Determination of binding affinity of proteins in the formation of protein
complexes requires sophisticated, expensive and time-consuming experimentation
which can be replaced with computational methods. Most computational prediction
techniques require protein structures which limit their applicability to
protein complexes with known structures. In this work, we explore sequence
based protein binding affinity prediction using machine learning. Our paper
highlights the fact that the generalization performance of even the state of
the art sequence-only predictor of binding affinity is far from satisfactory
and that the development of effective and practical methods in this domain is
still an open problem. We also propose a novel sequence-only predictor of
binding affinity called ISLAND which gives better accuracy than existing
methods over the same validation set as well as on external independent test
dataset. A cloud-based webserver implementation of ISLAND and its Python code
are available at the URL:
http://faculty.pieas.edu.pk/fayyaz/software.html#island.
|
Intent-Aware Contextual Recommendation System | Recommender systems take inputs from user history, use an internal ranking
algorithm to generate results and possibly optimize this ranking based on
feedback. However, often the recommender system is unaware of the actual intent
of the user and simply provides recommendations dynamically without properly
understanding the thought process of the user. An intelligent recommender
system is not only useful for the user but also for businesses which want to
learn the tendencies of their users. Finding out tendencies or intents of a
user is a difficult problem to solve.
Keeping this in mind, we sought out to create an intelligent system which
will keep track of the user's activity on a web-application as well as
determine the intent of the user in each session. We devised a way to encode
the user's activity through the sessions. Then, we have represented the
information seen by the user in a high dimensional format which is reduced to
lower dimensions using tensor factorization techniques. The aspect of intent
awareness (or scoring) is dealt with at this stage. Finally, combining the user
activity data with the contextual information gives the recommendation score.
The final recommendations are then ranked using filtering and collaborative
recommendation techniques to show the top-k recommendations to the user. A
provision for feedback is also envisioned in the current system which informs
the model to update the various weights in the recommender system. Our overall
model aims to combine both frequency-based and context-based recommendation
systems and quantify the intent of a user to provide better recommendations.
We ran experiments on real-world timestamped user activity data, in the
setting of recommending reports to the users of a business analytics tool and
the results are better than the baselines. We also tuned certain aspects of our
model to arrive at optimized results.
|
Physics Informed Deep Learning (Part I): Data-driven Solutions of
Nonlinear Partial Differential Equations | We introduce physics informed neural networks -- neural networks that are
trained to solve supervised learning tasks while respecting any given law of
physics described by general nonlinear partial differential equations. In this
two part treatise, we present our developments in the context of solving two
main classes of problems: data-driven solution and data-driven discovery of
partial differential equations. Depending on the nature and arrangement of the
available data, we devise two distinct classes of algorithms, namely continuous
time and discrete time models. The resulting neural networks form a new class
of data-efficient universal function approximators that naturally encode any
underlying physical laws as prior information. In this first part, we
demonstrate how these networks can be used to infer solutions to partial
differential equations, and obtain physics-informed surrogate models that are
fully differentiable with respect to all input coordinates and free parameters.
|
FearNet: Brain-Inspired Model for Incremental Learning | Incremental class learning involves sequentially learning classes in bursts
of examples from the same class. This violates the assumptions that underlie
methods for training standard deep neural networks, and will cause them to
suffer from catastrophic forgetting. Arguably, the best method for incremental
class learning is iCaRL, but it requires storing training examples for each
class, making it challenging to scale. Here, we propose FearNet for incremental
class learning. FearNet is a generative model that does not store previous
examples, making it memory efficient. FearNet uses a brain-inspired dual-memory
system in which new memories are consolidated from a network for recent
memories inspired by the mammalian hippocampal complex to a network for
long-term storage inspired by medial prefrontal cortex. Memory consolidation is
inspired by mechanisms that occur during sleep. FearNet also uses a module
inspired by the basolateral amygdala for determining which memory system to use
for recall. FearNet achieves state-of-the-art performance at incremental class
learning on image (CIFAR-100, CUB-200) and audio classification (AudioSet)
benchmarks.
|
Physics Informed Deep Learning (Part II): Data-driven Discovery of
Nonlinear Partial Differential Equations | We introduce physics informed neural networks -- neural networks that are
trained to solve supervised learning tasks while respecting any given law of
physics described by general nonlinear partial differential equations. In this
second part of our two-part treatise, we focus on the problem of data-driven
discovery of partial differential equations. Depending on whether the available
data is scattered in space-time or arranged in fixed temporal snapshots, we
introduce two main classes of algorithms, namely continuous time and discrete
time models. The effectiveness of our approach is demonstrated using a wide
range of benchmark problems in mathematical physics, including conservation
laws, incompressible fluid flow, and the propagation of nonlinear shallow-water
waves.
|
Contextual Outlier Interpretation | Outlier detection plays an essential role in many data-driven applications to
identify isolated instances that are different from the majority. While many
statistical learning and data mining techniques have been used for developing
more effective outlier detection algorithms, the interpretation of detected
outliers does not receive much attention. Interpretation is becoming
increasingly important to help people trust and evaluate the developed models
through providing intrinsic reasons why the certain outliers are chosen. It is
difficult, if not impossible, to simply apply feature selection for explaining
outliers due to the distinct characteristics of various detection models,
complicated structures of data in certain applications, and imbalanced
distribution of outliers and normal instances. In addition, the role of
contrastive contexts where outliers locate, as well as the relation between
outliers and contexts, are usually overlooked in interpretation. To tackle the
issues above, in this paper, we propose a novel Contextual Outlier
INterpretation (COIN) method to explain the abnormality of existing outliers
spotted by detectors. The interpretability for an outlier is achieved from
three aspects: outlierness score, attributes that contribute to the
abnormality, and contextual description of its neighborhoods. Experimental
results on various types of datasets demonstrate the flexibility and
effectiveness of the proposed framework compared with existing interpretation
approaches.
|
TensorFlow Distributions | The TensorFlow Distributions library implements a vision of probability
theory adapted to the modern deep-learning paradigm of end-to-end
differentiable computation. Building on two basic abstractions, it offers
flexible building blocks for probabilistic computation. Distributions provide
fast, numerically stable methods for generating samples and computing
statistics, e.g., log density. Bijectors provide composable volume-tracking
transformations with automatic caching. Together these enable modular
construction of high dimensional distributions and transformations not possible
with previous libraries (e.g., pixelCNNs, autoregressive flows, and reversible
residual networks). They are the workhorse behind deep probabilistic
programming systems like Edward and empower fast black-box inference in
probabilistic models built on deep-network components. TensorFlow Distributions
has proven an important part of the TensorFlow toolkit within Google and in the
broader deep learning community.
|
Private federated learning on vertically partitioned data via entity
resolution and additively homomorphic encryption | Consider two data providers, each maintaining private records of different
feature sets about common entities. They aim to learn a linear model jointly in
a federated setting, namely, data is local and a shared model is trained from
locally computed updates. In contrast with most work on distributed learning,
in this scenario (i) data is split vertically, i.e. by features, (ii) only one
data provider knows the target variable and (iii) entities are not linked
across the data providers. Hence, to the challenge of private learning, we add
the potentially negative consequences of mistakes in entity resolution. Our
contribution is twofold. First, we describe a three-party end-to-end solution
in two phases ---privacy-preserving entity resolution and federated logistic
regression over messages encrypted with an additively homomorphic scheme---,
secure against a honest-but-curious adversary. The system allows learning
without either exposing data in the clear or sharing which entities the data
providers have in common. Our implementation is as accurate as a naive
non-private solution that brings all data in one place, and scales to problems
with millions of entities with hundreds of features. Second, we provide what is
to our knowledge the first formal analysis of the impact of entity resolution's
mistakes on learning, with results on how optimal classifiers, empirical
losses, margins and generalisation abilities are affected. Our results bring a
clear and strong support for federated learning: under reasonable assumptions
on the number and magnitude of entity resolution's mistakes, it can be
extremely beneficial to carry out federated learning in the setting where each
peer's data provides a significant uplift to the other.
|
Predicting the Popularity of Online Videos via Deep Neural Networks | Predicting the popularity of online videos is important for video streaming
content providers. This is a challenging problem because of the following two
reasons. First, the problem is both "wide" and "deep". That is, it not only
depends on a wide range of features, but also be highly non-linear and complex.
Second, multiple competitors may be involved. In this paper, we propose a
general prediction model using the multi-task learning (MTL) module and the
relation network (RN) module, where MTL can reduce over-fitting and RN can
model the relations of multiple competitors. Experimental results show that our
proposed approach significantly increases the accuracy on predicting the total
view counts of TV series with RN and MTL modules.
|
Introduction to Tensor Decompositions and their Applications in Machine
Learning | Tensors are multidimensional arrays of numerical values and therefore
generalize matrices to multiple dimensions. While tensors first emerged in the
psychometrics community in the $20^{\text{th}}$ century, they have since then
spread to numerous other disciplines, including machine learning. Tensors and
their decompositions are especially beneficial in unsupervised learning
settings, but are gaining popularity in other sub-disciplines like temporal and
multi-relational data analysis, too.
The scope of this paper is to give a broad overview of tensors, their
decompositions, and how they are used in machine learning. As part of this, we
are going to introduce basic tensor concepts, discuss why tensors can be
considered more rigid than matrices with respect to the uniqueness of their
decomposition, explain the most important factorization algorithms and their
properties, provide concrete examples of tensor decomposition applications in
machine learning, conduct a case study on tensor-based estimation of mixture
models, talk about the current state of research, and provide references to
available software libraries.
|
Automating Vehicles by Deep Reinforcement Learning using Task Separation
with Hill Climbing | Within the context of autonomous driving a model-based reinforcement learning
algorithm is proposed for the design of neural network-parameterized
controllers. Classical model-based control methods, which include sampling- and
lattice-based algorithms and model predictive control, suffer from the
trade-off between model complexity and computational burden required for the
online solution of expensive optimization or search problems at every short
sampling time. To circumvent this trade-off, a 2-step procedure is motivated:
first learning of a controller during offline training based on an arbitrarily
complicated mathematical system model, before online fast feedforward
evaluation of the trained controller. The contribution of this paper is the
proposition of a simple gradient-free and model-based algorithm for deep
reinforcement learning using task separation with hill climbing (TSHC). In
particular, (i) simultaneous training on separate deterministic tasks with the
purpose of encoding many motion primitives in a neural network, and (ii) the
employment of maximally sparse rewards in combination with virtual velocity
constraints (VVCs) in setpoint proximity are advocated.
|
Efficient exploration with Double Uncertain Value Networks | This paper studies directed exploration for reinforcement learning agents by
tracking uncertainty about the value of each available action. We identify two
sources of uncertainty that are relevant for exploration. The first originates
from limited data (parametric uncertainty), while the second originates from
the distribution of the returns (return uncertainty). We identify methods to
learn these distributions with deep neural networks, where we estimate
parametric uncertainty with Bayesian drop-out, while return uncertainty is
propagated through the Bellman equation as a Gaussian distribution. Then, we
identify that both can be jointly estimated in one network, which we call the
Double Uncertain Value Network. The policy is directly derived from the learned
distributions based on Thompson sampling. Experimental results show that both
types of uncertainty may vastly improve learning in domains with a strong
exploration challenge.
|
Reinforcement Learning To Adapt Speech Enhancement to Instantaneous
Input Signal Quality | Today, the optimal performance of existing noise-suppression algorithms, both
data-driven and those based on classic statistical methods, is range bound to
specific levels of instantaneous input signal-to-noise ratios. In this paper,
we present a new approach to improve the adaptivity of such algorithms enabling
them to perform robustly across a wide range of input signal and noise types.
Our methodology is based on the dynamic control of algorithmic parameters via
reinforcement learning. Specifically, we model the noise-suppression module as
a black box, requiring no knowledge of the algorithmic mechanics except a
simple feedback from the output. We utilize this feedback as the reward signal
for a reinforcement-learning agent that learns a policy to adapt the
algorithmic parameters for every incoming audio frame (16 ms of data). Our
preliminary results show that such a control mechanism can substantially
increase the overall performance of the underlying noise-suppression algorithm;
42% and 16% improvements in output SNR and MSE, respectively, when compared to
no adaptivity.
|
Semi-Supervised and Active Few-Shot Learning with Prototypical Networks | We consider the problem of semi-supervised few-shot classification where a
classifier needs to adapt to new tasks using a few labeled examples and
(potentially many) unlabeled examples. We propose a clustering approach to the
problem. The features extracted with Prototypical Networks are clustered using
$K$-means with the few labeled examples guiding the clustering process. We note
that in many real-world applications the adaptation performance can be
significantly improved by requesting the few labels through user feedback. We
demonstrate good performance of the active adaptation strategy using image
data.
|
Deep Reinforcement Learning for De-Novo Drug Design | We propose a novel computational strategy for de novo design of molecules
with desired properties termed ReLeaSE (Reinforcement Learning for Structural
Evolution). Based on deep and reinforcement learning approaches, ReLeaSE
integrates two deep neural networks - generative and predictive - that are
trained separately but employed jointly to generate novel targeted chemical
libraries. ReLeaSE employs simple representation of molecules by their SMILES
strings only. Generative models are trained with stack-augmented memory network
to produce chemically feasible SMILES strings, and predictive models are
derived to forecast the desired properties of the de novo generated compounds.
In the first phase of the method, generative and predictive models are trained
separately with a supervised learning algorithm. In the second phase, both
models are trained jointly with the reinforcement learning approach to bias the
generation of new chemical structures towards those with the desired physical
and/or biological properties. In the proof-of-concept study, we have employed
the ReLeaSE method to design chemical libraries with a bias toward structural
complexity or biased toward compounds with either maximal, minimal, or specific
range of physical properties such as melting point or hydrophobicity, as well
as to develop novel putative inhibitors of JAK2. The approach proposed herein
can find a general use for generating targeted chemical libraries of novel
compounds optimized for either a single desired property or multiple
properties.
|
Causality Refined Diagnostic Prediction | Applying machine learning in the health care domain has shown promising
results in recent years. Interpretable outputs from learning algorithms are
desirable for decision making by health care personnel. In this work, we
explore the possibility of utilizing causal relationships to refine diagnostic
prediction. We focus on the task of diagnostic prediction using discomfort
drawings, and explore two ways to employ causal identification to improve the
diagnostic results. Firstly, we use causal identification to infer the causal
relationships among diagnostic labels which, by itself, provides interpretable
results to aid the decision making and training of health care personnel.
Secondly, we suggest a post-processing approach where the inferred causal
relationships are used to refine the prediction accuracy of a multi-view
probabilistic model. Experimental results show firstly that causal
identification is capable of detecting the causal relationships among
diagnostic labels correctly, and secondly that there is potential for improving
pain diagnostics prediction accuracy using the causal relationships.
|
NPC: Neighbors Progressive Competition Algorithm for Classification of
Imbalanced Data Sets | Learning from many real-world datasets is limited by a problem called the
class imbalance problem. A dataset is imbalanced when one class (the majority
class) has significantly more samples than the other class (the minority
class). Such datasets cause typical machine learning algorithms to perform
poorly on the classification task. To overcome this issue, this paper proposes
a new approach Neighbors Progressive Competition (NPC) for classification of
imbalanced datasets. Whilst the proposed algorithm is inspired by weighted
k-Nearest Neighbor (k-NN) algorithms, it has major differences from them.
Unlike k- NN, NPC does not limit its decision criteria to a preset number of
nearest neighbors. In contrast, NPC considers progressively more neighbors of
the query sample in its decision making until the sum of grades for one class
is much higher than the other classes. Furthermore, NPC uses a novel method for
grading the training samples to compensate for the imbalance issue. The grades
are calculated using both local and global information. In brief, the
contribution of this paper is an entirely new classifier for handling the
imbalance issue effectively without any manually-set parameters or any need for
expert knowledge. Experimental results compare the proposed approach with five
representative algorithms applied to fifteen imbalanced datasets and illustrate
this algorithms effectiveness.
|
Extreme Dimension Reduction for Handling Covariate Shift | In the covariate shift learning scenario, the training and test covariate
distributions differ, so that a predictor's average loss over the training and
test distributions also differ. In this work, we explore the potential of
extreme dimension reduction, i.e. to very low dimensions, in improving the
performance of importance weighting methods for handling covariate shift, which
fail in high dimensions due to potentially high train/test covariate divergence
and the inability to accurately estimate the requisite density ratios. We first
formulate and solve a problem optimizing over linear subspaces a combination of
their predictive utility and train/test divergence within. Applying it to
simulated and real data, we show extreme dimension reduction helps sometimes
but not always, due to a bias introduced by dimension reduction.
|
The Block Point Process Model for Continuous-Time Event-Based Dynamic
Networks | We consider the problem of analyzing timestamped relational events between a
set of entities, such as messages between users of an on-line social network.
Such data are often analyzed using static or discrete-time network models,
which discard a significant amount of information by aggregating events over
time to form network snapshots. In this paper, we introduce a block point
process model (BPPM) for continuous-time event-based dynamic networks. The BPPM
is inspired by the well-known stochastic block model (SBM) for static networks.
We show that networks generated by the BPPM follow an SBM in the limit of a
growing number of nodes. We use this property to develop principled and
efficient local search and variational inference procedures initialized by
regularized spectral clustering. We fit BPPMs with exponential Hawkes processes
to analyze several real network data sets, including a Facebook wall post
network with over 3,500 nodes and 130,000 events.
|
A Novel Data-Driven Framework for Risk Characterization and Prediction
from Electronic Medical Records: A Case Study of Renal Failure | Electronic medical records (EMR) contain longitudinal information about
patients that can be used to analyze outcomes. Typically, studies on EMR data
have worked with established variables that have already been acknowledged to
be associated with certain outcomes. However, EMR data may also contain
hitherto unrecognized factors for risk association and prediction of outcomes
for a disease. In this paper, we present a scalable data-driven framework to
analyze EMR data corpus in a disease agnostic way that systematically uncovers
important factors influencing outcomes in patients, as supported by data and
without expert guidance. We validate the importance of such factors by using
the framework to predict for the relevant outcomes. Specifically, we analyze
EMR data covering approximately 47 million unique patients to characterize
renal failure (RF) among type 2 diabetic (T2DM) patients. We propose a
specialized L1 regularized Cox Proportional Hazards (CoxPH) survival model to
identify the important factors from those available from patient encounter
history. To validate the identified factors, we use a specialized generalized
linear model (GLM) to predict the probability of renal failure for individual
patients within a specified time window. Our experiments indicate that the
factors identified via our data-driven method overlap with the patient
characteristics recognized by experts. Our approach allows for scalable,
repeatable and efficient utilization of data available in EMRs, confirms prior
medical knowledge and can generate new hypothesis without expert supervision.
|
Embedding Words as Distributions with a Bayesian Skip-gram Model | We introduce a method for embedding words as probability densities in a
low-dimensional space. Rather than assuming that a word embedding is fixed
across the entire text collection, as in standard word embedding methods, in
our Bayesian model we generate it from a word-specific prior density for each
occurrence of a given word. Intuitively, for each word, the prior density
encodes the distribution of its potential 'meanings'. These prior densities are
conceptually similar to Gaussian embeddings. Interestingly, unlike the Gaussian
embeddings, we can also obtain context-specific densities: they encode
uncertainty about the sense of a word given its context and correspond to
posterior distributions within our model. The context-dependent densities have
many potential applications: for example, we show that they can be directly
used in the lexical substitution task. We describe an effective estimation
method based on the variational autoencoding framework. We also demonstrate
that our embeddings achieve competitive results on standard benchmarks.
|
Gaussian Process Neurons Learn Stochastic Activation Functions | We propose stochastic, non-parametric activation functions that are fully
learnable and individual to each neuron. Complexity and the risk of overfitting
are controlled by placing a Gaussian process prior over these functions. The
result is the Gaussian process neuron, a probabilistic unit that can be used as
the basic building block for probabilistic graphical models that resemble the
structure of neural networks. The proposed model can intrinsically handle
uncertainties in its inputs and self-estimate the confidence of its
predictions. Using variational Bayesian inference and the central limit
theorem, a fully deterministic loss function is derived, allowing it to be
trained as efficiently as a conventional neural network using mini-batch
gradient descent. The posterior distribution of activation functions is
inferred from the training data alongside the weights of the network.
The proposed model favorably compares to deep Gaussian processes, both in
model complexity and efficiency of inference. It can be directly applied to
recurrent or convolutional network structures, allowing its use in audio and
image processing tasks.
As an preliminary empirical evaluation we present experiments on regression
and classification tasks, in which our model achieves performance comparable to
or better than a Dropout regularized neural network with a fixed activation
function. Experiments are ongoing and results will be added as they become
available.
|
Paradoxes in Fair Computer-Aided Decision Making | Computer-aided decision making--where a human decision-maker is aided by a
computational classifier in making a decision--is becoming increasingly
prevalent. For instance, judges in at least nine states make use of algorithmic
tools meant to determine "recidivism risk scores" for criminal defendants in
sentencing, parole, or bail decisions. A subject of much recent debate is
whether such algorithmic tools are "fair" in the sense that they do not
discriminate against certain groups (e.g., races) of people.
Our main result shows that for "non-trivial" computer-aided decision making,
either the classifier must be discriminatory, or a rational decision-maker
using the output of the classifier is forced to be discriminatory. We further
provide a complete characterization of situations where fair computer-aided
decision making is possible.
|
HSC: A Novel Method for Clustering Hierarchies of Networked Data | Hierarchical clustering is one of the most powerful solutions to the problem
of clustering, on the grounds that it performs a multi scale organization of
the data. In recent years, research on hierarchical clustering methods has
attracted considerable interest due to the demanding modern application
domains.
We present a novel divisive hierarchical clustering framework called
Hierarchical Stochastic Clustering (HSC), that acts in two stages. In the first
stage, it finds a primary hierarchy of clustering partitions in a dataset. In
the second stage, feeds a clustering algorithm with each one of the clusters of
the very detailed partition, in order to settle the final result. The output is
a hierarchy of clusters. Our method is based on the previous research of Meyer
and Weissel Stochastic Data Clustering and the theory of Simon and Ando on
Variable Aggregation.
Our experiments show that our framework builds a meaningful hierarchy of
clusters and benefits consistently the clustering algorithm that acts in the
second stage, not only computationally but also in terms of cluster quality.
This result suggest that HSC framework is ideal for obtaining hierarchical
solutions of large volumes of data.
|
Improving Latent User Models in Online Social Media | Modern social platforms are characterized by the presence of rich
user-behavior data associated with the publication, sharing and consumption of
textual content. Users interact with content and with each other in a complex
and dynamic social environment while simultaneously evolving over time. In
order to effectively characterize users and predict their future behavior in
such a setting, it is necessary to overcome several challenges. Content
heterogeneity and temporal inconsistency of behavior data result in severe
sparsity at the user level. In this paper, we propose a novel
mutual-enhancement framework to simultaneously partition and learn latent
activity profiles of users. We propose a flexible user partitioning approach to
effectively discover rare behaviors and tackle user-level sparsity. We
extensively evaluate the proposed framework on massive datasets from real-world
platforms including Q&A networks and interactive online courses (MOOCs). Our
results indicate significant gains over state-of-the-art behavior models ( 15%
avg ) in a varied range of tasks and our gains are further magnified for users
with limited interaction data. The proposed algorithms are amenable to
parallelization, scale linearly in the size of datasets, and provide
flexibility to model diverse facets of user behavior.
|
Easy High-Dimensional Likelihood-Free Inference | We introduce a framework using Generative Adversarial Networks (GANs) for
likelihood--free inference (LFI) and Approximate Bayesian Computation (ABC)
where we replace the black-box simulator model with an approximator network and
generate a rich set of summary features in a data driven fashion. On benchmark
data sets, our approach improves on others with respect to scalability, ability
to handle high dimensional data and complex probability distributions.
|
A Semantic Loss Function for Deep Learning with Symbolic Knowledge | This paper develops a novel methodology for using symbolic knowledge in deep
learning. From first principles, we derive a semantic loss function that
bridges between neural output vectors and logical constraints. This loss
function captures how close the neural network is to satisfying the constraints
on its output. An experimental evaluation shows that it effectively guides the
learner to achieve (near-)state-of-the-art results on semi-supervised
multi-class classification. Moreover, it significantly increases the ability of
the neural network to predict structured objects, such as rankings and paths.
These discrete concepts are tremendously difficult to learn, and benefit from a
tight integration of deep learning and symbolic reasoning methods.
|
Safe Exploration for Identifying Linear Systems via Robust Optimization | Safely exploring an unknown dynamical system is critical to the deployment of
reinforcement learning (RL) in physical systems where failures may have
catastrophic consequences. In scenarios where one knows little about the
dynamics, diverse transition data covering relevant regions of state-action
space is needed to apply either model-based or model-free RL. Motivated by the
cooling of Google's data centers, we study how one can safely identify the
parameters of a system model with a desired accuracy and confidence level. In
particular, we focus on learning an unknown linear system with Gaussian noise
assuming only that, initially, a nominal safe action is known. Define safety as
satisfying specific linear constraints on the state space (e.g., requirements
on process variable) that must hold over the span of an entire trajectory, and
given a Probably Approximately Correct (PAC) style bound on the estimation
error of model parameters, we show how to compute safe regions of action space
by gradually growing a ball around the nominal safe action. One can apply any
exploration strategy where actions are chosen from such safe regions.
Experiments on a stylized model of data center cooling dynamics show how
computing proper safe regions can increase the sample efficiency of safe
exploration.
|
State Space LSTM Models with Particle MCMC Inference | Long Short-Term Memory (LSTM) is one of the most powerful sequence models.
Despite the strong performance, however, it lacks the nice interpretability as
in state space models. In this paper, we present a way to combine the best of
both worlds by introducing State Space LSTM (SSL) models that generalizes the
earlier work \cite{zaheer2017latent} of combining topic models with LSTM.
However, unlike \cite{zaheer2017latent}, we do not make any factorization
assumptions in our inference algorithm. We present an efficient sampler based
on sequential Monte Carlo (SMC) method that draws from the joint posterior
directly. Experimental results confirms the superiority and stability of this
SMC inference algorithm on a variety of domains.
|
Variational Deep Q Network | We propose a framework that directly tackles the probability distribution of
the value function parameters in Deep Q Network (DQN), with powerful
variational inference subroutines to approximate the posterior of the
parameters. We will establish the equivalence between our proposed surrogate
objective and variational inference loss. Our new algorithm achieves efficient
exploration and performs well on large scale chain Markov Decision Process
(MDP).
|
Provably noise-robust, regularised $k$-means clustering | We consider the problem of clustering in the presence of noise. That is, when
on top of cluster structure, the data also contains a subset of
\emph{unstructured} points. Our goal is to detect the clusters despite the
presence of many unstructured points. Any algorithm that achieves this goal is
noise-robust. We consider a regularisation method which converts any
center-based clustering objective into a noise-robust one. We focus on the
$k$-means objective and we prove that the regularised version of $k$-means is
NP-Hard even for $k=1$. We consider two algorithms based on the convex (sdp and
lp) relaxation of the regularised objective and prove robustness guarantees for
both.
The sdp and lp relaxation of the standard (non-regularised) $k$-means
objective has been previously studied by [ABC+15]. Under the stochastic ball
model of the data they show that the sdp-based algorithm recovers the
underlying structure as long as the balls are separated by $\delta > 2\sqrt{2}
+ \epsilon$. We improve upon this result in two ways. First, we show recovery
even for $\delta > 2 + \epsilon$. Second, our regularised algorithm recovers
the balls even in the presence of noise so long as the number of noisy points
is not too large. We complement our theoretical analysis with simulations and
analyse the effect of various parameters like regularization constant,
noise-level etc. on the performance of our algorithm. In the presence of noise,
our algorithm performs better than $k$-means++ on MNIST.
|
Towards Accurate Binary Convolutional Neural Network | We introduce a novel scheme to train binary convolutional neural networks
(CNNs) -- CNNs with weights and activations constrained to {-1,+1} at run-time.
It has been known that using binary weights and activations drastically reduce
memory size and accesses, and can replace arithmetic operations with more
efficient bitwise operations, leading to much faster test-time inference and
lower power consumption. However, previous works on binarizing CNNs usually
result in severe prediction accuracy degradation. In this paper, we address
this issue with two major innovations: (1) approximating full-precision weights
with the linear combination of multiple binary weight bases; (2) employing
multiple binary activations to alleviate information loss. The implementation
of the resulting binary CNN, denoted as ABC-Net, is shown to achieve much
closer performance to its full-precision counterpart, and even reach the
comparable prediction accuracy on ImageNet and forest trail datasets, given
adequate binary weight bases and activations.
|
Multivariate Time Series Classification with WEASEL+MUSE | Multivariate time series (MTS) arise when multiple interconnected sensors
record data over time. Dealing with this high-dimensional data is challenging
for every classifier for at least two aspects: First, an MTS is not only
characterized by individual feature values, but also by the interplay of
features in different dimensions. Second, this typically adds large amounts of
irrelevant data and noise. We present our novel MTS classifier WEASEL+MUSE
which addresses both challenges. WEASEL+MUSE builds a multivariate feature
vector, first using a sliding-window approach applied to each dimension of the
MTS, then extracts discrete features per window and dimension. The feature
vector is subsequently fed through feature selection, removing
non-discriminative features, and analysed by a machine learning classifier. The
novelty of WEASEL+MUSE lies in its specific way of extracting and filtering
multivariate features from MTS by encoding context information into each
feature. Still the resulting feature set is small, yet very discriminative and
useful for MTS classification. Based on a popular benchmark of 20 MTS datasets,
we found that WEASEL+MUSE is among the most accurate classifiers, when compared
to the state of the art. The outstanding robustness of WEASEL+MUSE is further
confirmed based on motion gesture recognition data, where it out-of-the-box
achieved similar accuracies as domain-specific methods.
|
Learning to Learn from Weak Supervision by Full Supervision | In this paper, we propose a method for training neural networks when we have
a large set of data with weak labels and a small amount of data with true
labels. In our proposed model, we train two neural networks: a target network,
the learner and a confidence network, the meta-learner. The target network is
optimized to perform a given task and is trained using a large set of unlabeled
data that are weakly annotated. We propose to control the magnitude of the
gradient updates to the target network using the scores provided by the second
confidence network, which is trained on a small amount of supervised data. Thus
we avoid that the weight updates computed from noisy labels harm the quality of
the target network model.
|
Scalable synthesis of safety certificates from data with application to
learning-based control | The control of complex systems faces a trade-off between high performance and
safety guarantees, which in particular restricts the application of
learning-based methods to safety-critical systems. A recently proposed
framework to address this issue is the use of a safety controller, which
guarantees to keep the system within a safe region of the state space. This
paper introduces efficient techniques for the synthesis of a safe set and
control law, which offer improved scalability properties by relying on
approximations based on convex optimization problems. The first proposed method
requires only an approximate linear system model and Lipschitz continuity of
the unknown nonlinear dynamics. The second method extends the results by
showing how a Gaussian process prior on the unknown system dynamics can be used
in order to reduce conservatism of the resulting safe set. We demonstrate the
results with numerical examples, including an autonomous convoy of vehicles.
|
On reducing the communication cost of the diffusion LMS algorithm | The rise of digital and mobile communications has recently made the world
more connected and networked, resulting in an unprecedented volume of data
flowing between sources, data centers, or processes. While these data may be
processed in a centralized manner, it is often more suitable to consider
distributed strategies such as diffusion as they are scalable and can handle
large amounts of data by distributing tasks over networked agents. Although it
is relatively simple to implement diffusion strategies over a cluster, it
appears to be challenging to deploy them in an ad-hoc network with limited
energy budget for communication. In this paper, we introduce a diffusion LMS
strategy that significantly reduces communication costs without compromising
the performance. Then, we analyze the proposed algorithm in the mean and
mean-square sense. Next, we conduct numerical experiments to confirm the
theoretical findings. Finally, we perform large scale simulations to test the
algorithm efficiency in a scenario where energy is limited.
|
SyGuS-Comp 2017: Results and Analysis | Syntax-Guided Synthesis (SyGuS) is the computational problem of finding an
implementation f that meets both a semantic constraint given by a logical
formula phi in a background theory T, and a syntactic constraint given by a
grammar G, which specifies the allowed set of candidate implementations. Such a
synthesis problem can be formally defined in SyGuS-IF, a language that is built
on top of SMT-LIB.
The Syntax-Guided Synthesis Competition (SyGuS-Comp) is an effort to
facilitate, bring together and accelerate research and development of efficient
solvers for SyGuS by providing a platform for evaluating different synthesis
techniques on a comprehensive set of benchmarks. In this year's competition six
new solvers competed on over 1500 benchmarks. This paper presents and analyses
the results of SyGuS-Comp'17.
|
ConvNets and ImageNet Beyond Accuracy: Understanding Mistakes and
Uncovering Biases | ConvNets and Imagenet have driven the recent success of deep learning for
image classification. However, the marked slowdown in performance improvement
combined with the lack of robustness of neural networks to adversarial examples
and their tendency to exhibit undesirable biases question the reliability of
these methods. This work investigates these questions from the perspective of
the end-user by using human subject studies and explanations. The contribution
of this study is threefold. We first experimentally demonstrate that the
accuracy and robustness of ConvNets measured on Imagenet are vastly
underestimated. Next, we show that explanations can mitigate the impact of
misclassified adversarial examples from the perspective of the end-user. We
finally introduce a novel tool for uncovering the undesirable biases learned by
a model. These contributions also show that explanations are a valuable tool
both for improving our understanding of ConvNets' predictions and for designing
more reliable models.
|
Uncertainty Estimates for Efficient Neural Network-based Dialogue Policy
Optimisation | In statistical dialogue management, the dialogue manager learns a policy that
maps a belief state to an action for the system to perform. Efficient
exploration is key to successful policy optimisation. Current deep
reinforcement learning methods are very promising but rely on epsilon-greedy
exploration, thus subjecting the user to a random choice of action during
learning. Alternative approaches such as Gaussian Process SARSA (GPSARSA)
estimate uncertainties and are sample efficient, leading to better user
experience, but on the expense of a greater computational complexity. This
paper examines approaches to extract uncertainty estimates from deep Q-networks
(DQN) in the context of dialogue management. We perform an extensive benchmark
of deep Bayesian methods to extract uncertainty estimates, namely
Bayes-By-Backprop, dropout, its concrete variation, bootstrapped ensemble and
alpha-divergences, combining it with DQN algorithm.
|
Convolutional Networks with Adaptive Inference Graphs | Do convolutional networks really need a fixed feed-forward structure? What
if, after identifying the high-level concept of an image, a network could move
directly to a layer that can distinguish fine-grained differences? Currently, a
network would first need to execute sometimes hundreds of intermediate layers
that specialize in unrelated aspects. Ideally, the more a network already knows
about an image, the better it should be at deciding which layer to compute
next. In this work, we propose convolutional networks with adaptive inference
graphs (ConvNet-AIG) that adaptively define their network topology conditioned
on the input image. Following a high-level structure similar to residual
networks (ResNets), ConvNet-AIG decides for each input image on the fly which
layers are needed. In experiments on ImageNet we show that ConvNet-AIG learns
distinct inference graphs for different categories. Both ConvNet-AIG with 50
and 101 layers outperform their ResNet counterpart, while using 20% and 38%
less computations respectively. By grouping parameters into layers for related
classes and only executing relevant layers, ConvNet-AIG improves both
efficiency and overall classification quality. Lastly, we also study the effect
of adaptive inference graphs on the susceptibility towards adversarial
examples. We observe that ConvNet-AIG shows a higher robustness than ResNets,
complementing other known defense mechanisms.
|
Improved Linear Embeddings via Lagrange Duality | Near isometric orthogonal embeddings to lower dimensions are a fundamental
tool in data science and machine learning. In this paper, we present the
construction of such embeddings that minimizes the maximum distortion for a
given set of points. We formulate the problem as a non convex constrained
optimization problem. We first construct a primal relaxation and then use the
theory of Lagrange duality to create dual relaxation. We also suggest a
polynomial time algorithm based on the theory of convex optimization to solve
the dual relaxation provably. We provide a theoretical upper bound on the
approximation guarantees for our algorithm, which depends only on the spectral
properties of the dataset. We experimentally demonstrate the superiority of our
algorithm compared to baselines in terms of the scalability and the ability to
achieve lower distortion.
|
Predicting Severe Sepsis Using Text from the Electronic Health Record | Employing a machine learning approach we predict, up to 24 hours prior, a
diagnosis of severe sepsis. Strongly predictive models are possible that use
only text reports from the Electronic Health Record (EHR), and omit structured
numerical data. Unstructured text alone gives slightly better performance than
structured data alone, and the combination further improves performance. We
also discuss advantages of using unstructured EHR text for modeling, as
compared to structured EHR data.
|
Learning to Adapt by Minimizing Discrepancy | We explore whether useful temporal neural generative models can be learned
from sequential data without back-propagation through time. We investigate the
viability of a more neurocognitively-grounded approach in the context of
unsupervised generative modeling of sequences. Specifically, we build on the
concept of predictive coding, which has gained influence in cognitive science,
in a neural framework. To do so we develop a novel architecture, the Temporal
Neural Coding Network, and its learning algorithm, Discrepancy Reduction. The
underlying directed generative model is fully recurrent, meaning that it
employs structural feedback connections and temporal feedback connections,
yielding information propagation cycles that create local learning signals.
This facilitates a unified bottom-up and top-down approach for information
transfer inside the architecture. Our proposed algorithm shows promise on the
bouncing balls generative modeling problem. Further experiments could be
conducted to explore the strengths and weaknesses of our approach.
|
Embodied Question Answering | We present a new AI task -- Embodied Question Answering (EmbodiedQA) -- where
an agent is spawned at a random location in a 3D environment and asked a
question ("What color is the car?"). In order to answer, the agent must first
intelligently navigate to explore the environment, gather information through
first-person (egocentric) vision, and then answer the question ("orange").
This challenging task requires a range of AI skills -- active perception,
language understanding, goal-driven navigation, commonsense reasoning, and
grounding of language into actions. In this work, we develop the environments,
end-to-end-trained reinforcement learning agents, and evaluation protocols for
EmbodiedQA.
|
Measuring the tendency of CNNs to Learn Surface Statistical Regularities | Deep CNNs are known to exhibit the following peculiarity: on the one hand
they generalize extremely well to a test set, while on the other hand they are
extremely sensitive to so-called adversarial perturbations. The extreme
sensitivity of high performance CNNs to adversarial examples casts serious
doubt that these networks are learning high level abstractions in the dataset.
We are concerned with the following question: How can a deep CNN that does not
learn any high level semantics of the dataset manage to generalize so well? The
goal of this article is to measure the tendency of CNNs to learn surface
statistical regularities of the dataset. To this end, we use Fourier filtering
to construct datasets which share the exact same high level abstractions but
exhibit qualitatively different surface statistical regularities. For the SVHN
and CIFAR-10 datasets, we present two Fourier filtered variants: a low
frequency variant and a randomly filtered variant. Each of the Fourier
filtering schemes is tuned to preserve the recognizability of the objects. Our
main finding is that CNNs exhibit a tendency to latch onto the Fourier image
statistics of the training dataset, sometimes exhibiting up to a 28%
generalization gap across the various test sets. Moreover, we observe that
significantly increasing the depth of a network has a very marginal impact on
closing the aforementioned generalization gap. Thus we provide quantitative
evidence supporting the hypothesis that deep CNNs tend to learn surface
statistical regularities in the dataset rather than higher-level abstract
concepts.
|
Hybrid VAE: Improving Deep Generative Models using Partial Observations | Deep neural network models trained on large labeled datasets are the
state-of-the-art in a large variety of computer vision tasks. In many
applications, however, labeled data is expensive to obtain or requires a time
consuming manual annotation process. In contrast, unlabeled data is often
abundant and available in large quantities. We present a principled framework
to capitalize on unlabeled data by training deep generative models on both
labeled and unlabeled data. We show that such a combination is beneficial
because the unlabeled data acts as a data-driven form of regularization,
allowing generative models trained on few labeled samples to reach the
performance of fully-supervised generative models trained on much larger
datasets. We call our method Hybrid VAE (H-VAE) as it contains both the
generative and the discriminative parts. We validate H-VAE on three large-scale
datasets of different modalities: two face datasets: (MultiPIE, CelebA) and a
hand pose dataset (NYU Hand Pose). Our qualitative visualizations further
support improvements achieved by using partial observations.
|
Outlier-robust moment-estimation via sum-of-squares | We develop efficient algorithms for estimating low-degree moments of unknown
distributions in the presence of adversarial outliers. The guarantees of our
algorithms improve in many cases significantly over the best previous ones,
obtained in recent works of Diakonikolas et al, Lai et al, and Charikar et al.
We also show that the guarantees of our algorithms match information-theoretic
lower-bounds for the class of distributions we consider. These improved
guarantees allow us to give improved algorithms for independent component
analysis and learning mixtures of Gaussians in the presence of outliers.
Our algorithms are based on a standard sum-of-squares relaxation of the
following conceptually-simple optimization problem: Among all distributions
whose moments are bounded in the same way as for the unknown distribution, find
the one that is closest in statistical distance to the empirical distribution
of the adversarially-corrupted sample.
|
High-Resolution Image Synthesis and Semantic Manipulation with
Conditional GANs | We present a new method for synthesizing high-resolution photo-realistic
images from semantic label maps using conditional generative adversarial
networks (conditional GANs). Conditional GANs have enabled a variety of
applications, but the results are often limited to low-resolution and still far
from realistic. In this work, we generate 2048x1024 visually appealing results
with a novel adversarial loss, as well as new multi-scale generator and
discriminator architectures. Furthermore, we extend our framework to
interactive visual manipulation with two additional features. First, we
incorporate object instance segmentation information, which enables object
manipulations such as removing/adding objects and changing the object category.
Second, we propose a method to generate diverse results given the same input,
allowing users to edit the object appearance interactively. Human opinion
studies demonstrate that our method significantly outperforms existing methods,
advancing both the quality and the resolution of deep image synthesis and
editing.
|
Modeling Information Flow Through Deep Neural Networks | This paper proposes a principled information theoretic analysis of
classification for deep neural network structures, e.g. convolutional neural
networks (CNN). The output of convolutional filters is modeled as a random
variable Y conditioned on the object class C and network filter bank F. The
conditional entropy (CENT) H(Y |C,F) is shown in theory and experiments to be a
highly compact and class-informative code, that can be computed from the filter
outputs throughout an existing CNN and used to obtain higher classification
results than the original CNN itself. Experiments demonstrate the effectiveness
of CENT feature analysis in two separate CNN classification contexts. 1) In the
classification of neurodegeneration due to Alzheimer's disease (AD) and natural
aging from 3D magnetic resonance image (MRI) volumes, 3 CENT features result in
an AUC=94.6% for whole-brain AD classification, the highest reported accuracy
on the public OASIS dataset used and 12% higher than the softmax output of the
original CNN trained for the task. 2) In the context of visual object
classification from 2D photographs, transfer learning based on a small set of
CENT features identified throughout an existing CNN leads to AUC values
comparable to the 1000-feature softmax output of the original network when
classifying previously unseen object categories. The general information
theoretical analysis explains various recent CNN design successes, e.g. densely
connected CNN architectures, and provides insights for future research
directions in deep learning.
|
Learnings Options End-to-End for Continuous Action Tasks | We present new results on learning temporally extended actions for
continuoustasks, using the options framework (Suttonet al.[1999b], Precup
[2000]). In orderto achieve this goal we work with the option-critic
architecture (Baconet al.[2017])using a deliberation cost and train it with
proximal policy optimization (Schulmanet al.[2017]) instead of vanilla policy
gradient. Results on Mujoco domains arepromising, but lead to interesting
questions aboutwhena given option should beused, an issue directly connected to
the use of initiation sets.
|
Comparing Deep Reinforcement Learning and Evolutionary Methods in
Continuous Control | Reinforcement Learning and the Evolutionary Strategy are two major approaches
in addressing complicated control problems. Both are strong contenders and have
their own devotee communities. Both groups have been very active in developing
new advances in their own domain and devising, in recent years, leading-edge
techniques to address complex continuous control tasks. Here, in the context of
Deep Reinforcement Learning, we formulate a parallelized version of the
Proximal Policy Optimization method and a Deep Deterministic Policy Gradient
method. Moreover, we conduct a thorough comparison between the state-of-the-art
techniques in both camps fro continuous control; evolutionary methods and Deep
Reinforcement Learning methods. The results show there is no consistent winner.
|
Highrisk Prediction from Electronic Medical Records via Deep Attention
Networks | Predicting highrisk vascular diseases is a significant issue in the medical
domain. Most predicting methods predict the prognosis of patients from
pathological and radiological measurements, which are expensive and require
much time to be analyzed. Here we propose deep attention models that predict
the onset of the high risky vascular disease from symbolic medical histories
sequence of hypertension patients such as ICD-10 and pharmacy codes only,
Medical History-based Prediction using Attention Network (MeHPAN). We
demonstrate two types of attention models based on 1) bidirectional gated
recurrent unit (R-MeHPAN) and 2) 1D convolutional multilayer model (C-MeHPAN).
Two MeHPAN models are evaluated on approximately 50,000 hypertension patients
with respect to precision, recall, f1-measure and area under the curve (AUC).
Experimental results show that our MeHPAN methods outperform standard
classification models. Comparing two MeHPANs, R-MeHPAN provides more better
discriminative capability with respect to all metrics while C-MeHPAN presents
much shorter training time with competitive accuracy.
|
Feature discovery and visualization of robot mission data using
convolutional autoencoders and Bayesian nonparametric topic models | The gap between our ability to collect interesting data and our ability to
analyze these data is growing at an unprecedented rate. Recent algorithmic
attempts to fill this gap have employed unsupervised tools to discover
structure in data. Some of the most successful approaches have used
probabilistic models to uncover latent thematic structure in discrete data.
Despite the success of these models on textual data, they have not generalized
as well to image data, in part because of the spatial and temporal structure
that may exist in an image stream.
We introduce a novel unsupervised machine learning framework that
incorporates the ability of convolutional autoencoders to discover features
from images that directly encode spatial information, within a Bayesian
nonparametric topic model that discovers meaningful latent patterns within
discrete data. By using this hybrid framework, we overcome the fundamental
dependency of traditional topic models on rigidly hand-coded data
representations, while simultaneously encoding spatial dependency in our topics
without adding model complexity. We apply this model to the motivating
application of high-level scene understanding and mission summarization for
exploratory marine robots. Our experiments on a seafloor dataset collected by a
marine robot show that the proposed hybrid framework outperforms current
state-of-the-art approaches on the task of unsupervised seafloor terrain
characterization.
|
Paris-Lille-3D: a large and high-quality ground truth urban point cloud
dataset for automatic segmentation and classification | This paper introduces a new Urban Point Cloud Dataset for Automatic
Segmentation and Classification acquired by Mobile Laser Scanning (MLS). We
describe how the dataset is obtained from acquisition to post-processing and
labeling. This dataset can be used to learn classification algorithm, however,
given that a great attention has been paid to the split between the different
objects, this dataset can also be used to learn the segmentation. The dataset
consists of around 2km of MLS point cloud acquired in two cities. The number of
points and range of classes make us consider that it can be used to train
Deep-Learning methods. Besides we show some results of automatic segmentation
and classification. The dataset is available at:
http://caor-mines-paristech.fr/fr/paris-lille-3d-dataset/
|
Machine Learning and Manycore Systems Design: A Serendipitous Symbiosis | Tight collaboration between experts of machine learning and manycore system
design is necessary to create a data-driven manycore design framework that
integrates both learning and expert knowledge. Such a framework will be
necessary to address the rising complexity of designing large-scale manycore
systems and machine learning techniques.
|
Blind Gain and Phase Calibration via Sparse Spectral Methods | Blind gain and phase calibration (BGPC) is a bilinear inverse problem
involving the determination of unknown gains and phases of the sensing system,
and the unknown signal, jointly. BGPC arises in numerous applications, e.g.,
blind albedo estimation in inverse rendering, synthetic aperture radar
autofocus, and sensor array auto-calibration. In some cases, sparse structure
in the unknown signal alleviates the ill-posedness of BGPC. Recently there has
been renewed interest in solutions to BGPC with careful analysis of error
bounds. In this paper, we formulate BGPC as an eigenvalue/eigenvector problem,
and propose to solve it via power iteration, or in the sparsity or joint
sparsity case, via truncated power iteration. Under certain assumptions, the
unknown gains, phases, and the unknown signal can be recovered simultaneously.
Numerical experiments show that power iteration algorithms work not only in the
regime predicted by our main results, but also in regimes where theoretical
analysis is limited. We also show that our power iteration algorithms for BGPC
compare favorably with competing algorithms in adversarial conditions, e.g.,
with noisy measurement or with a bad initial estimate.
|
Towards Personalized Modeling of the Female Hormonal Cycle: Experiments
with Mechanistic Models and Gaussian Processes | In this paper, we introduce a novel task for machine learning in healthcare,
namely personalized modeling of the female hormonal cycle. The motivation for
this work is to model the hormonal cycle and predict its phases in time, both
for healthy individuals and for those with disorders of the reproductive
system. Because there are individual differences in the menstrual cycle, we are
particularly interested in personalized models that can account for individual
idiosyncracies, towards identifying phenotypes of menstrual cycles. As a first
step, we consider the hormonal cycle as a set of observations through time. We
use a previously validated mechanistic model to generate realistic hormonal
patterns, and experiment with Gaussian process regression to estimate their
values over time. Specifically, we are interested in the feasibility of
predicting menstrual cycle phases under varying learning conditions: number of
cycles used for training, hormonal measurement noise and sampling rates, and
informed vs. agnostic sampling of hormonal measurements. Our results indicate
that Gaussian processes can help model the female menstrual cycle. We discuss
the implications of our experiments in the context of modeling the female
menstrual cycle.
|
An interpretable latent variable model for attribute applicability in
the Amazon catalogue | Learning attribute applicability of products in the Amazon catalog (e.g.,
predicting that a shoe should have a value for size, but not for battery-type
at scale is a challenge. The need for an interpretable model is contingent on
(1) the lack of ground truth training data, (2) the need to utilise prior
information about the underlying latent space and (3) the ability to understand
the quality of predictions on new, unseen data. To this end, we develop the
MaxMachine, a probabilistic latent variable model that learns distributed
binary representations, associated to sets of features that are likely to
co-occur in the data. Layers of MaxMachines can be stacked such that higher
layers encode more abstract information. Any set of variables can be clamped to
encode prior information. We develop fast sampling based posterior inference.
Preliminary results show that the model improves over the baseline in 17 out of
19 product groups and provides qualitatively reasonable predictions.
|
Experimental learning of quantum states | The number of parameters describing a quantum state is well known to grow
exponentially with the number of particles. This scaling clearly limits our
ability to do tomography to systems with no more than a few qubits and has been
used to argue against the universal validity of quantum mechanics itself.
However, from a computational learning theory perspective, it can be shown
that, in a probabilistic setting, quantum states can be approximately learned
using only a linear number of measurements. Here we experimentally demonstrate
this linear scaling in optical systems with up to 6 qubits. Our results
highlight the power of computational learning theory to investigate quantum
information, provide the first experimental demonstration that quantum states
can be "probably approximately learned" with access to a number of copies of
the state that scales linearly with the number of qubits, and pave the way to
probing quantum states at new, larger scales.
|
Generative Adversarial Networks for Electronic Health Records: A
Framework for Exploring and Evaluating Methods for Predicting Drug-Induced
Laboratory Test Trajectories | Generative Adversarial Networks (GANs) represent a promising class of
generative networks that combine neural networks with game theory. From
generating realistic images and videos to assisting musical creation, GANs are
transforming many fields of arts and sciences. However, their application to
healthcare has not been fully realized, more specifically in generating
electronic health records (EHR) data. In this paper, we propose a framework for
exploring the value of GANs in the context of continuous laboratory time series
data. We devise an unsupervised evaluation method that measures the predictive
power of synthetic laboratory test time series. Further, we show that when it
comes to predicting the impact of drug exposure on laboratory test data,
incorporating representation learning of the training cohorts prior to training
GAN models is beneficial.
|
Audio Cover Song Identification using Convolutional Neural Network | In this paper, we propose a new approach to cover song identification using a
CNN (convolutional neural network). Most previous studies extract the feature
vectors that characterize the cover song relation from a pair of songs and used
it to compute the (dis)similarity between the two songs. Based on the
observation that there is a meaningful pattern between cover songs and that
this can be learned, we have reformulated the cover song identification problem
in a machine learning framework. To do this, we first build the CNN using as an
input a cross-similarity matrix generated from a pair of songs. We then
construct the data set composed of cover song pairs and non-cover song pairs,
which are used as positive and negative training samples, respectively. The
trained CNN outputs the probability of being in the cover song relation given a
cross-similarity matrix generated from any two pieces of music and identifies
the cover song by ranking on the probability. Experimental results show that
the proposed algorithm achieves performance better than or comparable to the
state-of-the-art.
|
Personalized Gaussian Processes for Future Prediction of Alzheimer's
Disease Progression | In this paper, we introduce the use of a personalized Gaussian Process model
(pGP) to predict the key metrics of Alzheimer's Disease progression (MMSE,
ADAS-Cog13, CDRSB and CS) based on each patient's previous visits. We start by
learning a population-level model using multi-modal data from previously seen
patients using the base Gaussian Process (GP) regression. Then, this model is
adapted sequentially over time to a new patient using domain adaptive GPs to
form the patient's pGP. We show that this new approach, together with an
auto-regressive formulation, leads to significant improvements in forecasting
future clinical status and cognitive scores for target patients when compared
to modeling the population with traditional GPs.
|
Distributed Stratified Locality Sensitive Hashing for Critical Event
Prediction in the Cloud | The availability of massive healthcare data repositories calls for efficient
tools for data-driven medicine. We introduce a distributed system for
Stratified Locality Sensitive Hashing to perform fast similarity-based
prediction on large medical waveform datasets. Our implementation, for an ICU
use case, prioritizes latency over throughput and is targeted at a cloud
environment. We demonstrate our system on Acute Hypotensive Episode prediction
from Arterial Blood Pressure waveforms. On a dataset of $1.37$ million points,
we show scaling up to $40$ processors and a $21\times$ speedup in number of
comparisons to parallel exhaustive search at the price of a $10\%$ Matthews
correlation coefficient (MCC) loss. Furthermore, if additional MCC loss can be
tolerated, our system achieves speedups up to two orders of magnitude.
|
Deep Learning for Metagenomic Data: using 2D Embeddings and
Convolutional Neural Networks | Deep learning (DL) techniques have had unprecedented success when applied to
images, waveforms, and texts to cite a few. In general, when the sample size
(N) is much greater than the number of features (d), DL outperforms previous
machine learning (ML) techniques, often through the use of convolution neural
networks (CNNs). However, in many bioinformatics ML tasks, we encounter the
opposite situation where d is greater than N. In these situations, applying DL
techniques (such as feed-forward networks) would lead to severe overfitting.
Thus, sparse ML techniques (such as LASSO e.g.) usually yield the best results
on these tasks. In this paper, we show how to apply CNNs on data which do not
have originally an image structure (in particular on metagenomic data). Our
first contribution is to show how to map metagenomic data in a meaningful way
to 1D or 2D images. Based on this representation, we then apply a CNN, with the
aim of predicting various diseases. The proposed approach is applied on six
different datasets including in total over 1000 samples from various diseases.
This approach could be a promising one for prediction tasks in the
bioinformatics field.
|
Faithful Inversion of Generative Models for Effective Amortized
Inference | Inference amortization methods share information across multiple
posterior-inference problems, allowing each to be carried out more efficiently.
Generally, they require the inversion of the dependency structure in the
generative model, as the modeller must learn a mapping from observations to
distributions approximating the posterior. Previous approaches have involved
inverting the dependency structure in a heuristic way that fails to capture
these dependencies correctly, thereby limiting the achievable accuracy of the
resulting approximations. We introduce an algorithm for faithfully, and
minimally, inverting the graphical model structure of any generative model.
Such inverses have two crucial properties: (a) they do not encode any
independence assertions that are absent from the model and; (b) they are local
maxima for the number of true independencies encoded. We prove the correctness
of our approach and empirically show that the resulting minimally faithful
inverses lead to better inference amortization than existing heuristic
approaches.
|
Deep Learning with Permutation-invariant Operator for Multi-instance
Histopathology Classification | The computer-aided analysis of medical scans is a longstanding goal in the
medical imaging field. Currently, deep learning has became a dominant
methodology for supporting pathologists and radiologist. Deep learning
algorithms have been successfully applied to digital pathology and radiology,
nevertheless, there are still practical issues that prevent these tools to be
widely used in practice. The main obstacles are low number of available cases
and large size of images (a.k.a. the small n, large p problem in machine
learning), and a very limited access to annotation at a pixel level that can
lead to severe overfitting and large computational requirements. We propose to
handle these issues by introducing a framework that processes a medical image
as a collection of small patches using a single, shared neural network. The
final diagnosis is provided by combining scores of individual patches using a
permutation-invariant operator (combination). In machine learning community
such approach is called a multi-instance learning (MIL).
|
Semi-Adversarial Networks: Convolutional Autoencoders for Imparting
Privacy to Face Images | In this paper, we design and evaluate a convolutional autoencoder that
perturbs an input face image to impart privacy to a subject. Specifically, the
proposed autoencoder transforms an input face image such that the transformed
image can be successfully used for face recognition but not for gender
classification. In order to train this autoencoder, we propose a novel training
scheme, referred to as semi-adversarial training in this work. The training is
facilitated by attaching a semi-adversarial module consisting of a pseudo
gender classifier and a pseudo face matcher to the autoencoder. The objective
function utilized for training this network has three terms: one to ensure that
the perturbed image is a realistic face image; another to ensure that the
gender attributes of the face are confounded; and a third to ensure that
biometric recognition performance due to the perturbed image is not impacted.
Extensive experiments confirm the efficacy of the proposed architecture in
extending gender privacy to face images.
|
Group Sparse Bayesian Learning for Active Surveillance on Epidemic
Dynamics | Predicting epidemic dynamics is of great value in understanding and
controlling diffusion processes, such as infectious disease spread and
information propagation. This task is intractable, especially when surveillance
resources are very limited. To address the challenge, we study the problem of
active surveillance, i.e., how to identify a small portion of system components
as sentinels to effect monitoring, such that the epidemic dynamics of an entire
system can be readily predicted from the partial data collected by such
sentinels. We propose a novel measure, the gamma value, to identify the
sentinels by modeling a sentinel network with row sparsity structure. We design
a flexible group sparse Bayesian learning algorithm to mine the sentinel
network suitable for handling both linear and non-linear dynamical systems by
using the expectation maximization method and variational approximation. The
efficacy of the proposed algorithm is theoretically analyzed and empirically
validated using both synthetic and real-world data.
|
Enabling Embodied Analogies in Intelligent Music Systems | The present methodology is aimed at cross-modal machine learning and uses
multidisciplinary tools and methods drawn from a broad range of areas and
disciplines, including music, systematic musicology, dance, motion capture,
human-computer interaction, computational linguistics and audio signal
processing. Main tasks include: (1) adapting wisdom-of-the-crowd approaches to
embodiment in music and dance performance to create a dataset of music and
music lyrics that covers a variety of emotions, (2) applying
audio/language-informed machine learning techniques to that dataset to identify
automatically the emotional content of the music and the lyrics, and (3)
integrating motion capture data from a Vicon system and dancers performing on
that music.
|
Deep Neural Network Detects Quantum Phase Transition | We detect the quantum phase transition of a quantum many-body system by
mapping the observed results of the quantum state onto a neural network. In the
present study, we utilized the simplest case of a quantum many-body system,
namely a one-dimensional chain of Ising spins with the transverse Ising model.
We prepared several spin configurations, which were obtained using repeated
observations of the model for a particular strength of the transverse field, as
input data for the neural network. Although the proposed method can be employed
using experimental observations of quantum many-body systems, we tested our
technique with spin configurations generated by a quantum Monte Carlo
simulation without initial relaxation. The neural network successfully
classified the strength of transverse field only from the spin configurations,
leading to consistent estimations of the critical point of our model $\Gamma_c
=J$.
|
Don't Just Assume; Look and Answer: Overcoming Priors for Visual
Question Answering | A number of studies have found that today's Visual Question Answering (VQA)
models are heavily driven by superficial correlations in the training data and
lack sufficient image grounding. To encourage development of models geared
towards the latter, we propose a new setting for VQA where for every question
type, train and test sets have different prior distributions of answers.
Specifically, we present new splits of the VQA v1 and VQA v2 datasets, which we
call Visual Question Answering under Changing Priors (VQA-CP v1 and VQA-CP v2
respectively). First, we evaluate several existing VQA models under this new
setting and show that their performance degrades significantly compared to the
original VQA setting. Second, we propose a novel Grounded Visual Question
Answering model (GVQA) that contains inductive biases and restrictions in the
architecture specifically designed to prevent the model from 'cheating' by
primarily relying on priors in the training data. Specifically, GVQA explicitly
disentangles the recognition of visual concepts present in the image from the
identification of plausible answer space for a given question, enabling the
model to more robustly generalize across different distributions of answers.
GVQA is built off an existing VQA model -- Stacked Attention Networks (SAN).
Our experiments demonstrate that GVQA significantly outperforms SAN on both
VQA-CP v1 and VQA-CP v2 datasets. Interestingly, it also outperforms more
powerful VQA models such as Multimodal Compact Bilinear Pooling (MCB) in
several cases. GVQA offers strengths complementary to SAN when trained and
evaluated on the original VQA v1 and VQA v2 datasets. Finally, GVQA is more
transparent and interpretable than existing VQA models.
|
Time Limits in Reinforcement Learning | In reinforcement learning, it is common to let an agent interact for a fixed
amount of time with its environment before resetting it and repeating the
process in a series of episodes. The task that the agent has to learn can
either be to maximize its performance over (i) that fixed period, or (ii) an
indefinite period where time limits are only used during training to diversify
experience. In this paper, we provide a formal account for how time limits
could effectively be handled in each of the two cases and explain why not doing
so can cause state aliasing and invalidation of experience replay, leading to
suboptimal policies and training instability. In case (i), we argue that the
terminations due to time limits are in fact part of the environment, and thus a
notion of the remaining time should be included as part of the agent's input to
avoid violation of the Markov property. In case (ii), the time limits are not
part of the environment and are only used to facilitate learning. We argue that
this insight should be incorporated by bootstrapping from the value of the
state at the end of each partial episode. For both cases, we illustrate
empirically the significance of our considerations in improving the performance
and stability of existing reinforcement learning algorithms, showing
state-of-the-art results on several control tasks.
|
Probabilistic Adaptive Computation Time | We present a probabilistic model with discrete latent variables that control
the computation time in deep learning models such as ResNets and LSTMs. A prior
on the latent variables expresses the preference for faster computation. The
amount of computation for an input is determined via amortized maximum a
posteriori (MAP) inference. MAP inference is performed using a novel stochastic
variational optimization method. The recently proposed Adaptive Computation
Time mechanism can be seen as an ad-hoc relaxation of this model. We
demonstrate training using the general-purpose Concrete relaxation of discrete
variables. Evaluation on ResNet shows that our method matches the
speed-accuracy trade-off of Adaptive Computation Time, while allowing for
evaluation with a simple deterministic procedure that has a lower memory
footprint.
|
Deep Learning Scaling is Predictable, Empirically | Deep learning (DL) creates impactful advances following a virtuous recipe:
model architecture search, creating large training data sets, and scaling
computation. It is widely believed that growing training sets and models should
improve accuracy and result in better products. As DL application domains grow,
we would like a deeper understanding of the relationships between training set
size, computational scale, and model accuracy improvements to advance the
state-of-the-art.
This paper presents a large scale empirical characterization of
generalization error and model size growth as training sets grow. We introduce
a methodology for this measurement and test four machine learning domains:
machine translation, language modeling, image processing, and speech
recognition. Our empirical results show power-law generalization error scaling
across a breadth of factors, resulting in power-law exponents---the "steepness"
of the learning curve---yet to be explained by theoretical work. Further, model
improvements only shift the error but do not appear to affect the power-law
exponent. We also show that model size scales sublinearly with data size. These
scaling relationships have significant implications on deep learning research,
practice, and systems. They can assist model debugging, setting accuracy
targets, and decisions about data set growth. They can also guide computing
system design and underscore the importance of continued computational scaling.
|
The reparameterization trick for acquisition functions | Bayesian optimization is a sample-efficient approach to solving global
optimization problems. Along with a surrogate model, this approach relies on
theoretically motivated value heuristics (acquisition functions) to guide the
search process. Maximizing acquisition functions yields the best performance;
unfortunately, this ideal is difficult to achieve since optimizing acquisition
functions per se is frequently non-trivial. This statement is especially true
in the parallel setting, where acquisition functions are routinely non-convex,
high-dimensional, and intractable. Here, we demonstrate how many popular
acquisition functions can be formulated as Gaussian integrals amenable to the
reparameterization trick and, ensuingly, gradient-based optimization. Further,
we use this reparameterized representation to derive an efficient Monte Carlo
estimator for the upper confidence bound acquisition function in the context of
parallel selection.
|
Deep Neural Network Architectures for Modulation Classification | In this work, we investigate the value of employing deep learning for the
task of wireless signal modulation recognition. Recently in [1], a framework
has been introduced by generating a dataset using GNU radio that mimics the
imperfections in a real wireless channel, and uses 10 different modulation
types. Further, a convolutional neural network (CNN) architecture was developed
and shown to deliver performance that exceeds that of expert-based approaches.
Here, we follow the framework of [1] and find deep neural network architectures
that deliver higher accuracy than the state of the art. We tested the
architecture of [1] and found it to achieve an accuracy of approximately 75% of
correctly recognizing the modulation type. We first tune the CNN architecture
of [1] and find a design with four convolutional layers and two dense layers
that gives an accuracy of approximately 83.8% at high SNR. We then develop
architectures based on the recently introduced ideas of Residual Networks
(ResNet [2]) and Densely Connected Networks (DenseNet [3]) to achieve high SNR
accuracies of approximately 83.5% and 86.6%, respectively. Finally, we
introduce a Convolutional Long Short-term Deep Neural Network (CLDNN [4]) to
achieve an accuracy of approximately 88.5% at high SNR.
|
Subject Selection on a Riemannian Manifold for Unsupervised
Cross-subject Seizure Detection | Inter-subject variability between individuals poses a challenge in
inter-subject brain signal analysis problems. A new algorithm for
subject-selection based on clustering covariance matrices on a Riemannian
manifold is proposed. After unsupervised selection of the subsets of relevant
subjects, data in a cluster is mapped to a tangent space at the mean point of
covariance matrices in that cluster and an SVM classifier on labeled data from
relevant subjects is trained. Experiment on an EEG seizure database shows that
the proposed method increases the accuracy over state-of-the-art from 86.83% to
89.84% and specificity from 87.38% to 89.64% while reducing the false positive
rate/hour from 0.8/hour to 0.77/hour.
|
Intelligent EHRs: Predicting Procedure Codes From Diagnosis Codes | In order to submit a claim to insurance companies, a doctor needs to code a
patient encounter with both the diagnosis (ICDs) and procedures performed
(CPTs) in an Electronic Health Record (EHR). Identifying and applying relevant
procedures code is a cumbersome and time-consuming task as a doctor has to
choose from around 13,000 procedure codes with no predefined one-to-one
mapping. In this paper, we propose a state-of-the-art deep learning method for
automatic and intelligent coding of procedures (CPTs) from the diagnosis codes
(ICDs) entered by the doctor. Precisely, we cast the learning problem as a
multi-label classification problem and use distributed representation to learn
the input mapping of high-dimensional sparse ICDs codes. Our final model
trained on 2.3 million claims is able to outperform existing rule-based
probabilistic and association-rule mining based methods and has a recall of
90@3.
|
Visual Features for Context-Aware Speech Recognition | Automatic transcriptions of consumer-generated multi-media content such as
"Youtube" videos still exhibit high word error rates. Such data typically
occupies a very broad domain, has been recorded in challenging conditions, with
cheap hardware and a focus on the visual modality, and may have been
post-processed or edited. In this paper, we extend our earlier work on adapting
the acoustic model of a DNN-based speech recognition system to an RNN language
model and show how both can be adapted to the objects and scenes that can be
automatically detected in the video. We are working on a corpus of "how-to"
videos from the web, and the idea is that an object that can be seen ("car"),
or a scene that is being detected ("kitchen") can be used to condition both
models on the "context" of the recording, thereby reducing perplexity and
improving transcription. We achieve good improvements in both cases and compare
and analyze the respective reductions in word error rate. We expect that our
results can be used for any type of speech processing in which "context"
information is available, for example in robotics, man-machine interaction, or
when indexing large audio-visual archives, and should ultimately help to bring
together the "video-to-text" and "speech-to-text" communities.
|
Prediction-Constrained Topic Models for Antidepressant Recommendation | Supervisory signals can help topic models discover low-dimensional data
representations that are more interpretable for clinical tasks. We propose a
framework for training supervised latent Dirichlet allocation that balances two
goals: faithful generative explanations of high-dimensional data and accurate
prediction of associated class labels. Existing approaches fail to balance
these goals by not properly handling a fundamental asymmetry: the intended task
is always predicting labels from data, not data from labels. Our new
prediction-constrained objective trains models that predict labels from heldout
data well while also producing good generative likelihoods and interpretable
topic-word parameters. In a case study on predicting depression medications
from electronic health records, we demonstrate improved recommendations
compared to previous supervised topic models and high- dimensional logistic
regression from words alone.
|
A Neural Stochastic Volatility Model | In this paper, we show that the recent integration of statistical models with
deep recurrent neural networks provides a new way of formulating volatility
(the degree of variation of time series) models that have been widely used in
time series analysis and prediction in finance. The model comprises a pair of
complementary stochastic recurrent neural networks: the generative network
models the joint distribution of the stochastic volatility process; the
inference network approximates the conditional distribution of the latent
variables given the observables. Our focus here is on the formulation of
temporal dynamics of volatility over time under a stochastic recurrent neural
network framework. Experiments on real-world stock price datasets demonstrate
that the proposed model generates a better volatility estimation and prediction
that outperforms mainstream methods, e.g., deterministic models such as GARCH
and its variants, and stochastic models namely the MCMC-based model
\emph{stochvol} as well as the Gaussian process volatility model \emph{GPVol},
on average negative log-likelihood.
|
An Elementary Analysis of the Probability That a Binomial Random
Variable Exceeds Its Expectation | We give an elementary proof of the fact that a binomial random variable $X$
with parameters $n$ and $0.29/n \le p < 1$ with probability at least $1/4$
strictly exceeds its expectation. We also show that for $1/n \le p < 1 - 1/n$,
$X$ exceeds its expectation by more than one with probability at least
$0.0370$. Both probabilities approach $1/2$ when $np$ and $n(1-p)$ tend to
infinity.
|
Adaptive Sampled Softmax with Kernel Based Sampling | Softmax is the most commonly used output function for multiclass problems and
is widely used in areas such as vision, natural language processing, and
recommendation. A softmax model has linear costs in the number of classes which
makes it too expensive for many real-world problems. A common approach to speed
up training involves sampling only some of the classes at each training step.
It is known that this method is biased and that the bias increases the more the
sampling distribution deviates from the output distribution. Nevertheless,
almost any recent work uses simple sampling distributions that require a large
sample size to mitigate the bias. In this work, we propose a new class of
kernel based sampling methods and develop an efficient sampling algorithm.
Kernel based sampling adapts to the model as it is trained, thus resulting in
low bias. Kernel based sampling can be easily applied to many models because it
relies only on the model's last hidden layer. We empirically study the
trade-off of bias, sampling distribution and sample size and show that kernel
based sampling results in low bias with few samples.
|
A global feature extraction model for the effective computer aided
diagnosis of mild cognitive impairment using structural MRI images | Multiple modalities of biomarkers have been proved to be very sensitive in
assessing the progression of Alzheimer's disease (AD), and using these
modalities and machine learning algorithms, several approaches have been
proposed to assist in the early diagnosis of AD. Among the recent investigated
state-of-the-art approaches, Gaussian discriminant analysis (GDA)-based
approaches have been demonstrated to be more effective and accurate in the
classification of AD, especially for delineating its prodromal stage of mild
cognitive impairment (MCI). Moreover, among those binary classification
investigations, the local feature extraction methods were mostly used, which
made them hardly be applied to a practical computer aided diagnosis system.
Therefore, this study presents a novel global feature extraction model taking
advantage of the recent proposed GDA-based dual high-dimensional decision
spaces, which can significantly improve the early diagnosis performance
comparing to those local feature extraction methods. In the true test using 20%
held-out data, for discriminating the most challenging MCI group from the
cognitively normal control (CN) group, an F1 score of 91.06%, an accuracy of
88.78%, a sensitivity of 91.80%, and a specificity of 83.78% were achieved that
can be considered as the best performance obtained so far.
|
Where Classification Fails, Interpretation Rises | An intriguing property of deep neural networks is their inherent
vulnerability to adversarial inputs, which significantly hinders their
application in security-critical domains. Most existing detection methods
attempt to use carefully engineered patterns to distinguish adversarial inputs
from their genuine counterparts, which however can often be circumvented by
adaptive adversaries. In this work, we take a completely different route by
leveraging the definition of adversarial inputs: while deceiving for deep
neural networks, they are barely discernible for human visions. Building upon
recent advances in interpretable models, we construct a new detection framework
that contrasts an input's interpretation against its classification. We
validate the efficacy of this framework through extensive experiments using
benchmark datasets and attacks. We believe that this work opens a new direction
for designing adversarial input detection methods.
|
Progressive Neural Architecture Search | We propose a new method for learning the structure of convolutional neural
networks (CNNs) that is more efficient than recent state-of-the-art methods
based on reinforcement learning and evolutionary algorithms. Our approach uses
a sequential model-based optimization (SMBO) strategy, in which we search for
structures in order of increasing complexity, while simultaneously learning a
surrogate model to guide the search through structure space. Direct comparison
under the same search space shows that our method is up to 5 times more
efficient than the RL method of Zoph et al. (2018) in terms of number of models
evaluated, and 8 times faster in terms of total compute. The structures we
discover in this way achieve state of the art classification accuracies on
CIFAR-10 and ImageNet.
|
Anesthesiologist-level forecasting of hypoxemia with only SpO2 data
using deep learning | We use a deep learning model trained only on a patient's blood oxygenation
data (measurable with an inexpensive fingertip sensor) to predict impending
hypoxemia (low blood oxygen) more accurately than trained anesthesiologists
with access to all the data recorded in a modern operating room. We also
provide a simple way to visualize the reason why a patient's risk is low or
high by assigning weight to the patient's past blood oxygen values. This work
has the potential to provide cutting-edge clinical decision support in
low-resource settings, where rates of surgical complication and death are
substantially greater than in high-resource areas.
|
Supervised Hashing based on Energy Minimization | Recently, supervised hashing methods have attracted much attention since they
can optimize retrieval speed and storage cost while preserving semantic
information. Because hashing codes learning is NP-hard, many methods resort to
some form of relaxation technique. But the performance of these methods can
easily deteriorate due to the relaxation. Luckily, many supervised hashing
formulations can be viewed as energy functions, hence solving hashing codes is
equivalent to learning marginals in the corresponding conditional random field
(CRF). By minimizing the KL divergence between a fully factorized distribution
and the Gibbs distribution of this CRF, a set of consistency equations can be
obtained, but updating them in parallel may not yield a local optimum since the
variational lower bound is not guaranteed to increase. In this paper, we use a
linear approximation of the sigmoid function to convert these consistency
equations to linear systems, which have a closed-form solution. By applying
this novel technique to two classical hashing formulations KSH and SPLH, we
obtain two new methods called EM (energy minimizing based)-KSH and EM-SPLH.
Experimental results on three datasets show the superiority of our methods.
|
Tracking the Best Expert in Non-stationary Stochastic Environments | We study the dynamic regret of multi-armed bandit and experts problem in
non-stationary stochastic environments. We introduce a new parameter $\Lambda$,
which measures the total statistical variance of the loss distributions over
$T$ rounds of the process, and study how this amount affects the regret. We
investigate the interaction between $\Lambda$ and $\Gamma$, which counts the
number of times the distributions change, as well as $\Lambda$ and $V$, which
measures how far the distributions deviates over time. One striking result we
find is that even when $\Gamma$, $V$, and $\Lambda$ are all restricted to
constant, the regret lower bound in the bandit setting still grows with $T$.
The other highlight is that in the full-information setting, a constant regret
becomes achievable with constant $\Gamma$ and $\Lambda$, as it can be made
independent of $T$, while with constant $V$ and $\Lambda$, the regret still has
a $T^{1/3}$ dependency. We not only propose algorithms with upper bound
guarantee, but prove their matching lower bounds as well.
|
Online Reinforcement Learning in Stochastic Games | We study online reinforcement learning in average-reward stochastic games
(SGs). An SG models a two-player zero-sum game in a Markov environment, where
state transitions and one-step payoffs are determined simultaneously by a
learner and an adversary. We propose the UCSG algorithm that achieves a
sublinear regret compared to the game value when competing with an arbitrary
opponent. This result improves previous ones under the same setting. The regret
bound has a dependency on the diameter, which is an intrinsic value related to
the mixing property of SGs. If we let the opponent play an optimistic best
response to the learner, UCSG finds an $\varepsilon$-maximin stationary policy
with a sample complexity of
$\tilde{\mathcal{O}}\left(\text{poly}(1/\varepsilon)\right)$, where
$\varepsilon$ is the gap to the best policy.
|
MAgent: A Many-Agent Reinforcement Learning Platform for Artificial
Collective Intelligence | We introduce MAgent, a platform to support research and development of
many-agent reinforcement learning. Unlike previous research platforms on single
or multi-agent reinforcement learning, MAgent focuses on supporting the tasks
and the applications that require hundreds to millions of agents. Within the
interactions among a population of agents, it enables not only the study of
learning algorithms for agents' optimal polices, but more importantly, the
observation and understanding of individual agent's behaviors and social
phenomena emerging from the AI society, including communication languages,
leaderships, altruism. MAgent is highly scalable and can host up to one million
agents on a single GPU server. MAgent also provides flexible configurations for
AI researchers to design their customized environments and agents. In this
demo, we present three environments designed on MAgent and show emerged
collective intelligence by learning from scratch.
|
PFAx: Predictable Feature Analysis to Perform Control | Predictable Feature Analysis (PFA) (Richthofer, Wiskott, ICMLA 2015) is an
algorithm that performs dimensionality reduction on high dimensional input
signal. It extracts those subsignals that are most predictable according to a
certain prediction model. We refer to these extracted signals as predictable
features.
In this work we extend the notion of PFA to take supplementary information
into account for improving its predictions. Such information can be a
multidimensional signal like the main input to PFA, but is regarded external.
That means it won't participate in the feature extraction - no features get
extracted or composed of it. Features will be exclusively extracted from the
main input such that they are most predictable based on themselves and the
supplementary information. We refer to this enhanced PFA as PFAx (PFA
extended).
Even more important than improving prediction quality is to observe the
effect of supplementary information on feature selection. PFAx transparently
provides insight how the supplementary information adds to prediction quality
and whether it is valuable at all. Finally we show how to invert that relation
and can generate the supplementary information such that it would yield a
certain desired outcome of the main signal.
We apply this to a setting inspired by reinforcement learning and let the
algorithm learn how to control an agent in an environment. With this method it
is feasible to locally optimize the agent's state, i.e. reach a certain goal
that is near enough. We are preparing a follow-up paper that extends this
method such that also global optimization is feasible.
|
Network Coding Based Evolutionary Network Formation for Dynamic Wireless
Networks | In this paper, we aim to find a robust network formation strategy that can
adaptively evolve the network topology against network dynamics in a
distributed manner. We consider a network coding deployed wireless ad hoc
network where source nodes are connected to terminal nodes with the help of
intermediate nodes. We show that mixing operations in network coding can induce
packet anonymity that allows the inter-connections in a network to be
decoupled. This enables each intermediate node to consider complex network
inter-connections as a node-environment interaction such that the Markov
decision process (MDP) can be employed at each intermediate node. The optimal
policy that can be obtained by solving the MDP provides each node with optimal
amount of changes in transmission range given network dynamics (e.g., the
number of nodes in the range and channel condition). Hence, the network can be
adaptively and optimally evolved by responding to the network dynamics. The
proposed strategy is used to maximize long-term utility, which is achieved by
considering both current network conditions and future network dynamics. We
define the utility of an action to include network throughput gain and the cost
of transmission power. We show that the resulting network of the proposed
strategy eventually converges to stationary networks, which maintain the states
of the nodes. Moreover, we propose to determine initial transmission ranges and
initial network topology that can expedite the convergence of the proposed
algorithm. Our simulation results confirm that the proposed strategy builds a
network which adaptively changes its topology in the presence of network
dynamics. Moreover, the proposed strategy outperforms existing strategies in
terms of system goodput and successful connectivity ratio.
|
Learning Sparse Adversarial Dictionaries For Multi-Class Audio
Classification | Audio events are quite often overlapping in nature, and more prone to noise
than visual signals. There has been increasing evidence for the superior
performance of representations learned using sparse dictionaries for
applications like audio denoising and speech enhancement. This paper
concentrates on modifying the traditional reconstructive dictionary learning
algorithms, by incorporating a discriminative term into the objective function
in order to learn class-specific adversarial dictionaries that are good at
representing samples of their own class at the same time poor at representing
samples belonging to any other class. We quantitatively demonstrate the
effectiveness of our learned dictionaries as a stand-alone solution for both
binary as well as multi-class audio classification problems.
|
Short-term Mortality Prediction for Elderly Patients Using Medicare
Claims Data | Risk prediction is central to both clinical medicine and public health. While
many machine learning models have been developed to predict mortality, they are
rarely applied in the clinical literature, where classification tasks typically
rely on logistic regression. One reason for this is that existing machine
learning models often seek to optimize predictions by incorporating features
that are not present in the databases readily available to providers and policy
makers, limiting generalizability and implementation. Here we tested a number
of machine learning classifiers for prediction of six-month mortality in a
population of elderly Medicare beneficiaries, using an administrative claims
database of the kind available to the majority of health care payers and
providers. We show that machine learning classifiers substantially outperform
current widely-used methods of risk prediction but only when used with an
improved feature set incorporating insights from clinical medicine, developed
for this study. Our work has applications to supporting patient and provider
decision making at the end of life, as well as population health-oriented
efforts to identify patients at high risk of poor outcomes.
|
Representation and Reinforcement Learning for Personalized Glycemic
Control in Septic Patients | Glycemic control is essential for critical care. However, it is a challenging
task because there has been no study on personalized optimal strategies for
glycemic control. This work aims to learn personalized optimal glycemic
trajectories for severely ill septic patients by learning data-driven policies
to identify optimal targeted blood glucose levels as a reference for
clinicians. We encoded patient states using a sparse autoencoder and adopted a
reinforcement learning paradigm using policy iteration to learn the optimal
policy from data. We also estimated the expected return following the policy
learned from the recorded glycemic trajectories, which yielded a function
indicating the relationship between real blood glucose values and 90-day
mortality rates. This suggests that the learned optimal policy could reduce the
patients' estimated 90-day mortality rate by 6.3%, from 31% to 24.7%. The
result demonstrates that reinforcement learning with appropriate patient state
encoding can potentially provide optimal glycemic trajectories and allow
clinicians to design a personalized strategy for glycemic control in septic
patients.
|
An Asymptotically Optimal Algorithm for Communicating Multiplayer
Multi-Armed Bandit Problems | We consider a decentralized stochastic multi-armed bandit problem with
multiple players. Each player aims to maximize his/her own reward by pulling an
arm. The arms give rewards based on i.i.d. stochastic Bernoulli distributions.
Players are not aware about the probability distributions of the arms. At the
end of each turn, the players inform their neighbors about the arm he/she
pulled and the reward he/she got. Neighbors of players are determined according
to an Erd{\H{o}}s-R{\'e}nyi graph with connectivity $\alpha$. This graph is
reproduced in the beginning of every turn with the same connectivity. When more
than one player choose the same arm in a turn, we assume that only one of the
players who is randomly chosen gets the reward where the others get nothing. We
first start by assuming players are not aware of the collision model and offer
an asymptotically optimal algorithm for $\alpha = 1$ case. Then, we extend our
prior work and offer an asymptotically optimal algorithm for any connectivity
but zero, assuming players aware of the collision model. We also study the
effect of $\alpha$, the degree of communication between players, empirically on
the cumulative regret by comparing them with traditional multi-armed bandit
algorithms.
|
Mix-and-Match Tuning for Self-Supervised Semantic Segmentation | Deep convolutional networks for semantic image segmentation typically require
large-scale labeled data, e.g. ImageNet and MS COCO, for network pre-training.
To reduce annotation efforts, self-supervised semantic segmentation is recently
proposed to pre-train a network without any human-provided labels. The key of
this new form of learning is to design a proxy task (e.g. image colorization),
from which a discriminative loss can be formulated on unlabeled data. Many
proxy tasks, however, lack the critical supervision signals that could induce
discriminative representation for the target image segmentation task. Thus
self-supervision's performance is still far from that of supervised
pre-training. In this study, we overcome this limitation by incorporating a
"mix-and-match" (M&M) tuning stage in the self-supervision pipeline. The
proposed approach is readily pluggable to many self-supervision methods and
does not use more annotated samples than the original process. Yet, it is
capable of boosting the performance of target image segmentation task to
surpass fully-supervised pre-trained counterpart. The improvement is made
possible by better harnessing the limited pixel-wise annotations in the target
dataset. Specifically, we first introduce the "mix" stage, which sparsely
samples and mixes patches from the target set to reflect rich and diverse local
patch statistics of target images. A "match" stage then forms a class-wise
connected graph, which can be used to derive a strong triplet-based
discriminative loss for fine-tuning the network. Our paradigm follows the
standard practice in existing self-supervised studies and no extra data or
label is required. With the proposed M&M approach, for the first time, a
self-supervision method can achieve comparable or even better performance
compared to its ImageNet pre-trained counterpart on both PASCAL VOC2012 dataset
and CityScapes dataset.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.