title
stringlengths 6
244
| abstract
stringlengths 19
5.09k
| label
stringclasses 10
values |
|---|---|---|
Causal Inference Using Linear Time-Varying Filters with Additive Noise | Causal inference using the restricted structural causal model framework
hinges largely on the asymmetry between cause and effect from the data
generating mechanisms. For linear non-Gaussian noise models and nonlinear
additive noise models, the asymmetry arises from non-Gaussianity or
nonlinearity, respectively. Despite the fact that this methodology can be
adapted to stationary time series, inferring causal relationships from
nonstationary time series remains a challenging task. In this work, we focus on
slowly-varying nonstationary processes and propose to break the symmetry by
exploiting the nonstationarity of the data. Our main theoretical result shows
that the causal direction is identifiable in generic cases when cause and
effect are connected via a time-varying filter. We propose a causal discovery
procedure by leveraging powerful estimates of the bivariate evolutionary
spectra. Both synthetic and real-world data simulations that involve high-order
and non-smooth filters are provided to demonstrate the effectiveness of our
proposed methodology.
| stat |
Quantized Fisher Discriminant Analysis | This paper proposes a new subspace learning method, named Quantized Fisher
Discriminant Analysis (QFDA), which makes use of both machine learning and
information theory. There is a lack of literature for combination of machine
learning and information theory and this paper tries to tackle this gap. QFDA
finds a subspace which discriminates the uniformly quantized images in the
Discrete Cosine Transform (DCT) domain at least as well as discrimination of
non-quantized images by Fisher Discriminant Analysis (FDA) while the images
have been compressed. This helps the user to throw away the original images and
keep the compressed images instead without noticeable loss of classification
accuracy. We propose a cost function whose minimization can be interpreted as
rate-distortion optimization in information theory. We also propose quantized
Fisherfaces for facial analysis in QFDA. Our experiments on AT&T face dataset
and Fashion MNIST dataset show the effectiveness of this subspace learning
method.
| stat |
Quality analysis in acyclic production networks | The production network under examination consists of a number of
workstations. Each workstation is a parallel configuration of machines
performing the same kind of tasks on a given part. Parts move from one
workstation to another and at each workstation a part is assigned randomly to a
machine. We assume that the production network is acyclic, that is, a part does
not return to a workstation where it previously received service. Furthermore,
we assume that the quality of the end product is additive, that is, the sum of
the quality contributions of the machines along the production path. The
contribution of each machine is modeled by a separate random variable.
Our main result is the construction of estimators that allow pairwise and
multiple comparison of the means and variances of machines in the same
workstation. These comparisons then may lead to the identification of
unreliable machines. We also discuss the asymptotic distributions of the
estimators that allow the use of standard statistical tests and decision
making.
| stat |
Landscape Complexity for the Empirical Risk of Generalized Linear Models | We present a method to obtain the average and the typical value of the number
of critical points of the empirical risk landscape for generalized linear
estimation problems and variants. This represents a substantial extension of
previous applications of the Kac-Rice method since it allows to analyze the
critical points of high dimensional non-Gaussian random functions. We obtain a
rigorous explicit variational formula for the annealed complexity, which is the
logarithm of the average number of critical points at fixed value of the
empirical risk. This result is simplified, and extended, using the non-rigorous
Kac-Rice replicated method from theoretical physics. In this way we find an
explicit variational formula for the quenched complexity, which is generally
different from its annealed counterpart, and allows to obtain the number of
critical points for typical instances up to exponential accuracy.
| stat |
A new SVD approach to optimal topic estimation | In the probabilistic topic models, the quantity of interest---a low-rank
matrix consisting of topic vectors---is hidden in the text corpus matrix,
masked by noise, and Singular Value Decomposition (SVD) is a potentially useful
tool for learning such a matrix. However, different rows and columns of the
matrix are usually in very different scales and the connection between this
matrix and the singular vectors of the text corpus matrix are usually
complicated and hard to spell out, so how to use SVD for learning topic models
faces challenges.
We overcome the challenges by introducing a proper Pre-SVD normalization of
the text corpus matrix and a proper column-wise scaling for the matrix of
interest, and by revealing a surprising Post-SVD low-dimensional {\it simplex}
structure. The simplex structure, together with the Pre-SVD normalization and
column-wise scaling, allows us to conveniently reconstruct the matrix of
interest, and motivates a new SVD-based approach to learning topic models.
We show that under the popular probabilistic topic model \citep{hofmann1999},
our method has a faster rate of convergence than existing methods in a wide
variety of cases. In particular, for cases where documents are long or $n$ is
much larger than $p$, our method achieves the optimal rate. At the heart of the
proofs is a tight element-wise bound on singular vectors of a multinomially
distributed data matrix, which do not exist in literature and we have to derive
by ourself.
We have applied our method to two data sets, Associated Process (AP) and
Statistics Literature Abstract (SLA), with encouraging results. In particular,
there is a clear simplex structure associated with the SVD of the data
matrices, which largely validates our discovery.
| stat |
Performance of variable and function selection methods for estimating
the non-linear health effects of correlated chemical mixtures: a simulation
study | Statistical methods for identifying harmful chemicals in a correlated mixture
often assume linearity in exposure-response relationships. Non-monotonic
relationships are increasingly recognised (e.g., for endocrine-disrupting
chemicals); however, the impact of non-monotonicity on exposure selection has
not been evaluated. In a simulation study, we assessed the performance of
Bayesian kernel machine regression (BKMR), Bayesian additive regression trees
(BART), Bayesian structured additive regression with spike-slab priors
(BSTARSS), and lasso penalised regression. We used data on exposure to 12
phthalates and phenols in pregnant women from the U.S. National Health and
Nutrition Examination Survey to simulate realistic exposure data using a
multivariate copula. We simulated datasets of size N = 250 and compared methods
across 32 scenarios, varying by model size and sparsity, signal-to-noise ratio,
correlation structure, and exposure-response relationship shapes. We compared
methods in terms of their sensitivity, specificity, and estimation accuracy. In
most scenarios, BKMR and BSTARSS achieved moderate to high specificity
(0.56--0.91 and 0.57--0.96, respectively) and sensitivity (0.49--0.98 and
0.25--0.97, respectively). BART achieved high specificity ($\geq$ 0.96), but
low to moderate sensitivity (0.13--0.66). Lasso was highly sensitive
(0.75--0.99), except for symmetric inverse-U-shaped relationships ($\leq$ 0.2).
Performance was affected by the signal-to-noise ratio, but not substantially by
the correlation structure. Penalised regression methods that assume linearity,
such as lasso, may not be suitable for studies of environmental chemicals
hypothesised to have non-monotonic relationships with outcomes. Instead, BKMR
and BSTARSS are attractive methods for flexibly estimating the shapes of
exposure-response relationships and selecting among correlated exposures.
| stat |
Point Process Regression | Point processes in time have a wide range of applications that include the
claims arrival process in insurance or the analysis of queues in operations
research. Due to advances in technology, such samples of point processes are
increasingly encountered. A key object of interest is the local intensity
function. It has a straightforward interpretation that allows to understand and
explore point process data. We consider functional approaches for point
processes, where one has a sample of repeated realizations of the point
process. This situation is inherently connected with Cox processes, where the
intensity functions of the replications are modeled as random functions. Here
we study a situation where one records covariates for each replication of the
process, such as the daily temperature for bike rentals. For modeling point
processes as responses with vector covariates as predictors we propose a novel
regression approach for the intensity function that is intrinsically
nonparametric. While the intensity function of a point process that is only
observed once on a fixed domain cannot be identified, we show how covariates
and repeated observations of the process can be utilized to make consistent
estimation possible, and we also derive asymptotic rates of convergence without
invoking parametric assumptions.
| stat |
Causal Inference in high dimensions: A marriage between Bayesian
modeling and good frequentist properties | We introduce a framework for estimating causal effects of binary and
continuous treatments in high dimensions. We show how posterior distributions
of treatment and outcome models can be used together with doubly robust
estimators. We propose an approach to uncertainty quantification for the doubly
robust estimator which utilizes posterior distributions of model parameters and
(1) results in good frequentist properties in small samples, (2) is based on a
single MCMC, and (3) improves over frequentist measures of uncertainty which
rely on asymptotic properties. We show that our proposed variance estimation
strategy is consistent when both models are correctly specified and that it is
conservative in finite samples or when one or both models are misspecified. We
consider a flexible framework for modeling the treatment and outcome processes
within the Bayesian paradigm that reduces model dependence, accommodates
nonlinearity, and achieves dimension reduction of the covariate space. We
illustrate the ability of the proposed approach to flexibly estimate causal
effects in high dimensions and appropriately quantify uncertainty, and show
that it performs well relative to existing approaches. Finally, we estimate the
effect of continuous environmental exposures on cholesterol and triglyceride
levels. An R package is available at github.com/jantonelli111/DoublyRobustHD.
| stat |
A Law of Robustness for Weight-bounded Neural Networks | Robustness of deep neural networks against adversarial perturbations is a
pressing concern motivated by recent findings showing the pervasive nature of
such vulnerabilities. One method of characterizing the robustness of a neural
network model is through its Lipschitz constant, which forms a robustness
certificate. A natural question to ask is, for a fixed model class (such as
neural networks) and a dataset of size $n$, what is the smallest achievable
Lipschitz constant among all models that fit the dataset? Recently, (Bubeck et
al., 2020) conjectured that when using two-layer networks with $k$ neurons to
fit a generic dataset, the smallest Lipschitz constant is
$\Omega(\sqrt{\frac{n}{k}})$. This implies that one would require one neuron
per data point to robustly fit the data. In this work we derive a lower bound
on the Lipschitz constant for any arbitrary model class with bounded Rademacher
complexity. Our result coincides with that conjectured in (Bubeck et al., 2020)
for two-layer networks under the assumption of bounded weights. However, due to
our result's generality, we also derive bounds for multi-layer neural networks,
discovering that one requires $\log n$ constant-sized layers to robustly fit
the data. Thus, our work establishes a law of robustness for weight bounded
neural networks and provides formal evidence on the necessity of
over-parametrization in deep learning.
| stat |
Exclusive Topic Modeling | We propose an Exclusive Topic Modeling (ETM) for unsupervised text
classification, which is able to 1) identify the field-specific keywords though
less frequently appeared and 2) deliver well-structured topics with exclusive
words. In particular, a weighted Lasso penalty is imposed to reduce the
dominance of the frequently appearing yet less relevant words automatically,
and a pairwise Kullback-Leibler divergence penalty is used to implement topics
separation. Simulation studies demonstrate that the ETM detects the
field-specific keywords, while LDA fails. When applying to the benchmark NIPS
dataset, the topic coherence score on average improves by 22% and 10% for the
model with weighted Lasso penalty and pairwise Kullback-Leibler divergence
penalty, respectively.
| stat |
Robustness on Networks | We adopt the statistical framework on robustness proposed by Watson and
Holmes in 2016 and then tackle the practical challenges that hinder its
applicability to network models. The goal is to evaluate how the quality of an
inference for a network feature degrades when the assumed model is
misspecified. Decision theory methods aimed to identify model
missespecification are applied in the context of network data with the goal of
investigating the stability of optimal actions to perturbations to the assumed
model. Here the modified versions of the model are contained within a well
defined neighborhood of model space. Our main challenge is to combine
stochastic optimization and graph limits tools to explore the model space. As a
result, a method for robustness on exchangeable random networks is developed.
Our approach is inspired by recent developments in the context of robustness
and recent works in the robust control, macroeconomics and financial
mathematics literature and more specifically and is based on the concept of
graphon approximation through its empirical graphon.
| stat |
A Fast Linear Regression via SVD and Marginalization | We describe a numerical scheme for evaluating the posterior moments of
Bayesian linear regression models with partial pooling of the coefficients. The
principal analytical tool of the evaluation is a change of basis from
coefficient space to the space of singular vectors of the matrix of predictors.
After this change of basis and an analytical integration, we reduce the problem
of finding moments of a density over k + m dimensions, to finding moments of an
m-dimensional density, where k is the number of coefficients and k + m is the
dimension of the posterior. Moments can then be computed using, for example,
MCMC, the trapezoid rule, or adaptive Gaussian quadrature. An evaluation of the
SVD of the matrix of predictors is the dominant computational cost and is
performed once during the precomputation stage. We demonstrate numerical
results of the algorithm. The scheme described in this paper generalizes
naturally to multilevel and multi-group hierarchical regression models where
normal-normal parameters appear.
| stat |
Estimation of underreporting in Brazilian tuberculosis data, 2012-2014 | Analysis of burden of underregistration in tuberculosis data in Brazil, from
2012 to 2014. Approches of Oliveira et al. (2020) and Stoner et al. (2019) are
applied. The main focus is to illustrated how the approach of Oliveira et al.
(2020) can be applied when the clustering structure is not previously
available.
| stat |
A Dynamic Choice Model with Heterogeneous Decision Rules: Application in
Estimating the User Cost of Rail Crowding | Crowding valuation of subway riders is an important input to various
supply-side decisions of transit operators. The crowding cost perceived by a
transit rider is generally estimated by capturing the trade-off that the rider
makes between crowding and travel time while choosing a route. However,
existing studies rely on static compensatory choice models and fail to account
for inertia and the learning behaviour of riders. To address these challenges,
we propose a new dynamic latent class model (DLCM) which (i) assigns riders to
latent compensatory and inertia/habit classes based on different decision
rules, (ii) enables transitions between these classes over time, and (iii)
adopts instance-based learning theory to account for the learning behaviour of
riders. We use the expectation-maximisation algorithm to estimate DLCM, and the
most probable sequence of latent classes for each rider is retrieved using the
Viterbi algorithm. The proposed DLCM can be applied in any choice context to
capture the dynamics of decision rules used by a decision-maker. We demonstrate
its practical advantages in estimating the crowding valuation of an Asian
metro's riders. To calibrate the model, we recover the daily route preferences
and in-vehicle crowding experiences of regular metro riders using a
two-month-long smart card and vehicle location data. The results indicate that
the average rider follows the compensatory rule on only 25.5% of route choice
occasions. DLCM estimates also show an increase of 47% in metro riders'
valuation of travel time under extremely crowded conditions relative to that
under uncrowded conditions.
| stat |
Hierarchical spatial log Gaussian Cox process model for replicated point
patterns | We propose a conditional log Gaussian Cox process (LGCP) model to investigate
the effect of a realization y of a point process Y on the intensity of a point
process X. In the motivating forestry example, the point pattern y represents
large trees and the point pattern x, a realization of X, seedlings. In the
model, every point in Y has a parametric influence kernel or signal, which
together form an influence field. Conditionally on the parameters, the
influence field acts as a spatial covariate in the (log) intensity of the LGCP
model, and the (log) intensity itself is a non-linear function of the
parameters. Unlike in the typical unconditional LGCP situation, points of Y
outside the observation window may affect the intensity of X inside the window.
Therefore, we propose a simple edge correction method to account for this edge
effect. The parameters of the model are estimated in a Bayesian framework using
Markov chain Monte Carlo (MCMC) where a Laplace approximation is used for the
Gaussian field of the LGCP model. Since forest data are often measured in small
sample plots, we present the estimation procedure based on replicates. The
proposed model is fitted to uneven-aged forest stands in Finland to study the
effect of large trees on the success of regeneration.
| stat |
A kernel log-rank test of independence for right-censored data | We introduce a general non-parametric independence test between
right-censored survival times and covariates, which may be multivariate. Our
test statistic has a dual interpretation, first in terms of the supremum of a
potentially infinite collection of weight-indexed log-rank tests, with weight
functions belonging to a reproducing kernel Hilbert space (RKHS) of functions;
and second, as the norm of the difference of embeddings of certain finite
measures into the RKHS, similar to the Hilbert-Schmidt Independence Criterion
(HSIC) test-statistic. We study the asymptotic properties of the test, finding
sufficient conditions to ensure our test correctly rejects the null hypothesis
under any alternative. The test statistic can be computed straightforwardly,
and the rejection threshold is obtained via an asymptotically consistent Wild
Bootstrap procedure. Extensive simulations demonstrate that our testing
procedure generally performs better than competing approaches in detecting
complex non-linear dependence.
| stat |
SAFER: Sparse Secure Aggregation for Federated Learning | Federated learning enables one to train a common machine learning model
across separate, privately-held datasets via distributed model training. During
federated training, only intermediate model parameters are transmitted to a
central server which aggregates these parameters to create a new common model,
thus exposing only intermediate parameters rather than the training data
itself. However, some attacks (e.g. membership inference) are able to infer
properties of local data from these intermediate model parameters. Hence,
performing the aggregation of these client-specific model parameters in a
secure way is required. Additionally, the communication cost is often the
bottleneck of the federated systems, especially for large neural networks. So,
limiting the number and the size of communications is necessary to efficiently
train large neural architectures. In this article, we present an efficient and
secure protocol for performing secure aggregation over compressed model updates
in the context of collaborative, few-party federated learning, a context common
in the medical, healthcare, and biotechnical use-cases of federated systems. By
making compression-based federated techniques amenable to secure computation,
we develop a secure aggregation protocol between multiple servers with very low
communication and computation costs and without preprocessing overhead. Our
experiments demonstrate the efficiency of this new approach for secure
federated training of deep convolutional neural networks.
| stat |
A clusterwise supervised learning procedure based on aggregation of
distances | Nowadays, many machine learning procedures are available on the shelve and
may be used easily to calibrate predictive models on supervised data. However,
when the input data consists of more than one unknown cluster, and when
different underlying predictive models exist, fitting a model is a more
challenging task. We propose, in this paper, a procedure in three steps to
automatically solve this problem. The KFC procedure aggregates different models
adaptively on data. The first step of the procedure aims at catching the
clustering structure of the input data, which may be characterized by several
statistical distributions. It provides several partitions, given the
assumptions on the distributions. For each partition, the second step fits a
specific predictive model based on the data in each cluster. The overall model
is computed by a consensual aggregation of the models corresponding to the
different partitions. A comparison of the performances on different simulated
and real data assesses the excellent performance of our method in a large
variety of prediction problems.
| stat |
Dying ReLU and Initialization: Theory and Numerical Examples | The dying ReLU refers to the problem when ReLU neurons become inactive and
only output 0 for any input. There are many empirical and heuristic
explanations of why ReLU neurons die. However, little is known about its
theoretical analysis. In this paper, we rigorously prove that a deep ReLU
network will eventually die in probability as the depth goes to infinite.
Several methods have been proposed to alleviate the dying ReLU. Perhaps, one of
the simplest treatments is to modify the initialization procedure. One common
way of initializing weights and biases uses symmetric probability
distributions, which suffers from the dying ReLU. We thus propose a new
initialization procedure, namely, a randomized asymmetric initialization. We
prove that the new initialization can effectively prevent the dying ReLU. All
parameters required for the new initialization are theoretically designed.
Numerical examples are provided to demonstrate the effectiveness of the new
initialization procedure.
| stat |
On generating random Gaussian graphical models | Structure learning methods for covariance and concentration graphs are often
validated on synthetic models, usually obtained by randomly generating: (i) an
undirected graph, and (ii) a compatible symmetric positive definite (SPD)
matrix. In order to ensure positive definiteness in (ii), a dominant diagonal
is usually imposed. In this work we investigate different methods to generate
random symmetric positive definite matrices with undirected graphical
constraints. We show that if the graph is chordal it is possible to sample
uniformly from the set of correlation matrices compatible with the graph, while
for general undirected graphs we rely on a partial orthogonalization method.
| stat |
Conditional Information and Inference in Response-Adaptive Allocation
Designs | Response-adaptive allocation designs refer to a class of designs where the
probability an observation is assigned to a treatment is changed throughout an
experiment based on the accrued responses. Such procedures result in random
treatment sample sizes. Most of the current literature considers unconditional
inference procedures in the analysis of response-adaptive allocation designs.
The focus of this work is inference conditional on the observed treatment
sample sizes. The inverse of information is a description of the large sample
variance of the parameter estimates. A simple form for the conditional
information relative to unconditional information is derived. It is found that
conditional information can be greater than unconditional information. It is
also shown that the variance of the conditional maximum likelihood estimate can
be less than the variance of the unconditional maximum likelihood estimate.
Finally, a conditional bootstrap procedure is developed that, in the majority
of cases examined, resulted in narrower confidence intervals than relevant
unconditional procedures.
| stat |
Cross-Component Registration for Multivariate Functional Data, With
Application to Growth Curves | Multivariate functional data are becoming ubiquitous with advances in modern
technology and are substantially more complex than univariate functional data.
We propose and study a novel model for multivariate functional data where the
component processes are subject to mutual time warping. That is, the component
processes exhibit a similar shape but are subject to systematic phase variation
across their time domains. To address this previously unconsidered mode of
warping, we propose new registration methodology which is based on a
shift-warping model. Our method differs from all existing registration methods
for functional data in a fundamental way. Namely, instead of focusing on the
traditional approach to warping, where one aims to recover individual-specific
registration, we focus on shift registration across the components of a
multivariate functional data vector on a population-wide level. Our proposed
estimates for these shifts are identifiable, enjoy parametric rates of
convergence and often have intuitive physical interpretations, all in contrast
to traditional curve-specific registration approaches. We demonstrate the
implementation and interpretation of the proposed method by applying our
methodology to the Z\"urich Longitudinal Growth data and study its finite
sample properties in simulations.
| stat |
Noise Contrastive Priors for Functional Uncertainty | Obtaining reliable uncertainty estimates of neural network predictions is a
long standing challenge. Bayesian neural networks have been proposed as a
solution, but it remains open how to specify their prior. In particular, the
common practice of an independent normal prior in weight space imposes
relatively weak constraints on the function posterior, allowing it to
generalize in unforeseen ways on inputs outside of the training distribution.
We propose noise contrastive priors (NCPs) to obtain reliable uncertainty
estimates. The key idea is to train the model to output high uncertainty for
data points outside of the training distribution. NCPs do so using an input
prior, which adds noise to the inputs of the current mini batch, and an output
prior, which is a wide distribution given these inputs. NCPs are compatible
with any model that can output uncertainty estimates, are easy to scale, and
yield reliable uncertainty estimates throughout training. Empirically, we show
that NCPs prevent overfitting outside of the training distribution and result
in uncertainty estimates that are useful for active learning. We demonstrate
the scalability of our method on the flight delays data set, where we
significantly improve upon previously published results.
| stat |
Efficient design of geographically-defined clusters with spatial
autocorrelation | Clusters form the basis of a number of research study designs including
survey and experimental studies. Cluster-based designs can be less costly but
also less efficient than individual-based designs due to correlation between
individuals within the same cluster. Their design typically relies on
\textit{ad hoc} choices of correlation parameters, and is insensitive to
variations in cluster design. This article examines how to efficiently design
clusters where they are geographically defined by demarcating areas
incorporating individuals and households or other units. Using geostatistical
models for spatial autocorrelation we generate approximations to within cluster
average covariance in order to estimate the effective sample size given
particular cluster design parameters. We show how the number of enumerated
locations, cluster area, proportion sampled, and sampling method affect the
efficiency of the design and consider the optimization problem of choosing the
most efficient design subject to budgetary constraints. We also consider how
the parameters from these approximations can be interpreted simply in terms of
`real-world' quantities and used in design analysis.
| stat |
Bayesian inference for high-dimensional nonstationary Gaussian processes | In spite of the diverse literature on nonstationary spatial modeling and
approximate Gaussian process (GP) methods, there are no general approaches for
conducting fully Bayesian inference for moderately sized nonstationary spatial
data sets on a personal laptop. For statisticians and data scientists who wish
to learn about spatially-referenced data and conduct posterior inference and
prediction with appropriate uncertainty quantification, the lack of such
approaches and corresponding software is a significant limitation. In this
paper, we develop methodology for implementing formal Bayesian inference for a
general class of nonstationary GPs. Our novel approach uses pre-existing
frameworks for characterizing nonstationarity in a new way that is applicable
for small to moderately sized data sets via modern GP likelihood
approximations. Posterior sampling is implemented using flexible MCMC methods,
with nonstationary posterior prediction conducted as a post-processing step. We
demonstrate our novel methods on two data sets, ranging from several hundred to
several thousand locations, and compare our methodology with related
statistical methods that provide off-the-shelf software. All of our methods are
implemented in the freely available BayesNSGP software package for R.
| stat |
Targeting Solutions in Bayesian Multi-Objective Optimization: Sequential
and Batch Versions | Multi-objective optimization aims at finding trade-off solutions to
conflicting objectives. These constitute the Pareto optimal set. In the context
of expensive-to-evaluate functions, it is impossible and often non-informative
to look for the entire set. As an end-user would typically prefer a certain
part of the objective space, we modify the Bayesian multi-objective
optimization algorithm which uses Gaussian Processes to maximize the Expected
Hypervolume Improvement, to focus the search in the preferred region. The
cumulated effects of the Gaussian Processes and the targeting strategy lead to
a particularly efficient convergence to the desired part of the Pareto set. To
take advantage of parallel computing, a multi-point extension of the targeting
criterion is proposed and analyzed.
| stat |
Solve fraud detection problem by using graph based learning methods | The credit cards' fraud transactions detection is the important problem in
machine learning field. To detect the credit cards's fraud transactions help
reduce the significant loss of the credit cards' holders and the banks. To
detect the credit cards' fraud transactions, data scientists normally employ
the unsupervised learning techniques and supervised learning techniques. In
this paper, we employ the graph p-Laplacian based semi-supervised learning
methods combined with the undersampling techniques such as Cluster Centroids to
solve the credit cards' fraud transactions detection problem. Experimental
results show that the graph p-Laplacian semi-supervised learning methods
outperform the current state of the art graph Laplacian based semi-supervised
learning method (p=2).
| stat |
Profile and Globe Tests of Mean Surfaces for Two-Sample Bivariate
Functional Data | Multivariate functional data has received considerable attention but testing
for equality of mean surfaces and its profile has limited progress. The
existing literature has tested equality of either mean curves of univariate
functional samples directly, or mean surfaces of bivariate functional data
samples but turn into functional curves comparison again. In this paper, we aim
to develop both the profile and globe tests of mean surfaces for two-sample
bivariate functional data. We present valid approaches of tests by employing
the idea of pooled projection and by developing a novel profile functional
principal component analysis tool. The proposed methodology enjoys the merit of
readily interpretability and implementation. Under mild conditions, we derive
the asymptotic behaviors of test statistics under null and alternative
hypotheses. Simulations show that the proposed tests have a good control of the
type I error by the size and can detect difference in mean surfaces and its
profile effectively in terms of power in finite samples. Finally, we apply the
testing procedures to two real data sets associated with the precipitation
change affected jointly by time and locations in the Midwest of USA, and the
trends in human mortality from European period life tables.
| stat |
Continuous-Time Bayesian Networks with Clocks | Structured stochastic processes evolving in continuous time present a widely
adopted framework to model phenomena occurring in nature and engineering.
However, such models are often chosen to satisfy the Markov property to
maintain tractability. One of the more popular of such memoryless models are
Continuous Time Bayesian Networks (CTBNs). In this work, we lift its
restriction to exponential survival times to arbitrary distributions. Current
extensions achieve this via auxiliary states, which hinder tractability. To
avoid that, we introduce a set of node-wise clocks to construct a collection of
graph-coupled semi-Markov chains. We provide algorithms for parameter and
structure inference, which make use of local dependencies and conduct
experiments on synthetic data and a data-set generated through a benchmark tool
for gene regulatory networks. In doing so, we point out advantages compared to
current CTBN extensions.
| stat |
Hybrid Confidence Intervals for Informative Uniform Asymptotic Inference
After Model Selection | I propose a new type of confidence interval for correct asymptotic inference
after using data to select a model of interest without assuming any model is
correctly specified. This hybrid confidence interval is constructed by
combining techniques from the selective inference and post-selection inference
literatures to yield a short confidence interval across a wide range of data
realizations. I show that hybrid confidence intervals have correct asymptotic
coverage, uniformly over a large class of probability distributions. I
illustrate the use of these confidence intervals in the problem of inference
after using the LASSO objective function to select a regression model of
interest and provide evidence of their desirable length properties in finite
samples via a set of Monte Carlo exercises that is calibrated to real-world
data.
| stat |
The Holdout Randomization Test for Feature Selection in Black Box Models | We propose the holdout randomization test (HRT), an approach to feature
selection using black box predictive models. The HRT is a specialized version
of the conditional randomization test (CRT; Candes et al., 2018) that uses data
splitting for feasible computation. The HRT works with any predictive model and
produces a valid $p$-value for each feature. To make the HRT more practical, we
propose a set of extensions to maximize power and speed up computation. In
simulations, these extensions lead to greater power than a competing
knockoffs-based approach, without sacrificing control of the error rate. We
apply the HRT to two case studies from the scientific literature where
heuristics were originally used to select important features for predictive
models. The results illustrate how such heuristics can be misleading relative
to principled methods like the HRT. Code is available at
https://github.com/tansey/hrt.
| stat |
Sparse learning with CART | Decision trees with binary splits are popularly constructed using
Classification and Regression Trees (CART) methodology. For regression models,
this approach recursively divides the data into two near-homogenous daughter
nodes according to a split point that maximizes the reduction in sum of squares
error (the impurity) along a particular variable. This paper aims to study the
statistical properties of regression trees constructed with CART methodology.
In doing so, we find that the training error is governed by the Pearson
correlation between the optimal decision stump and response data in each node,
which we bound by constructing a prior distribution on the split points and
solving a nonlinear optimization problem. We leverage this connection between
the training error and Pearson correlation to show that CART with
cost-complexity pruning achieves an optimal complexity/goodness-of-fit tradeoff
when the depth scales with the logarithm of the sample size. Data dependent
quantities, which adapt to the dimensionality and latent structure of the
regression model, are seen to govern the rates of convergence of the prediction
error.
| stat |
Predicting Student Performance in an Educational Game Using a Hidden
Markov Model | Contributions: Prior studies on education have mostly followed the model of
the cross sectional study, namely, examining the pretest and the posttest
scores. This paper shows that students' knowledge throughout the intervention
can be estimated by time series analysis using a hidden Markov model.
Background: Analyzing time series and the interaction between the students and
the game data can result in valuable information that cannot be gained by only
cross sectional studies of the exams. Research Questions: Can a hidden Markov
model be used to analyze the educational games? Can a hidden Markov model be
used to make a prediction of the students' performance? Methodology: The study
was conducted on (N=854) students who played the Save Patch game. Students were
divided into class 1 and class 2. Class 1 students are those who scored lower
in the test than class 2 students. The analysis is done by choosing various
features of the game as the observations. Findings: The state trajectories can
predict the students' performance accurately for both class 1 and class 2.
| stat |
A unified machine learning approach to time series forecasting applied
to demand at emergency departments | There were 25.6 million attendances at Emergency Departments (EDs) in England
in 2019 corresponding to an increase of 12 million attendances over the past
ten years. The steadily rising demand at EDs creates a constant challenge to
provide adequate quality of care while maintaining standards and productivity.
Managing hospital demand effectively requires an adequate knowledge of the
future rate of admission. Using 8 years of electronic admissions data from two
major acute care hospitals in London, we develop a novel ensemble methodology
that combines the outcomes of the best performing time series and machine
learning approaches in order to make highly accurate forecasts of demand, 1, 3
and 7 days in the future. Both hospitals face an average daily demand of 208
and 106 attendances respectively and experience considerable volatility around
this mean. However, our approach is able to predict attendances at these
emergency departments one day in advance up to a mean absolute error of +/- 14
and +/- 10 patients corresponding to a mean absolute percentage error of 6.8%
and 8.6% respectively. Our analysis compares machine learning algorithms to
more traditional linear models. We find that linear models often outperform
machine learning methods and that the quality of our predictions for any of the
forecasting horizons of 1, 3 or 7 days are comparable as measured in MAE. In
addition to comparing and combining state-of-the-art forecasting methods to
predict hospital demand, we consider two different hyperparameter tuning
methods, enabling a faster deployment of our models without compromising
performance. We believe our framework can readily be used to forecast a wide
range of policy relevant indicators.
| stat |
BayesTime: Bayesian Functional Principal Components for Sparse
Longitudinal Data | Modeling non-linear temporal trajectories is of fundamental interest in many
application areas, such as in longitudinal microbiome analysis. Many existing
methods focus on estimating mean trajectories, but it is also often of value to
assess temporal patterns of individual subjects. Sparse principal components
analysis (SFPCA) serves as a useful tool for assessing individual variation in
non-linear trajectories; however its application to real data often requires
careful model selection criteria and diagnostic tools. Here, we propose a
Bayesian approach to SFPCA, which allows users to use the efficient
leave-one-out cross-validation (LOO) with Pareto-smoothed importance sampling
(PSIS) for model selection, and to utilize the estimated shape parameter from
PSIS-LOO and also the posterior predictive checks for graphical model
diagnostics. This Bayesian implementation thus enables careful application of
SFPCA to a wide range of longitudinal data applications.
| stat |
Ensemble MCMC: Accelerating Pseudo-Marginal MCMC for State Space Models
using the Ensemble Kalman Filter | Particle Markov chain Monte Carlo (pMCMC) is now a popular method for
performing Bayesian statistical inference on challenging state space models
(SSMs) with unknown static parameters. It uses a particle filter (PF) at each
iteration of an MCMC algorithm to unbiasedly estimate the likelihood for a
given static parameter value. However, pMCMC can be computationally intensive
when a large number of particles in the PF is required, such as when the data
is highly informative, the model is misspecified and/or the time series is
long. In this paper we exploit the ensemble Kalman filter (EnKF) developed in
the data assimilation literature to speed up pMCMC. We replace the unbiased PF
likelihood with the biased EnKF likelihood estimate within MCMC to sample over
the space of the static parameter. On a wide class of different non-linear SSM
models, we demonstrate that our new ensemble MCMC (eMCMC) method can
significantly reduce the computational cost whilst maintaining reasonable
accuracy. We also propose several extensions of the vanilla eMCMC algorithm to
further improve computational efficiency. Computer code to implement our
methods on all the examples can be downloaded from
https://github.com/cdrovandi/Ensemble-MCMC.
| stat |
Adversarial Variational Optimization of Non-Differentiable Simulators | Complex computer simulators are increasingly used across fields of science as
generative models tying parameters of an underlying theory to experimental
observations. Inference in this setup is often difficult, as simulators rarely
admit a tractable density or likelihood function. We introduce Adversarial
Variational Optimization (AVO), a likelihood-free inference algorithm for
fitting a non-differentiable generative model incorporating ideas from
generative adversarial networks, variational optimization and empirical Bayes.
We adapt the training procedure of generative adversarial networks by replacing
the differentiable generative network with a domain-specific simulator. We
solve the resulting non-differentiable minimax problem by minimizing
variational upper bounds of the two adversarial objectives. Effectively, the
procedure results in learning a proposal distribution over simulator
parameters, such that the JS divergence between the marginal distribution of
the synthetic data and the empirical distribution of observed data is
minimized. We evaluate and compare the method with simulators producing both
discrete and continuous data.
| stat |
Visualizing class specific heterogeneous tendencies in categorical data | In multiple correspondence analysis, both individuals (observations) and
categories can be represented in a biplot that jointly depicts the
relationships across categories or individuals, as well as the associations
between them. Additional information about the individuals can enhance
interpretation capacities, such as by including categorical variables for which
the interdependencies are not of immediate concern, but that facilitate the
interpretation of the plot with respect to relationships between individuals
and categories. This article proposes a new method for adding such information,
according to a multiple-set cluster correspondence analysis approach that
identifies clusters specific to classes, or subsets of the data that correspond
to the categories of the additional variables. The proposed method can
construct a biplot that depicts heterogeneous tendencies of individual members,
as well as their relationships with the original categorical variables. A
simulation study to investigate the performance of this proposed method and an
application to data regarding road accidents in the United Kingdom confirms the
viability of this approach.
| stat |
Exploiting Convexification for Bayesian Optimal Sensor Placement by
Maximization of Mutual Information | Bayesian optimal sensor placement, in its full generality, seeks to maximize
the mutual information between uncertain model parameters and the predicted
data to be collected from the sensors for the purpose of performing Bayesian
inference. Equivalently, the expected information entropy of the posterior of
the model parameters is minimized over all possible sensor configurations for a
given sensor budget. In the context of structural dynamical systems, this
minimization is computationally expensive because of the large number of
possible sensor configurations. Here, a very efficient convex relaxation scheme
is presented to determine informative and possibly-optimal solutions to the
problem, thereby bypassing the necessity for an exhaustive, and often
infeasible, combinatorial search. The key idea is to relax the binary sensor
location vector so that its components corresponding to all possible sensor
locations lie in the unit interval. Then, the optimization over this vector is
a convex problem that can be efficiently solved. This method always yields a
unique solution for the relaxed problem, which is often binary and therefore
the optimal solution to the original problem. When not binary, the relaxed
solution is often suggestive of what the optimal solution for the original
problem is. An illustrative example using a fifty-story shear building model
subject to sinusoidal ground motion is presented, including a case where there
are over 47 trillion possible sensor configurations. The solutions and
computational effort are compared to greedy and heuristic methods.
| stat |
Matching with Text Data: An Experimental Evaluation of Methods for
Matching Documents and of Measuring Match Quality | Matching for causal inference is a well-studied problem, but standard methods
fail when the units to match are text documents: the high-dimensional and rich
nature of the data renders exact matching infeasible, causes propensity scores
to produce incomparable matches, and makes assessing match quality difficult.
In this paper, we characterize a framework for matching text documents that
decomposes existing methods into: (1) the choice of text representation, and
(2) the choice of distance metric. We investigate how different choices within
this framework affect both the quantity and quality of matches identified
through a systematic multifactor evaluation experiment using human subjects.
Altogether we evaluate over 100 unique text matching methods along with 5
comparison methods taken from the literature. Our experimental results identify
methods that generate matches with higher subjective match quality than current
state-of-the-art techniques. We enhance the precision of these results by
developing a predictive model to estimate the match quality of pairs of text
documents as a function of our various distance scores. This model, which we
find successfully mimics human judgment, also allows for approximate and
unsupervised evaluation of new procedures. We then employ the identified best
method to illustrate the utility of text matching in two applications. First,
we engage with a substantive debate in the study of media bias by using text
matching to control for topic selection when comparing news articles from
thirteen news sources. We then show how conditioning on text data leads to more
precise causal inferences in an observational study examining the effects of a
medical intervention.
| stat |
Multimodal Sparse Bayesian Dictionary Learning | This paper addresses the problem of learning dictionaries for multimodal
datasets, i.e. datasets collected from multiple data sources. We present an
algorithm called multimodal sparse Bayesian dictionary learning (MSBDL). MSBDL
leverages information from all available data modalities through a joint
sparsity constraint. The underlying framework offers a considerable amount of
flexibility to practitioners and addresses many of the shortcomings of existing
multimodal dictionary learning approaches. In particular, the procedure
includes the automatic tuning of hyperparameters and is unique in that it
allows the dictionaries for each data modality to have different cardinality, a
significant feature in cases when the dimensionality of data differs across
modalities. MSBDL is scalable and can be used in supervised learning settings.
Theoretical results relating to the convergence of MSBDL are presented and the
numerical results provide evidence of the superior performance of MSBDL on
synthetic and real datasets compared to existing methods.
| stat |
Multimodal Data Fusion of Non-Gaussian Spatial Fields in Sensor Networks | We develop a robust data fusion algorithm for field reconstruction of
multiple physical phenomena. The contribution of this paper is twofold: First,
we demonstrate how multi-spatial fields which can have any marginal
distributions and exhibit complex dependence structures can be constructed. To
this end we develop a model where a latent process of these physical phenomena
is modelled as Multiple Gaussian Process (MGP), and the dependence structure
between these phenomena is captured through a Copula process. This model has
the advantage of allowing one to choose any marginal distributions for the
physical phenomenon. Second, we develop an efficient and robust linear
estimation algorithm to predict the mean behaviour of the physical phenomena
using rank correlation instead of the conventional linear Pearson correlation.
Our approach has the advantage of avoiding the need to derive intractable
predictive posterior distribution and also has a tractable solution for the
rank correlation values. We show that our model outperforms the model which
uses the conventional linear Pearson correlation metric in terms of the
prediction mean-squared-errors (MSE). This provides the motivation for using
our models for multimodal data fusion.
| stat |
Particle-based Energetic Variational Inference | We introduce a new variational inference (VI) framework, called energetic
variational inference (EVI). It minimizes the VI objective function based on a
prescribed energy-dissipation law. Using the EVI framework, we can derive many
existing Particle-based Variational Inference (ParVI) methods, including the
popular Stein Variational Gradient Descent (SVGD) approach. More importantly,
many new ParVI schemes can be created under this framework. For illustration,
we propose a new particle-based EVI scheme, which performs the particle-based
approximation of the density first and then uses the approximated density in
the variational procedure, or "Approximation-then-Variation" for short. Thanks
to this order of approximation and variation, the new scheme can maintain the
variational structure at the particle level, and can significantly decrease the
KL-divergence in each iteration. Numerical experiments show the proposed method
outperforms some existing ParVI methods in terms of fidelity to the target
distribution.
| stat |
Practical and Consistent Estimation of f-Divergences | The estimation of an f-divergence between two probability distributions based
on samples is a fundamental problem in statistics and machine learning. Most
works study this problem under very weak assumptions, in which case it is
provably hard. We consider the case of stronger structural assumptions that are
commonly satisfied in modern machine learning, including representation
learning and generative modelling with autoencoder architectures. Under these
assumptions we propose and study an estimator that can be easily implemented,
works well in high dimensions, and enjoys faster rates of convergence. We
verify the behavior of our estimator empirically in both synthetic and
real-data experiments, and discuss its direct implications for total
correlation, entropy, and mutual information estimation.
| stat |
Spatial and Spatiotemporal GARCH Models -- A Unified Approach | In time-series analyses, particularly for finance, generalized autoregressive
conditional heteroscedasticity (GARCH) models are widely applied statistical
tools for modelling volatility clusters (i.e., periods of increased or
decreased risk). In contrast, it has not been considered to be of critical
importance until now to model spatial dependence in the conditional second
moments. Only a few models have been proposed for modelling local clusters of
increased risks. In this paper, we introduce novel spatial GARCH and
exponential GARCH processes in a unified spatial and spatiotemporal GARCH-type
model, which also covers all previously proposed spatial ARCH models as well as
time-series GARCH models. For this common modelling framework, estimators are
derived based on nonlinear least squares and on the maximum-likelihood
approach. In addition to the theoretical contributions of this paper, we
suggest a model selection strategy that is verified by a series of Monte Carlo
simulation studies. Eventually, the use of the unified model is demonstrated by
an empirical example that focuses on real estate prices from 1995 to 2014
across the ZIP-Code areas of Berlin. A spatial autoregressive model is applied
to the data to illustrate how locally varying model uncertainties can be
captured by the spatial GARCH-type models.
| stat |
Improving performances of MCMC for Nearest Neighbor Gaussian Process
models with full data augmentation | Even though Nearest Neighbor Gaussian Processes (NNGP) alleviate considerably
MCMC implementation of Bayesian space-time models, they do not solve the
convergence problems caused by high model dimension. Frugal alternatives such
as response or collapsed algorithms are an answer.gree Our approach is to keep
full data augmentation but to try and make it more efficient. We present two
strategies to do so. The first scheme is to pay a particular attention to the
seemingly trivial fixed effects of the model. We show empirically that
re-centering the latent field on the intercept critically improves chain
behavior. We extend this approach to other fixed effects that may interfere
with a coherent spatial field. We propose a simple method that requires no
tuning while remaining affordable thanks to NNGP's sparsity. The second scheme
accelerates the sampling of the random field using Chromatic samplers. This
method makes long sequential simulation boil down to group-parallelized or
group-vectorized sampling. The attractive possibility to parallelize NNGP
likelihood can therefore be carried over to field sampling. We present a R
implementation of our methods for Gaussian fields in the public repository
https://github.com/SebastienCoube/Improving_NNGP_full_augmentation . An
extensive vignette is provided. We run our implementation on two synthetic toy
examples along with the state of the art package spNNGP. Finally, we apply our
method on a real data set of lead contamination in the United States of America
mainland.
| stat |
IVFS: Simple and Efficient Feature Selection for High Dimensional
Topology Preservation | Feature selection is an important tool to deal with high dimensional data. In
unsupervised case, many popular algorithms aim at maintaining the structure of
the original data. In this paper, we propose a simple and effective feature
selection algorithm to enhance sample similarity preservation through a new
perspective, topology preservation, which is represented by persistent diagrams
from the context of computational topology. This method is designed upon a
unified feature selection framework called IVFS, which is inspired by random
subset method. The scheme is flexible and can handle cases where the problem is
analytically intractable. The proposed algorithm is able to well preserve the
pairwise distances, as well as topological patterns, of the full data. We
demonstrate that our algorithm can provide satisfactory performance under a
sharp sub-sampling rate, which supports efficient implementation of our
proposed method to large scale datasets. Extensive experiments validate the
effectiveness of the proposed feature selection scheme.
| stat |
GP-LVM of categorical data from test-positive COVID-19 pregnant women | Novel coronavirus disease 2019 (COVID-19) is rapidly spreading throughout the
world and while pregnant women present the same adverse outcome rates, they are
underrepresented in clinical research. In this paper, we model categorical
variables of 89 test-positive COVID-19 pregnant women within the unsupervised
Bayesian framework. We model the data using latent Gaussian processes for
density estimation of multivariate categorical data. The results show that the
model can find latent patterns in the data, which in turn could provide
additional insights into the study of pregnant women that are COVID-19
positive.
| stat |
Bayesian non-asymptotic extreme value models for environmental data | Motivated by the analysis of extreme rainfall data, we introduce a general
Bayesian hierarchical model for estimating the probability distribution of
extreme values of intermittent random sequences, a common problem in
geophysical and environmental science settings. The approach presented here
relaxes the asymptotic assumption typical of the traditional extreme value (EV)
theory, and accounts for the possible underlying variability in the
distribution of event magnitudes and occurrences, which are described through a
latent temporal process. Focusing on daily rainfall extremes, the structure of
the proposed model lends itself to incorporating prior geo-physical
understanding of the rainfall process. By means of an extensive simulation
study, we show that this methodology can significantly reduce estimation
uncertainty with respect to Bayesian formulations of traditional asymptotic EV
methods, particularly in the case of relatively small samples. The benefits of
the approach are further illustrated with an application to a large data set of
479 long daily rainfall historical records from across the continental United
States. By comparing measures of in-sample and out-of-sample predictive
accuracy, we find that the model structure developed here, combined with the
use of all available observations for inference, significantly improves
robustness with respect to overfitting to the specific sample.
| stat |
Adaptive Experience Selection for Policy Gradient | Policy gradient reinforcement learning (RL) algorithms have achieved
impressive performance in challenging learning tasks such as continuous
control, but suffer from high sample complexity. Experience replay is a
commonly used approach to improve sample efficiency, but gradient estimators
using past trajectories typically have high variance. Existing sampling
strategies for experience replay like uniform sampling or prioritised
experience replay do not explicitly try to control the variance of the gradient
estimates. In this paper, we propose an online learning algorithm, adaptive
experience selection (AES), to adaptively learn an experience sampling
distribution that explicitly minimises this variance. Using a regret
minimisation approach, AES iteratively updates the experience sampling
distribution to match the performance of a competitor distribution assumed to
have optimal variance. Sample non-stationarity is addressed by proposing a
dynamic (i.e. time changing) competitor distribution for which a closed-form
solution is proposed. We demonstrate that AES is a low-regret algorithm with
reasonable sample complexity. Empirically, AES has been implemented for deep
deterministic policy gradient and soft actor critic algorithms, and tested on 8
continuous control tasks from the OpenAI Gym library. Ours results show that
AES leads to significantly improved performance compared to currently available
experience sampling strategies for policy gradient.
| stat |
Sequential change point test in the presence of outliers | In this study, we consider a problem of monitoring parameter changes
particularly in the presence of outliers. To propose a sequential procedure
that is robust against outliers, we use the density power divergence to derive
a detector and stopping time that make up our procedure. We first investigate
the asymptotic properties of our sequential procedure for i.i.d. sequences, and
then extend the proposed procedure to stationary time series models, where we
provide a set of sufficient conditions under which the proposed procedure has
an asymptotically controlled size and consistency in power. As an application,
our procedure is applied to the GARCH models. We demonstrate the validity and
robustness of the proposed procedure through a simulation study. Finally, two
real data analyses are provided to illustrate the usefulness of the proposed
sequential procedure.
| stat |
Bayesian Measurement Error Models Using Finite Mixtures of Scale
Mixtures of Skew-Normal Distributions | We present a proposal to deal with the non-normality issue in the context of
regression models with measurement errors when both the response and the
explanatory variable are observed with error. We extend the normal model by
jointly modeling the unobserved covariate and the random errors by a finite
mixture of scale mixture of skew-normal distributions. This approach allows us
to model data with great flexibility, accommodating skewness, heavy tails, and
multi-modality.
| stat |
Castor: Contextual IoT Time Series Data and Model Management at Scale | We demonstrate Castor, a cloud-based system for contextual IoT time series
data and model management at scale. Castor is designed to assist Data
Scientists in (a) exploring and retrieving all relevant time series and
contextual information that is required for their predictive modelling tasks;
(b) seamlessly storing and deploying their predictive models in a cloud
production environment; (c) monitoring the performance of all predictive models
in production and (semi-)automatically retraining them in case of performance
deterioration. The main features of Castor are: (1) an efficient pipeline for
ingesting IoT time series data in real time; (2) a scalable, hybrid data
management service for both time series and contextual data; (3) a versatile
semantic model for contextual information which can be easily adopted to
different application domains; (4) an abstract framework for developing and
storing predictive models in R or Python; (5) deployment services which
automatically train and/or score predictive models upon user-defined
conditions. We demonstrate Castor for a real-world Smart Grid use case and
discuss how it can be adopted to other application domains such as Smart
Buildings, Telecommunication, Retail or Manufacturing.
| stat |
Kernel Assisted Learning for Personalized Dose Finding | An individualized dose rule recommends a dose level within a continuous safe
dose range based on patient level information such as physical conditions,
genetic factors and medication histories. Traditionally, personalized dose
finding process requires repeating clinical visits of the patient and frequent
adjustments of the dosage. Thus the patient is constantly exposed to the risk
of underdosing and overdosing during the process. Statistical methods for
finding an optimal individualized dose rule can lower the costs and risks for
patients. In this article, we propose a kernel assisted learning method for
estimating the optimal individualized dose rule. The proposed methodology can
also be applied to all other continuous decision-making problems. Advantages of
the proposed method include robustness to model misspecification and capability
of providing statistical inference for the estimated parameters. In the
simulation studies, we show that this method is capable of identifying the
optimal individualized dose rule and produces favorable expected outcomes in
the population. Finally, we illustrate our approach using data from a warfarin
dosing study for thrombosis patients.
| stat |
Using statistical control charts to monitor duration-based performance
of project | Monitoring of project performance is a crucial task of project managers that
significantly affect the project success or failure. Earned Value Management
(EVM) is a well-known tool to evaluate project performance and effective
technique for identifying delays and proposing appropriate corrective actions.
The original EVM analysis is a monetary-based method and it can be misleading
in the evaluation of the project schedule performance and estimation of the
project duration. Earned Duration Management (EDM) is a more recent method
which introduces metrics for the project schedule performance evaluation and
improves EVM analysis. In this paper, we apply statistical control charts on
EDM indices to better investigate the variations of project schedule
performance. Control charts are decision support tools to detect the out of
control performance. Usually project performance measurements are
auto-correlated and not following the normal distribution. Hence, in this
paper, a two-step adjustment framework is proposed to make the control charts
applicable to non-normal and auto-correlated measurements. The case study
project illustrates how the new method can be implemented in practice. The
numerical results conclude that that employing control chart method along with
analyzing the actual values of EDM indices increase the capability of project
management teams to detect cost and schedule problems on time
| stat |
Longitudinal Mediation Analysis with Latent Growth Curves | The paper considers mediation analysis with longitudinal data under latent
growth curve models within a counterfactual framework. Estimators and their
standard errors are derived for natural direct and indirect effects when the
mediator, the outcome, and possibly also the exposure can be modeled by an
underlying latent variable giving rise to a growth curve. Settings are also
considered in which the exposure is instead fixed at a single point in time.
| stat |
TopRank: A practical algorithm for online stochastic ranking | Online learning to rank is a sequential decision-making problem where in each
round the learning agent chooses a list of items and receives feedback in the
form of clicks from the user. Many sample-efficient algorithms have been
proposed for this problem that assume a specific click model connecting
rankings and user behavior. We propose a generalized click model that
encompasses many existing models, including the position-based and cascade
models. Our generalization motivates a novel online learning algorithm based on
topological sort, which we call TopRank. TopRank is (a) more natural than
existing algorithms, (b) has stronger regret guarantees than existing
algorithms with comparable generality, (c) has a more insightful proof that
leaves the door open to many generalizations, (d) outperforms existing
algorithms empirically.
| stat |
Simultaneous Inference for Massive Data: Distributed Bootstrap | In this paper, we propose a bootstrap method applied to massive data
processed distributedly in a large number of machines. This new method is
computationally efficient in that we bootstrap on the master machine without
over-resampling, typically required by existing methods
\cite{kleiner2014scalable,sengupta2016subsampled}, while provably achieving
optimal statistical efficiency with minimal communication. Our method does not
require repeatedly re-fitting the model but only applies multiplier bootstrap
in the master machine on the gradients received from the worker machines.
Simulations validate our theory.
| stat |
You Only Propagate Once: Accelerating Adversarial Training via Maximal
Principle | Deep learning achieves state-of-the-art results in many tasks in computer
vision and natural language processing. However, recent works have shown that
deep networks can be vulnerable to adversarial perturbations, which raised a
serious robustness issue of deep networks. Adversarial training, typically
formulated as a robust optimization problem, is an effective way of improving
the robustness of deep networks. A major drawback of existing adversarial
training algorithms is the computational overhead of the generation of
adversarial examples, typically far greater than that of the network training.
This leads to the unbearable overall computational cost of adversarial
training. In this paper, we show that adversarial training can be cast as a
discrete time differential game. Through analyzing the Pontryagin's Maximal
Principle (PMP) of the problem, we observe that the adversary update is only
coupled with the parameters of the first layer of the network. This inspires us
to restrict most of the forward and back propagation within the first layer of
the network during adversary updates. This effectively reduces the total number
of full forward and backward propagation to only one for each group of
adversary updates. Therefore, we refer to this algorithm YOPO (You Only
Propagate Once). Numerical experiments demonstrate that YOPO can achieve
comparable defense accuracy with approximately 1/5 ~ 1/4 GPU time of the
projected gradient descent (PGD) algorithm. Our codes are available at
https://https://github.com/a1600012888/YOPO-You-Only-Propagate-Once.
| stat |
ChemGrapher: Optical Graph Recognition of Chemical Compounds by Deep
Learning | In drug discovery, knowledge of the graph structure of chemical compounds is
essential. Many thousands of scientific articles in chemistry and
pharmaceutical sciences have investigated chemical compounds, but in cases the
details of the structure of these chemical compounds is published only as an
images. A tool to analyze these images automatically and convert them into a
chemical graph structure would be useful for many applications, such drug
discovery. A few such tools are available and they are mostly derived from
optical character recognition. However, our evaluation of the performance of
those tools reveals that they make often mistakes in detecting the correct bond
multiplicity and stereochemical information. In addition, errors sometimes even
lead to missing atoms in the resulting graph. In our work, we address these
issues by developing a compound recognition method based on machine learning.
More specifically, we develop a deep neural network model for optical compound
recognition. The deep learning solution presented here consists of a
segmentation model, followed by three classification models that predict atom
locations, bonds and charges. Furthermore, this model not only predicts the
graph structure of the molecule but also produces all information necessary to
relate each component of the resulting graph to the source image. This solution
is scalable and could rapidly process thousands of images. Finally, we compare
empirically the proposed method to a well-established tool and observe
significant error reductions.
| stat |
Optimal Representative Sample Weighting | We consider the problem of assigning weights to a set of samples or data
records, with the goal of achieving a representative weighting, which happens
when certain sample averages of the data are close to prescribed values. We
frame the problem of finding representative sample weights as an optimization
problem, which in many cases is convex and can be efficiently solved. Our
formulation includes as a special case the selection of a fixed number of the
samples, with equal weights, i.e., the problem of selecting a smaller
representative subset of the samples. While this problem is combinatorial and
not convex, heuristic methods based on convex optimization seem to perform very
well. We describe rsw, an open-source implementation of the ideas described in
this paper, and apply it to a skewed sample of the CDC BRFSS dataset.
| stat |
A New Framework for Inference on Markov Population Models | In this work we construct a joint Gaussian likelihood for approximate
inference on Markov population models. We demonstrate that Markov population
models can be approximated by a system of linear stochastic differential
equations with time-varying coefficients. We show that the system of stochastic
differential equations converges to a set of ordinary differential equations.
We derive our proposed joint Gaussian deterministic limiting approximation
(JGDLA) model from the limiting system of ordinary differential equations. The
results is a method for inference on Markov population models that relies
solely on the solution to a system deterministic equations. We show that our
method requires no stochastic infill and exhibits improved predictive power in
comparison to the Euler-Maruyama scheme on simulated
susceptible-infected-recovered data sets. We use the JGDLA to fit a stochastic
susceptible-exposed-infected-recovered system to the Princess Diamond COVID-19
cruise ship data set.
| stat |
Tensor Graphical Lasso (TeraLasso) | This paper introduces a multi-way tensor generalization of the Bigraphical
Lasso (BiGLasso), which uses a two-way sparse Kronecker-sum multivariate-normal
model for the precision matrix to parsimoniously model conditional dependence
relationships of matrix-variate data based on the Cartesian product of graphs.
We call this generalization the {\bf Te}nsor g{\bf ra}phical Lasso (TeraLasso).
We demonstrate using theory and examples that the TeraLasso model can be
accurately and scalably estimated from very limited data samples of high
dimensional variables with multiway coordinates such as space, time and
replicates. Statistical consistency and statistical rates of convergence are
established for both the BiGLasso and TeraLasso estimators of the precision
matrix and estimators of its support (non-sparsity) set, respectively. We
propose a scalable composite gradient descent algorithm and analyze the
computational convergence rate, showing that the composite gradient descent
algorithm is guaranteed to converge at a geometric rate to the global minimizer
of the TeraLasso objective function. Finally, we illustrate the TeraLasso using
both simulation and experimental data from a meteorological dataset, showing
that we can accurately estimate precision matrices and recover meaningful
conditional dependency graphs from high dimensional complex datasets.
| stat |
Decentralized Bayesian Learning over Graphs | We propose a decentralized learning algorithm over a general social network.
The algorithm leaves the training data distributed on the mobile devices while
utilizing a peer to peer model aggregation method. The proposed algorithm
allows agents with local data to learn a shared model explaining the global
training data in a decentralized fashion. The proposed algorithm can be viewed
as a Bayesian and peer-to-peer variant of federated learning in which each
agent keeps a "posterior probability distribution" over a global model
parameters. The agent update its "posterior" based on 1) the local training
data and 2) the asynchronous communication and model aggregation with their
1-hop neighbors. This Bayesian formulation allows for a systematic treatment of
model aggregation over any arbitrary connected graph. Furthermore, it provides
strong analytic guarantees on converge in the realizable case as well as a
closed form characterization of the rate of convergence. We also show that our
methodology can be combined with efficient Bayesian inference techniques to
train Bayesian neural networks in a decentralized manner. By empirical studies
we show that our theoretical analysis can guide the design of network/social
interactions and data partitioning to achieve convergence.
| stat |
D\'ej\`a vu: A data-centric forecasting approach through time series
cross-similarity | Accurate forecasts are vital for supporting the decisions of modern
companies. Forecasters typically select the most appropriate statistical model
for each time series. However, statistical models usually presume some data
generation process while making strong assumptions about the errors. In this
paper, we present a novel data-centric approach -- `forecasting with
similarity', which tackles model uncertainty in a model-free manner. Existing
similarity-based methods focus on identifying similar patterns within the
series, i.e., `self-similarity'. In contrast, we propose searching for similar
patterns from a reference set, i.e., `cross-similarity'. Instead of
extrapolating, the future paths of the similar series are aggregated to obtain
the forecasts of the target series. Building on the cross-learning concept, our
approach allows the application of similarity-based forecasting on series with
limited lengths. We evaluate the approach using a rich collection of real data
and show that it yields competitive accuracy in both points forecasts and
prediction intervals.
| stat |
Debiasing Evaluations That are Biased by Evaluations | It is common to evaluate a set of items by soliciting people to rate them.
For example, universities ask students to rate the teaching quality of their
instructors, and conference organizers ask authors of submissions to evaluate
the quality of the reviews. However, in these applications, students often give
a higher rating to a course if they receive higher grades in a course, and
authors often give a higher rating to the reviews if their papers are accepted
to the conference. In this work, we call these external factors the "outcome"
experienced by people, and consider the problem of mitigating these
outcome-induced biases in the given ratings when some information about the
outcome is available. We formulate the information about the outcome as a known
partial ordering on the bias. We propose a debiasing method by solving a
regularized optimization problem under this ordering constraint, and also
provide a carefully designed cross-validation method that adaptively chooses
the appropriate amount of regularization. We provide theoretical guarantees on
the performance of our algorithm, as well as experimental evaluations.
| stat |
The huge Package for High-dimensional Undirected Graph Estimation in R | We describe an R package named huge which provides easy-to-use functions for
estimating high dimensional undirected graphs from data. This package
implements recent results in the literature, including Friedman et al. (2007),
Liu et al. (2009, 2012) and Liu et al. (2010). Compared with the existing graph
estimation package glasso, the huge package provides extra features: (1)
instead of using Fortan, it is written in C, which makes the code more portable
and easier to modify; (2) besides fitting Gaussian graphical models, it also
provides functions for fitting high dimensional semiparametric Gaussian copula
models; (3) more functions like data-dependent model selection, data generation
and graph visualization; (4) a minor convergence problem of the graphical lasso
algorithm is corrected; (5) the package allows the user to apply both lossless
and lossy screening rules to scale up large-scale problems, making a tradeoff
between computational and statistical efficiency.
| stat |
Evaluation of the performance of Euro-CORDEX RCMs for assessing
hydrological climate change impacts in Great Britain: a comparison of
different spatial resolutions and quantile mapping bias correction methods | Regional Climate Models (RCMs) are an essential tool for analysing regional
climate change impacts as they provide simulations with more small-scale
details and expected smaller errors than global climate models. There has been
much effort to increase the spatial resolution and simulation skill of RCMs,
yet the extent to which this improves the projection of hydrological change is
unclear. Here, we evaluate the skill of five reanalysis-driven Euro-CORDEX RCMs
in simulating precipitation and temperature, and as drivers of a hydrological
model to simulate river flow on four UK catchments covering different physical,
climatic and hydrological characteristics. We test whether high-resolution RCMs
provide added value, through analysis of two RCM resolutions, 50 km and 12.5
km, which are also bias-corrected employing the parametric quantile-mapping
(QM) method, using the normal distribution for temperature, and the Gamma (GQM)
and Double Gamma (DGQM) distributions for precipitation. In a small catchment
with complex topography, the 12.5 km RCMs outperform their 50 km version for
precipitation and temperature, but when used in combination with the
hydrological model, fail to capture the observed river flow distribution. In
the other (larger) catchments, only one high-resolution RCM consistently
outperforms its low-resolution version, implying that in general there is no
added value from using the high-resolution RCMs in those catchments. GQM
decreases most of the simulation biases, except for extreme precipitation and
high flows, which are further decreased by DGQM. Bias correction does not
improve the representation of daily temporal variability, but it does for
monthly variability, in particular when applying DGQM. Overall, an increase in
RCM resolution does not imply a better simulation of hydrology and
bias-correction represents an alternative to ease decision-making.
| stat |
RKHSMetaMod: An R package to estimate the Hoeffding decomposition of a
complex model by solving RKHS ridge group sparse optimization problem | We propose an R package, called RKHSMetaMod, that implements a procedure for
estimating a meta-model of a complex model $m$. The meta-model approximates the
Hoeffding decomposition of $m$ and allows to perform sensitivity analysis on
it. It belongs to a reproducing kernel Hilbert space that is constructed as a
direct sum of Hilbert spaces. The estimator of the meta-model is the solution
of a penalized empirical least-squares minimization with the sum of the Hilbert
norm and the empirical $L^2$-norm. This procedure, called RKHS ridge group
sparse, allows both to select and estimate the terms in the Hoeffding
decomposition, and therefore, to select and estimate the Sobol indices that are
non-zero. The RKHSMetaMod package provides an interface from R statistical
computing environment to the C++ libraries Eigen and GSL. In order to speed up
the execution time and optimize the storage memory, except for a function that
is written in R, all of the functions of this package are written using the
efficient C++ libraries through RcppEigen and RcppGSL packages. These functions
are then interfaced in the R environment in order to propose an user friendly
package.
| stat |
Concentric mixtures of Mallows models for top-$k$ rankings: sampling and
identifiability | In this paper, we consider mixtures of two Mallows models for top-$k$
rankings, both with the same location parameter but with different scale
parameters, i.e., a mixture of concentric Mallows models. This situation arises
when we have a heterogeneous population of voters formed by two homogeneous
populations, one of which is a subpopulation of expert voters while the other
includes the non-expert voters. We propose efficient sampling algorithms for
Mallows top-$k$ rankings. We show the identifiability of both components, and
the learnability of their respective parameters in this setting by, first,
bounding the sample complexity for the Borda algorithm with top-$k$ rankings
and second, proposing polynomial time algorithm for the separation of the
rankings in each component. Finally, since the rank aggregation will suffer
from a large amount of noise introduced by the non-expert voters, we adapt the
Borda algorithm to be able to recover the ground truth consensus ranking which
is especially consistent with the expert rankings.
| stat |
Rao-Blackwellizing the Straight-Through Gumbel-Softmax Gradient
Estimator | Gradient estimation in models with discrete latent variables is a challenging
problem, because the simplest unbiased estimators tend to have high variance.
To counteract this, modern estimators either introduce bias, rely on multiple
function evaluations, or use learned, input-dependent baselines. Thus, there is
a need for estimators that require minimal tuning, are computationally cheap,
and have low mean squared error. In this paper, we show that the variance of
the straight-through variant of the popular Gumbel-Softmax estimator can be
reduced through Rao-Blackwellization without increasing the number of function
evaluations. This provably reduces the mean squared error. We empirically
demonstrate that this leads to variance reduction, faster convergence, and
generally improved performance in two unsupervised latent variable models.
| stat |
On the implied weights of linear regression for causal inference | In this paper, we derive and analyze the implied weights of linear regression
methods for causal inference. We obtain new closed-form, finite-sample
expressions of the weights for various types of estimators based on
multivariate linear regression models. In finite samples, we show that the
implied weights have minimum variance, exactly balance the means of the
covariates (or transformations thereof) included in the model, and produce
estimators that may not be sample bounded. Furthermore, depending on the
specification of the regression model, we show that the implied weights may
distort the structure of the sample in such a way that the resulting estimator
is biased for the average treatment effect for a given target population. In
large samples, we demonstrate that, under certain functional form assumptions,
the implied weights are consistent estimators of the true inverse probability
weights. We examine doubly robust properties of regression estimators from the
perspective of their implied weights. We also derive and analyze the implied
weights of weighted least squares regression. The equivalence between
minimizing regression residuals and optimizing for certain weights allows us to
bridge ideas from the regression modeling and causal inference literatures. As
a result, we propose a set of regression diagnostics for causal inference. We
discuss the connection of the implied weights to existing matching and
weighting approaches. As special cases, we analyze the implied weights in
common settings such as multi-valued treatments, regression after matching, and
two-stage least squares regression with instrumental variables.
| stat |
Deep Sigma Point Processes | We introduce Deep Sigma Point Processes, a class of parametric models
inspired by the compositional structure of Deep Gaussian Processes (DGPs). Deep
Sigma Point Processes (DSPPs) retain many of the attractive features of
(variational) DGPs, including mini-batch training and predictive uncertainty
that is controlled by kernel basis functions. Importantly, since DSPPs admit a
simple maximum likelihood inference procedure, the resulting predictive
distributions are not degraded by any posterior approximations. In an extensive
empirical comparison on univariate and multivariate regression tasks we find
that the resulting predictive distributions are significantly better calibrated
than those obtained with other probabilistic methods for scalable regression,
including variational DGPs--often by as much as a nat per datapoint.
| stat |
Byzantine-Robust Federated Machine Learning through Adaptive Model
Averaging | Federated learning enables training collaborative machine learning models at
scale with many participants whilst preserving the privacy of their datasets.
Standard federated learning techniques are vulnerable to Byzantine failures,
biased local datasets, and poisoning attacks. In this paper we introduce
Adaptive Federated Averaging, a novel algorithm for robust federated learning
that is designed to detect failures, attacks, and bad updates provided by
participants in a collaborative model. We propose a Hidden Markov Model to
model and learn the quality of model updates provided by each participant
during training. In contrast to existing robust federated learning schemes, we
propose a robust aggregation rule that detects and discards bad or malicious
local model updates at each training iteration. This includes a mechanism that
blocks unwanted participants, which also increases the computational and
communication efficiency. Our experimental evaluation on 4 real datasets show
that our algorithm is significantly more robust to faulty, noisy and malicious
participants, whilst being computationally more efficient than other
state-of-the-art robust federated learning methods such as Multi-KRUM and
coordinate-wise median.
| stat |
Tangent functional canonical correlation analysis for densities and
shapes, with applications to multimodal imaging data | It is quite common for functional data arising from imaging data to assume
values in infinite-dimensional manifolds. Uncovering associations between two
or more such nonlinear functional data extracted from the same object across
medical imaging modalities can assist development of personalized treatment
strategies. We propose a method for canonical correlation analysis between
paired probability densities or shapes of closed planar curves, routinely used
in biomedical studies, which combines a convenient linearization and dimension
reduction of the data using tangent space coordinates. Leveraging the fact that
the corresponding manifolds are submanifolds of unit Hilbert spheres, we
describe how finite-dimensional representations of the functional data objects
can be easily computed, which then facilitates use of standard multivariate
canonical correlation analysis methods. We further construct and visualize
canonical variate directions directly on the space of densities or shapes.
Utility of the method is demonstrated through numerical simulations and
performance on a magnetic resonance imaging dataset of Glioblastoma Multiforme
brain tumors.
| stat |
New Non-parametric Tests for Multivariate Paired Data | Paired data are common in many fields, such as medical diagnosis and
longitudinal data analysis, where measurements are taken for the same set of
objects under different conditions. In many of these studies, the number of
measurements could be large. Existing methods either impose strong assumptions
or have power decrease fast as the number of measurements increases. In this
work, we propose new non-parametric tests for paired data. These tests exhibit
substantial power improvements over existing methods under moderate- to high-
dimensional data. We also derived asymptotic distributions of the new tests and
the approximate $p$-values based on them are reasonably accurate under finite
samples through simulation studies, making the new tests easy-off-the-shelf
tools for real applications. The proposed tests are illustrated through the
analysis of a real data set on the Alzheimer's disease research.
| stat |
Is There an Analog of Nesterov Acceleration for MCMC? | We formulate gradient-based Markov chain Monte Carlo (MCMC) sampling as
optimization on the space of probability measures, with Kullback-Leibler (KL)
divergence as the objective functional. We show that an underdamped form of the
Langevin algorithm performs accelerated gradient descent in this metric. To
characterize the convergence of the algorithm, we construct a Lyapunov
functional and exploit hypocoercivity of the underdamped Langevin algorithm. As
an application, we show that accelerated rates can be obtained for a class of
nonconvex functions with the Langevin algorithm.
| stat |
Robust Semantic Interpretability: Revisiting Concept Activation Vectors | Interpretability methods for image classification assess model
trustworthiness by attempting to expose whether the model is systematically
biased or attending to the same cues as a human would. Saliency methods for
feature attribution dominate the interpretability literature, but these methods
do not address semantic concepts such as the textures, colors, or genders of
objects within an image. Our proposed Robust Concept Activation Vectors (RCAV)
quantifies the effects of semantic concepts on individual model predictions and
on model behavior as a whole. RCAV calculates a concept gradient and takes a
gradient ascent step to assess model sensitivity to the given concept. By
generalizing previous work on concept activation vectors to account for model
non-linearity, and by introducing stricter hypothesis testing, we show that
RCAV yields interpretations which are both more accurate at the image level and
robust at the dataset level. RCAV, like saliency methods, supports the
interpretation of individual predictions. To evaluate the practical use of
interpretability methods as debugging tools, and the scientific use of
interpretability methods for identifying inductive biases (e.g. texture over
shape), we construct two datasets and accompanying metrics for realistic
benchmarking of semantic interpretability methods. Our benchmarks expose the
importance of counterfactual augmentation and negative controls for quantifying
the practical usability of interpretability methods.
| stat |
Simultaneous disease mapping and hot spot detection with application to
childhood obesity surveillance from electronic health records | Electronic health records (EHRs) have become a platform for data-driven
surveillance on a granular level in recent years. In this paper, we make use of
EHRs for early prevention of childhood obesity. The proposed method
simultaneously provides smooth disease mapping and outlier information for
obesity prevalence, which are useful for raising public awareness and
facilitating targeted intervention. More precisely, we consider a penalized
multilevel generalized linear model. We decompose regional contribution into
smooth and sparse signals, which are automatically identified by a combination
of fusion and sparse penalties imposed on the likelihood function. In addition,
we weigh the proposed likelihood to account for the missingness and potential
non-representativeness arising from the EHR data. We develop a novel
alternating minimization algorithm, which is computationally efficient, easy to
implement, and guarantees convergence. Simulation studies demonstrate superior
performance of the proposed method. Finally, we apply our method to the
University of Wisconsin Population Health Information Exchange database.
| stat |
Generative Adversarial Networks for geometric surfaces prediction in
injection molding | Geometrical and appearance quality requirements set the limits of the current
industrial performance in injection molding. To guarantee the product's
quality, it is necessary to adjust the process settings in a closed loop. Those
adjustments cannot rely on the final quality because a part takes days to be
geometrically stable. Thus, the final part geometry must be predicted from
measurements on hot parts. In this paper, we use recent success of Generative
Adversarial Networks (GAN) with the pix2pix network architecture to predict the
final part geometry, using only hot parts thermographic images, measured right
after production. Our dataset is really small, and the GAN learns to translate
thermography to geometry. We firstly study prediction performances using
different image similarity comparison algorithms. Moreover, we introduce the
innovative use of Discrete Modal Decomposition (DMD) to analyze network
predictions. The DMD is a geometrical parameterization technique using a modal
space projection to geometrically describe surfaces. We study GAN performances
to retrieve geometrical parameterization of surfaces.
| stat |
Large-scale benchmark study of survival prediction methods using
multi-omics data | Multi-omics data, that is, datasets containing different types of
high-dimensional molecular variables (often in addition to classical clinical
variables), are increasingly generated for the investigation of various
diseases. Nevertheless, questions remain regarding the usefulness of
multi-omics data for the prediction of disease outcomes such as survival time.
It is also unclear which methods are most appropriate to derive such prediction
models. We aim to give some answers to these questions by means of a
large-scale benchmark study using real data. Different prediction methods from
machine learning and statistics were applied on 18 multi-omics cancer datasets
from the database "The Cancer Genome Atlas", containing from 35 to 1,000
observations and from 60,000 to 100,000 variables. The considered outcome was
the (censored) survival time. Twelve methods based on boosting, penalized
regression and random forest were compared, comprising both methods that do and
that do not take the group structure of the omics variables into account. The
Kaplan-Meier estimate and a Cox model using only clinical variables were used
as reference methods. The methods were compared using several repetitions of
5-fold cross-validation. Uno's C-index and the integrated Brier-score served as
performance metrics. The results show that, although multi-omics data can
improve the prediction performance, this is not generally the case. Only the
method block forest slightly outperformed the Cox model on average over all
datasets. Taking into account the multi-omics structure improves the predictive
performance and protects variables in low-dimensional groups - especially
clinical variables - from not being included in the model. All analyses are
reproducible using freely available R code.
| stat |
Metric learning by Similarity Network for Deep Semi-Supervised Learning | Deep semi-supervised learning has been widely implemented in the real-world
due to the rapid development of deep learning. Recently, attention has shifted
to the approaches such as Mean-Teacher to penalize the inconsistency between
two perturbed input sets. Although these methods may achieve positive results,
they ignore the relationship information between data instances. To solve this
problem, we propose a novel method named Metric Learning by Similarity Network
(MLSN), which aims to learn a distance metric adaptively on different domains.
By co-training with the classification network, similarity network can learn
more information about pairwise relationships and performs better on some
empirical tasks than state-of-art methods.
| stat |
Multiscale Jump Testing and Estimation Under Complex Temporal Dynamics | We consider the problem of detecting jumps in an otherwise smoothly evolving
trend whilst the covariance and higher-order structures of the system can
experience both smooth and abrupt changes over time. The number of jump points
is allowed to diverge to infinity with the jump sizes possibly shrinking to
zero. The method is based on a multiscale application of an optimal jump-pass
filter to the time series, where the scales are dense between admissible lower
and upper bounds. For a wide class of non-stationary time series models and
associated trend functions, the proposed method is shown to be able to detect
all jump points within a nearly optimal range with a prescribed probability
asymptotically. For a time series of length $n$, the computational complexity
of the proposed method is $O(n)$ for each scale and $O(n\log^{1+\epsilon} n)$
overall, where $\epsilon$ is an arbitrarily small positive constant.
Simulations and data analysis show that the proposed jump testing and
estimation method performs robustly and accurately under complex temporal
dynamics.
| stat |
Geographically Weighted Regression Analysis for Spatial Economics Data:
a Bayesian Recourse | The geographically weighted regression (GWR) is a well-known statistical
approach to explore spatial non-stationarity of the regression relationship in
spatial data analysis. In this paper, we discuss a Bayesian recourse of GWR.
Bayesian variable selection based on spike-and-slab prior, bandwidth selection
based on range prior, and model assessment using a modified deviance
information criterion and a modified logarithm of pseudo-marginal likelihood
are fully discussed in this paper. Usage of the graph distance in modeling
areal data is also introduced. Extensive simulation studies are carried out to
examine the empirical performance of the proposed methods with both small and
large number of location scenarios, and comparison with the classical
frequentist GWR is made. The performance of variable selection and estimation
of the proposed methodology under different circumstances are satisfactory. We
further apply the proposed methodology in analysis of a province-level
macroeconomic data of 30 selected provinces in China. The estimation and
variable selection results reveal insights about China's economy that are
convincing and agree with previous studies and facts.
| stat |
A causal framework for classical statistical estimands in failure time
settings with competing events | In failure-time settings, a competing risk event is any event that makes it
impossible for the event of interest to occur. For example, cardiovascular
disease death is a competing event for prostate cancer death because an
individual cannot die of prostate cancer once he has died of cardiovascular
disease. Various statistical estimands have been defined as possible targets of
inference in the classical competing risks literature. Many reviews have
described these statistical estimands and their estimating procedures with
recommendations about their use. However, this previous work has not used a
formal framework for characterizing causal effects and their identifying
conditions, which makes it difficult to interpret effect estimates and assess
recommendations regarding analytic choices. Here we use a counterfactual
framework to explicitly define each of these classical estimands. We clarify
that, depending on whether competing events are defined as censoring events,
contrasts of risks can define a total effect of the treatment on the event of
interest, or a direct effect of the treatment on the event of interest not
mediated through the competing event. In contrast, regardless of whether
competing events are defined as censoring events, counterfactual hazard
contrasts cannot generally be interpreted as causal effects. We illustrate how
identifying assumptions for all of these counterfactual estimands can be
represented in causal diagrams in which competing events are depicted as
time-varying covariates. We present an application of these ideas to data from
a randomized trial designed to estimate the effect of estrogen therapy on
prostate cancer mortality.
| stat |
Restless dependent bandits with fading memory | We study the stochastic multi-armed bandit problem in the case when the arm
samples are dependent over time and generated from so-called weak $\cC$-mixing
processes. We establish a $\cC-$Mix Improved UCB agorithm and provide both
problem-dependent and independent regret analysis in two different scenarios.
In the first, so-called fast-mixing scenario, we show that pseudo-regret enjoys
the same upper bound (up to a factor) as for i.i.d. observations; whereas in
the second, slow mixing scenario, we discover a surprising effect, that the
regret upper bound is similar to the independent case, with an incremental {\em
additive} term which does not depend on the number of arms. The analysis of
slow mixing scenario is supported with a minmax lower bound, which (up to a
$\log(T)$ factor) matches the obtained upper bound.
| stat |
Provably Efficient Neural Estimation of Structural Equation Model: An
Adversarial Approach | Structural equation models (SEMs) are widely used in sciences, ranging from
economics to psychology, to uncover causal relationships underlying a complex
system under consideration and estimate structural parameters of interest. We
study estimation in a class of generalized SEMs where the object of interest is
defined as the solution to a linear operator equation. We formulate the linear
operator equation as a min-max game, where both players are parameterized by
neural networks (NNs), and learn the parameters of these neural networks using
the stochastic gradient descent. We consider both 2-layer and multi-layer NNs
with ReLU activation functions and prove global convergence in an
overparametrized regime, where the number of neurons is diverging. The results
are established using techniques from online learning and local linearization
of NNs, and improve in several aspects the current state-of-the-art. For the
first time we provide a tractable estimation procedure for SEMs based on NNs
with provable convergence and without the need for sample splitting.
| stat |
A Method Expanding 2 by 2 Contingency Table by Obtaining Tendencies of
Boolean Operators: Boolean Monte Carlo Method | Accuracy of medical test and diagnosis are often discussed by 2 by 2
contingency tables. However, it is difficult to apply a 2 by 2 contingency
table to multivariate cases because the number of possible categories increases
exponentially. The aims of this study is introducing a method by which we can
obtain trends of boolean operators and construct a 2 by 2 contingency tables to
multivariate cases. In this method, we randomly assigned boolean operators
between binary variables and focused on frequencies of boolean operators which
could explain outcomes correctly. We determined trends of boolean operators by
chi-square test. As an application, we performed this method for a dataset
which included patient's age, body mass index (BMI), mean arterial pressure
(MAP) and the quantitative measure of diabetes progression ($Y$). We set cutoff
to each variable and considered these as binary variables. Interactions of age,
BMI and MAP were determined as age $and$ BMI $and$ MAP ($p<0.0001$ for both
operators) in estimating $Y$ and sensitivity, specificity, positive and
negative predicting value were 0.50, 0.85. 0.59, 0.80 respectively. We may be
able to detemine trends of boolean operators and construct a 2 by 2 contingency
table in multivariate situation by this method.
| stat |
On ratio measures of population heterogeneity for meta-analyses | Popular measures of meta-analysis heterogeneity, such as $I^2$, cannot be
considered measures of population heterogeneity since they are dependant on
samples sizes within studies. The coefficient of variation (CV) recently
introduced and defined to be the heterogeneity variance divided by the absolute
value of the overall mean effect does not suffer such shortcomings. However,
very large CV values can occur when the effect is small making interpretation
difficult. The purpose of this paper is two-fold. Firstly, we consider variants
of the CV that exist in the interval (0, 1] making interpretation simpler.
Secondly, we provide interval estimators for the CV and its variants with
excellent coverage properties. We perform simulation studies based on simulated
and real data sets and draw comparisons between the methods.
| stat |
Elastic Priors to Dynamically Borrow Information from Historical Data in
Clinical Trials | Use of historical data and real-world evidence holds great potential to
improve the efficiency of clinical trials. One major challenge is how to
effectively borrow information from historical data while maintaining a
reasonable type I error. We propose the elastic prior approach to address this
challenge and achieve dynamic information borrowing. Unlike existing
approaches, this method proactively controls the behavior of dynamic
information borrowing and type I errors by incorporating a well-known concept
of clinically meaningful difference through an elastic function, defined as a
monotonic function of a congruence measure between historical data and trial
data. The elastic function is constructed to satisfy a set of
information-borrowing constraints prespecified by researchers or regulatory
agencies, such that the prior will borrow information when historical and trial
data are congruent, but refrain from information borrowing when historical and
trial data are incongruent. In doing so, the elastic prior improves power and
reduces the risk of data dredging and bias. The elastic prior is information
borrowing consistent, i.e. asymptotically controls type I and II errors at the
nominal values when historical data and trial data are not congruent, a unique
characteristics of the elastic prior approach. Our simulation study that
evaluates the finite sample characteristic confirms that, compared to existing
methods, the elastic prior has better type I error control and yields
competitive or higher power.
| stat |
Statistical monitoring of functional data using the notion of Fr\'echet
mean combined with the framework of the deformation models | The aim of this paper is to investigate possible advances obtained by the
implementation of the framework of Fr\'echet mean and the generalized sense of
mean that it offers, in the field of statistical process monitoring and
control. In particular, the case of non-linear profiles which are described by
data in functional form is considered and a framework combining the notion of
Fr\'echet mean and deformation models is developed. The proposed monitoring
approach is implemented to the intra-day air pollution monitoring task in the
city of Athens where the capabilities and advantages of the method are
illustrated.
| stat |
Learning from eXtreme Bandit Feedback | We study the problem of batch learning from bandit feedback in the setting of
extremely large action spaces. Learning from extreme bandit feedback is
ubiquitous in recommendation systems, in which billions of decisions are made
over sets consisting of millions of choices in a single day, yielding massive
observational data. In these large-scale real-world applications, supervised
learning frameworks such as eXtreme Multi-label Classification (XMC) are widely
used despite the fact that they incur significant biases due to the mismatch
between bandit feedback and supervised labels. Such biases can be mitigated by
importance sampling techniques, but these techniques suffer from impractical
variance when dealing with a large number of actions. In this paper, we
introduce a selective importance sampling estimator (sIS) that operates in a
significantly more favorable bias-variance regime. The sIS estimator is
obtained by performing importance sampling on the conditional expectation of
the reward with respect to a small subset of actions for each instance (a form
of Rao-Blackwellization). We employ this estimator in a novel algorithmic
procedure -- named Policy Optimization for eXtreme Models (POXM) -- for
learning from bandit feedback on XMC tasks. In POXM, the selected actions for
the sIS estimator are the top-p actions of the logging policy, where p is
adjusted from the data and is significantly smaller than the size of the action
space. We use a supervised-to-bandit conversion on three XMC datasets to
benchmark our POXM method against three competing methods: BanditNet, a
previously applied partial matching pruning strategy, and a supervised learning
baseline. Whereas BanditNet sometimes improves marginally over the logging
policy, our experiments show that POXM systematically and significantly
improves over all baselines.
| stat |
Latent Embeddings of Point Process Excitations | When specific events seem to spur others in their wake, marked Hawkes
processes enable us to reckon with their statistics. The underdetermined
empirical nature of these event-triggering mechanisms hinders estimation in the
multivariate setting. Spatiotemporal applications alleviate this obstacle by
allowing relationships to depend only on relative distances in real Euclidean
space; we employ the framework as a vessel for embedding arbitrary event types
in a new latent space. By performing synthetic experiments on short records as
well as an investigation into options markets and pathogens, we demonstrate
that learning the embedding alongside a point process model uncovers the
coherent, rather than spurious, interactions.
| stat |
Elliptical Symmetry Tests in \proglang{R} | The assumption of elliptical symmetry has an important role in many
theoretical developments and applications, hence it is of primary importance to
be able to test whether that assumption actually holds true or not. Various
tests have been proposed in the literature for this problem. To the best of our
knowledge, none of them has been implemented in R. The focus of this paper is
the implementation of several well-known tests for elliptical symmetry together
with some recent tests. We demonstrate the testing procedures with a real data
example.
| stat |
Reducing Noise in GAN Training with Variance Reduced Extragradient | We study the effect of the stochastic gradient noise on the training of
generative adversarial networks (GANs) and show that it can prevent the
convergence of standard game optimization methods, while the batch version
converges. We address this issue with a novel stochastic variance-reduced
extragradient (SVRE) optimization algorithm, which for a large class of games
improves upon the previous convergence rates proposed in the literature. We
observe empirically that SVRE performs similarly to a batch method on MNIST
while being computationally cheaper, and that SVRE yields more stable GAN
training on standard datasets.
| stat |
A Bayesian Nonparametric Analysis of the 2003 Outbreak of Highly
Pathogenic Avian Influenza in the Netherlands | Infectious diseases on farms pose both public and animal health risks, so
understanding how they spread between farms is crucial for developing disease
control strategies to prevent future outbreaks. We develop novel Bayesian
nonparametric methodology to fit spatial stochastic transmission models in
which the infection rate between any two farms is a function that depends on
the distance between them, but without assuming a specified parametric form.
Making nonparametric inference in this context is challenging since the
likelihood function of the observed data is intractable because the underlying
transmission process is unobserved. We adopt a fully Bayesian approach by
assigning a transformed Gaussian Process prior distribution to the infection
rate function, and then develop an efficient data augmentation Markov Chain
Monte Carlo algorithm to perform Bayesian inference. We use the posterior
predictive distribution to simulate the effect of different disease control
methods and their economic impact. We analyse a large outbreak of Avian
Influenza in the Netherlands and infer the between-farm infection rate, as well
as the unknown infection status of farms which were pre-emptively culled. We
use our results to analyse ring-culling strategies, and conclude that although
effective, ring-culling has limited impact in high density areas.
| stat |
Developing a predictive signature for two trial endpoints using the
cross-validated risk scores method | The existing cross-validated risk scores (CVRS) design has been proposed for
developing and testing the efficacy of a treatment in a high-efficacy patient
group (the sensitive group) using high-dimensional data (such as genetic data).
The design is based on computing a risk score for each patient and dividing
them into clusters using a non-parametric clustering procedure. In some
settings it is desirable to consider the trade-off between two outcomes, such
as efficacy and toxicity, or cost and effectiveness. With this motivation, we
extend the CVRS design (CVRS2) to consider two outcomes. The design employs
bivariate risk scores that are divided into clusters. We assess the properties
of the CVRS2 using simulated data and illustrate its application on a
randomised psychiatry trial. We show that CVRS2 is able to reliably identify
the sensitive group (the group for which the new treatment provides benefit on
both outcomes) in the simulated data. We apply the CVRS2 design to a psychology
clinical trial that had offender status and substance use status as two
outcomes and collected a large number of baseline covariates. The CVRS2 design
yields a significant treatment effect for both outcomes, while the CVRS
approach identified a significant effect for the offender status only after
pre-filtering the covariates.
| stat |
Learning from DPPs via Sampling: Beyond HKPV and symmetry | Determinantal point processes (DPPs) have become a significant tool for
recommendation systems, feature selection, or summary extraction, harnessing
the intrinsic ability of these probabilistic models to facilitate sample
diversity. The ability to sample from DPPs is paramount to the empirical
investigation of these models. Most exact samplers are variants of a spectral
meta-algorithm due to Hough, Krishnapur, Peres and Vir\'ag (henceforth HKPV),
which is in general time and resource intensive. For DPPs with symmetric
kernels, scalable HKPV samplers have been proposed that either first downsample
the ground set of items, or force the kernel to be low-rank, using e.g.
Nystr\"om-type decompositions.
In the present work, we contribute a radically different approach than HKPV.
Exploiting the fact that many statistical and learning objectives can be
effectively accomplished by only sampling certain key observables of a DPP
(so-called linear statistics), we invoke an expression for the Laplace
transform of such an observable as a single determinant, which holds in
complete generality. Combining traditional low-rank approximation techniques
with Laplace inversion algorithms from numerical analysis, we show how to
directly approximate the distribution function of a linear statistic of a DPP.
This distribution function can then be used in hypothesis testing or to
actually sample the linear statistic, as per requirement. Our approach is
scalable and applies to very general DPPs, beyond traditional symmetric
kernels.
| stat |
Dynamic Risk Prediction Using Survival Tree Ensembles with Application
to Cystic Fibrosis | With the availability of massive amounts of data from electronic health
records and registry databases, incorporating time-varying patient information
to improve risk prediction has attracted great attention. To exploit the
growing amount of predictor information over time, we develop a unified
framework for landmark prediction using survival tree ensembles, where an
updated prediction can be performed when new information becomes available.
Compared to the conventional landmark prediction, our framework enjoys great
flexibility in that the landmark times can be subject-specific and triggered by
an intermediate clinical event. Moreover, the nonparametric approach
circumvents the thorny issue in model incompatibility at different landmark
times. When both the longitudinal predictors and the outcome event time are
subject to right censoring, existing tree-based approaches cannot be directly
applied. To tackle the analytical challenges, we consider a risk-set-based
ensemble procedure by averaging martingale estimating equations from individual
trees. Extensive simulation studies are conducted to evaluate the performance
of our methods. The methods are applied to the Cystic Fibrosis Patient Registry
(CFFPR) data to perform dynamic prediction of lung disease in cystic fibrosis
patients and to identify important prognosis factors.
| stat |
Partitioned hybrid learning of Bayesian network structures | We develop a novel hybrid method for Bayesian network structure learning
called partitioned hybrid greedy search (pHGS), composed of three distinct yet
compatible new algorithms: Partitioned PC (pPC) accelerates skeleton learning
via a divide-and-conquer strategy, $p$-value adjacency thresholding (PATH)
effectively accomplishes parameter tuning with a single execution, and hybrid
greedy initialization (HGI) maximally utilizes constraint-based information to
obtain a high-scoring and well-performing initial graph for greedy search. We
establish structure learning consistency of our algorithms in the large-sample
limit, and empirically validate our methods individually and collectively
through extensive numerical comparisons. The combined merits of pPC and PATH
achieve significant computational reductions compared to the PC algorithm
without sacrificing the accuracy of estimated structures, and our generally
applicable HGI strategy reliably improves the estimation structural accuracy of
popular hybrid algorithms with negligible additional computational expense. Our
empirical results demonstrate the superior empirical performance of pHGS
against many state-of-the-art structure learning algorithms.
| stat |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.