title
stringlengths 6
244
| abstract
stringlengths 19
5.09k
| label
stringclasses 10
values |
|---|---|---|
A Data-Adaptive Targeted Learning Approach of Evaluating Viscoelastic
Assay Driven Trauma Treatment Protocols | Estimating the impact of trauma treatment protocols is complicated by the
high dimensional yet finite sample nature of trauma data collected from
observational studies. Viscoelastic assays are highly predictive measures of
hemostasis. However, the effectiveness of thromboelastography(TEG) based
treatment protocols has not been statistically evaluated.To conduct robust and
reliable estimation with sparse data, we built an estimation "machine" for
estimating causal impacts of candidate variables using the collaborative
targeted maximum loss-based estimation(CTMLE) framework.The computational
efficiency is achieved by using the scalable version of CTMLE such that the
covariates are pre-ordered by summary statistics of their importance before
proceeding to the estimation steps.To extend the application of the estimator
in practice, we used super learning in combination with CTMLE to flexibly
choose the best convex combination of algorithms. By selecting the optimal
covariates set in high dimension and reducing constraints in choosing
pre-ordering algorithms, we are able to construct a robust and data-adaptive
model to estimate the parameter of interest.Under this estimation framework,
CTMLE outperformed the other doubly robust estimators(IPW,AIPW,stabilized
IPW,TMLE) in the simulation study. CTMLE demonstrated very accurate estimation
of the target parameter (ATE). Applying CTMLE on the real trauma data, the
treatment protocol (using TEG values immediately after injury) showed
significant improvement in trauma patient hemostasis status (control of
bleeding), and a decrease in mortality rate at 6h compared to standard care.The
estimation results did not show significant change in mortality rate at 24h
after arrival.
| stat |
How does topology influence gradient propagation and model performance
of deep networks with DenseNet-type skip connections? | DenseNets introduce concatenation-type skip connections that achieve
state-of-the-art accuracy in several computer vision tasks. In this paper, we
reveal that the topology of the concatenation-type skip connections is closely
related to the gradient propagation which, in turn, enables a predictable
behavior of DNNs' test performance. To this end, we introduce a new metric
called NN-Mass to quantify how effectively information flows through DNNs.
Moreover, we empirically show that NN-Mass also works for other types of skip
connections, e.g., for ResNets, Wide-ResNets (WRNs), and MobileNets, which
contain addition-type skip connections (i.e., residuals or inverted residuals).
As such, for both DenseNet-like CNNs and ResNets/WRNs/MobileNets, our
theoretically grounded NN-Mass can identify models with similar accuracy,
despite having significantly different size/compute requirements. Detailed
experiments on both synthetic and real datasets (e.g., MNIST, CIFAR-10,
CIFAR-100, ImageNet) provide extensive evidence for our insights. Finally, the
closed-form equation of our NN-Mass enables us to design significantly
compressed DenseNets (for CIFAR-10) and MobileNets (for ImageNet) directly at
initialization without time-consuming training and/or searching.
| stat |
Invertible Particle Flow-based Sequential MCMC with extension to
Gaussian Mixture noise models | Sequential state estimation in non-linear and non-Gaussian state spaces has a
wide range of applications in statistics and signal processing. One of the most
effective non-linear filtering approaches, particle filtering, suffers from
weight degeneracy in high-dimensional filtering scenarios. Several avenues have
been pursued to address high-dimensionality. Among these, particle flow
particle filters construct effective proposal distributions by using invertible
flow to migrate particles continuously from the prior distribution to the
posterior, and sequential Markov chain Monte Carlo (SMCMC) methods use a
Metropolis-Hastings (MH) accept-reject approach to improve filtering
performance. In this paper, we propose to combine the strengths of invertible
particle flow and SMCMC by constructing a composite Metropolis-Hastings (MH)
kernel within the SMCMC framework using invertible particle flow. In addition,
we propose a Gaussian mixture model (GMM)-based particle flow algorithm to
construct effective MH kernels for multi-modal distributions. Simulation
results show that for high-dimensional state estimation example problems the
proposed kernels significantly increase the acceptance rate with minimal
additional computational overhead and improve estimation accuracy compared with
state-of-the-art filtering algorithms.
| stat |
A regularized hidden Markov model for analyzing the 'hot shoe' in
football | Although academic research on the 'hot hand' effect (in particular, in
sports, especially in basketball) has been going on for more than 30 years, it
still remains a central question in different areas of research whether such an
effect exists. In this contribution, we investigate the potential occurrence of
a 'hot shoe' effect for the performance of penalty takers in football based on
data from the German Bundesliga. For this purpose, we consider hidden Markov
models (HMMs) to model the (latent) forms of players. To further account for
individual heterogeneity of the penalty taker as well as the opponent's
goalkeeper, player-specific abilities are incorporated in the model formulation
together with a LASSO penalty. Our results suggest states which can be tied to
different forms of players, thus providing evidence for the hot shoe effect,
and shed some light on exceptionally well-performing goalkeepers, which are of
potential interest to managers and sports fans.
| stat |
Generic Machine Learning Inference on Heterogenous Treatment Effects in
Randomized Experiments | We propose strategies to estimate and make inference on key features of
heterogeneous effects in randomized experiments. These key features include
best linear predictors of the effects on machine learning proxies, average
effects sorted by impact groups, and average characteristics of most and least
impacted units. The approach is valid in high dimensional settings, where the
effects are proxied by machine learning methods. We post-process these proxies
into the estimates of the key features. Our approach is generic, it can be used
in conjunction with penalized methods, deep and shallow neural networks,
canonical and new random forests, boosted trees, and ensemble methods.
Estimation and inference are based on repeated data splitting to avoid
overfitting and achieve validity. For inference, we take medians of p-values
and medians of confidence intervals, resulting from many different data splits,
and then adjust their nominal level to guarantee uniform validity. This
variational inference method, which quantifies the uncertainty coming from both
parameter estimation and data splitting, is shown to be uniformly valid for a
large class of data generating processes. We illustrate the use of the approach
with a randomized field experiment that evaluated a combination of nudges to
stimulate demand for immunization in India.
| stat |
Highly Efficient Stepped Wedge Designs for Clusters of Unequal Size | The Stepped Wedge Design (SWD) is a form of cluster randomized trial, usually
comparing two treatments, which is divided into time periods and sequences,
with clusters allocated to sequences. Typically all sequences start with the
standard treatment and end with the new treatment, with the change happening at
different times in the different sequences. The clusters will usually differ in
size but this is overlooked in much of the existing literature. This paper
considers the case when clusters have different sizes and determines how
efficient designs can be found. The approach uses an approximation to the the
variance of the treatment effect which is expressed in terms of the proportions
of clusters and of individuals allocated to each sequence of the design. The
roles of these sets of proportions in determining an efficient design are
discussed and illustrated using two SWDs, one in the treatment of sexually
transmitted diseases and one in renal replacement therapy. Cluster-balanced
designs, which allocate equal numbers of clusters to each sequence, are shown
to have excellent statistical and practical properties; suggestions are made
about the practical application of the results for these designs. The paper
concentrates on the cross-sectional case, where subjects are measured once, but
it is briefly indicated how the methods can be extended to the closed-cohort
design.
| stat |
Ice Model Calibration Using Semi-continuous Spatial Data | Rapid changes in Earth's cryosphere caused by human activity can lead to
significant environmental impacts. Computer models provide a useful tool for
understanding the behavior and projecting the future of Arctic and Antarctic
ice sheets. However, these models are typically subject to large parametric
uncertainties due to poorly constrained model input parameters that govern the
behavior of simulated ice sheets. Computer model calibration provides a formal
statistical framework to infer parameters using observational data, and to
quantify the uncertainty in projections due to the uncertainty in these
parameters. Calibration of ice sheet models is often challenging because the
relevant model output and observational data take the form of semi-continuous
spatial data, with a point mass at zero and a right-skewed continuous
distribution for positive values. Current calibration approaches cannot handle
such data. Here we introduce a hierarchical latent variable model that handles
binary spatial patterns and positive continuous spatial patterns as separate
components. To overcome challenges due to high-dimensionality we use
likelihood-based generalized principal component analysis to impose
low-dimensional structures on the latent variables for spatial dependence. We
apply our methodology to calibrate a physical model for the Antarctic ice sheet
and demonstrate that we can overcome the aforementioned modeling and
computational challenges. As a result of our calibration, we obtain improved
future ice-volume change projections.
| stat |
Kumaraswamy autoregressive moving average models for double bounded
environmental data | In this paper we introduce the Kumaraswamy autoregressive moving average
models (KARMA), which is a dynamic class of models for time series taking
values in the double bounded interval $(a,b)$ following the Kumaraswamy
distribution. The Kumaraswamy family of distribution is widely applied in many
areas, especially hydrology and related fields. Classical examples are time
series representing rates and proportions observed over time. In the proposed
KARMA model, the median is modeled by a dynamic structure containing
autoregressive and moving average terms, time-varying regressors, unknown
parameters and a link function. We introduce the new class of models and
discuss conditional maximum likelihood estimation, hypothesis testing
inference, diagnostic analysis and forecasting. In particular, we provide
closed-form expressions for the conditional score vector and conditional Fisher
information matrix. An application to environmental real data is presented and
discussed.
| stat |
Learning Mixtures of Low-Rank Models | We study the problem of learning mixtures of low-rank models, i.e.
reconstructing multiple low-rank matrices from unlabelled linear measurements
of each. This problem enriches two widely studied settings -- low-rank matrix
sensing and mixed linear regression -- by bringing latent variables (i.e.
unknown labels) and structural priors (i.e. low-rank structures) into
consideration. To cope with the non-convexity issues arising from unlabelled
heterogeneous data and low-complexity structure, we develop a three-stage
meta-algorithm that is guaranteed to recover the unknown matrices with
near-optimal sample and computational complexities under Gaussian designs. In
addition, the proposed algorithm is provably stable against random noise. We
complement the theoretical studies with empirical evidence that confirms the
efficacy of our algorithm.
| stat |
Optimization of High-dimensional Simulation Models Using Synthetic Data | Simulation models are valuable tools for resource usage estimation and
capacity planning. In many situations, reliable data is not available. We
introduce the BuB simulator, which requires only the specification of plausible
intervals for the simulation parameters. By performing a surrogate-model based
optimization, improved simulation model parameters can be determined.
Furthermore, a detailed statistical analysis can be performed, which allows
deep insights into the most important model parameters and their interactions.
This information can be used to screen the parameters that should be further
investigated. To exemplify our approach, a capacity and resource planning task
for a hospital was simulated and optimized. The study explicitly covers
difficulties caused by the COVID-19 pandemic. It can be shown, that even if
only limited real-world data is available, the BuB simulator can be
beneficially used to consider worst- and best-case scenarios. The BuB simulator
can be extended in many ways, e.g., by adding further resources (personal
protection equipment, staff, pharmaceuticals) or by specifying several cohorts
(based on age, health status, etc.).
Keywords: Synthetic data, discrete-event simulation, surrogate-model-based
optimization, COVID-19, machine learning, artificial intelligence, hospital
resource planning, prediction tool, capacity planning.
| stat |
Theoretical Guarantees for Bridging Metric Measure Embedding and Optimal
Transport | We propose a novel approach for comparing distributions whose supports do not
necessarily lie on the same metric space. Unlike Gromov-Wasserstein (GW)
distance which compares pairwise distances of elements from each distribution,
we consider a method allowing to embed the metric measure spaces in a common
Euclidean space and compute an optimal transport (OT) on the embedded
distributions. This leads to what we call a sub-embedding robust Wasserstein
(SERW) distance. Under some conditions, SERW is a distance that considers an OT
distance of the (low-distorted) embedded distributions using a common metric.
In addition to this novel proposal that generalizes several recent OT works,
our contributions stand on several theoretical analyses: (i) we characterize
the embedding spaces to define SERW distance for distribution alignment; (ii)
we prove that SERW mimics almost the same properties of GW distance, and we
give a cost relation between GW and SERW. The paper also provides some
numerical illustrations of how SERW behaves on matching problems.
| stat |
Beyond Lazy Training for Over-parameterized Tensor Decomposition | Over-parametrization is an important technique in training neural networks.
In both theory and practice, training a larger network allows the optimization
algorithm to avoid bad local optimal solutions. In this paper we study a
closely related tensor decomposition problem: given an $l$-th order tensor in
$(R^d)^{\otimes l}$ of rank $r$ (where $r\ll d$), can variants of gradient
descent find a rank $m$ decomposition where $m > r$? We show that in a lazy
training regime (similar to the NTK regime for neural networks) one needs at
least $m = \Omega(d^{l-1})$, while a variant of gradient descent can find an
approximate tensor when $m = O^*(r^{2.5l}\log d)$. Our results show that
gradient descent on over-parametrized objective could go beyond the lazy
training regime and utilize certain low-rank structure in the data.
| stat |
Statistical inference of probabilistic origin-destination demand using
day-to-day traffic data | Recent transportation network studies on uncertainty and reliability call for
modeling the probabilistic O-D demand and probabilistic network flow. Making
the best use of day-to-day traffic data collected over many years, this paper
develops a novel theoretical framework for estimating the mean and
variance/covariance matrix of O-D demand considering the day-to-day variation
induced by travelers' independent route choices. It also estimates the
probability distributions of link/path flow and their travel cost where the
variance stems from three sources, O-D demand, route choice and unknown errors.
The framework estimates O-D demand mean and variance/covariance matrix
iteratively, also known as iterative generalized least squares (IGLS) in
statistics. Lasso regularization is employed to obtain sparse covariance matrix
for better interpretation and computational efficiency. Though the
probabilistic O-D estimation (ODE) works with a much larger solution space than
the deterministic ODE, we show that its estimator for O-D demand mean is no
worse than the best possible estimator by an error that reduces with the
increase in sample size. The probabilistic ODE is examined on two small
networks and two real-world large-scale networks. The solution converges
quickly under the IGLS framework. In all those experiments, the results of the
probabilistic ODE are compelling, satisfactory and computationally plausible.
Lasso regularization on the covariance matrix estimation leans to underestimate
most of variance/covariance entries. A proper Lasso penalty ensures a good
trade-off between bias and variance of the estimation.
| stat |
Consistently estimating graph statistics using Aggregated Relational
Data | Aggregated Relational Data, known as ARD, capture information about a social
network by asking about the number of connections between a person and a group
with a particular characteristic, rather than asking about connections between
each pair of individuals directly. Breza et al. (Forthcoming) and McCormick and
Zheng (2015) relate ARD questions, consisting of survey items of the form "How
many people with characteristic X do you know?" to parametric statistical
models for complete graphs. In this paper, we propose criteria for consistent
estimation of individual and graph level statistics from ARD data.
| stat |
A Novel Statistical Test for Treatment Differences in Clinical Trials
using a Response Adaptive Forward Looking Gittins Index Rule | The most common objective for response adaptive clinical trials is to seek to
ensure that patients within a trial have a high chance of receiving the best
treatment available by altering the chance of allocation on the basis of
accumulating data. Approaches which yield good patient benefit properties
suffer from low power from a frequentist perspective when testing for a
treatment difference at the end of the study due to the high imbalance in
treatment allocations. In this work we develop an alternative pairwise test for
treatment difference on the basis of allocation probabilities of the
covariate-adjusted response-adaptive randomization with forward looking Gittins
index rule (CARA-FLGI). The performance of the novel test is evaluated in
simulations for two-armed studies and then its applications to multi-armed
studies is illustrated. The proposed test has markedly improved power over the
traditional Fisher exact test when this class of non-myopic response adaptation
is used. We also find that the test's power is close to the power of a Fisher
exact test under equal randomization.
| stat |
Compressing Large Sample Data for Discriminant Analysis | Large-sample data became prevalent as data acquisition became cheaper and
easier. While a large sample size has theoretical advantages for many
statistical methods, it presents computational challenges. Sketching, or
compression, is a well-studied approach to address these issues in regression
settings, but considerably less is known about its performance in
classification settings. Here we consider the computational issues due to large
sample size within the discriminant analysis framework. We propose a new
compression approach for reducing the number of training samples for linear and
quadratic discriminant analysis, in contrast to existing compression methods
which focus on reducing the number of features. We support our approach with a
theoretical bound on the misclassification error rate compared to the Bayes
classifier. Empirical studies confirm the significant computational gains of
the proposed method and its superior predictive ability compared to random
sub-sampling.
| stat |
Success-Odds: An improved Win-Ratio | Multiple and combined endpoints involving also non-normal outcomes appear in
many clinical trials in various areas in medicine. In some cases, the outcome
can be observed only on an ordinal or dichotomous scale. Then the success of
two therapies is assessed by comparing the outcome of two randomly selected
patients from the two therapy groups by 'better', 'equal' or 'worse'. These
outcomes can be described by the probabilities $p^-=P(X<Y)$, $p_0=P(X=Y)$, and
$p^+ =P(X~>~Y)$. For a clinician, however, these quantities are less intuitive.
Therefore, Noether (1987) introduced the quantity $\lambda=p^+ / p^-$ assuming
continuous distributions. The same quantity was used by Pocock et al. (2012)
and by Wang and Pocock (2016) also for general non-normal outcomes and has been
called 'win-ratio' $\lambda_{WR}$. Unlike Noether (1987), Wang and Pocock
(2016) explicitly allowed for ties in the data. It is the aim of this
manuscript to investigate the properties of $\lambda_{WR}$ in case of ties. It
turns out that it has the strange property of becoming larger if the data are
observed less accurately, i.e. include more ties. Thus, in case of ties, the
win-ratio looses its appealing property to describe and quantify an intuitive
and well interpretable treatment effect. Therefore, a slight modification of
$\lambda_{WR} = \theta / (1-\theta)$ is suggested, namely the so-called
'success-odds' where $\theta=p^+ + \frac12 p_0$ is called a success of a
therapy if $\theta>\frac12$. In the case of no ties, $\lambda_{SO}$ is
identical to $\lambda_{WR}$. A test for the hypothesis $\lambda_{SO}=1$ and
range preserving confidence intervals for $\lambda_{SO}$ are derived. By two
counterexamples it is demonstrated that generalizations of both the win-ratio
and the success-odds to more than two treatments or to stratified designs are
not straightforward and need more detailed considerations.
| stat |
Estimation of conditional mixture Weibull distribution with
right-censored data using neural network for time-to-event analysis | In this paper, we consider survival analysis with right-censored data which
is a common situation in predictive maintenance and health field. We propose a
model based on the estimation of two-parameter Weibull distribution
conditionally to the features. To achieve this result, we describe a neural
network architecture and the associated loss functions that takes into account
the right-censored data. We extend the approach to a finite mixture of
two-parameter Weibull distributions. We first validate that our model is able
to precisely estimate the right parameters of the conditional Weibull
distribution on synthetic datasets. In numerical experiments on two real-word
datasets (METABRIC and SEER), our model outperforms the state-of-the-art
methods. We also demonstrate that our approach can consider any survival time
horizon.
| stat |
Convergent stochastic algorithm for parameter estimation in frailty
models using integrated partial likelihood | Frailty models are often the model of choice for heterogeneous survival data.
A frailty model contains both random effects and fixed effects, with the random
effects accommodating for the correlation in the data. Different estimation
procedures have been proposed for the fixed effects and the variances of and
covariances between the random effects. Especially with an unspecified baseline
hazard, i.e., the Cox model, the few available methods deal only with a
specific correlation structure. In this paper, an estimation procedure, based
on the integrated partial likelihood, is introduced, which can generally deal
with any kind of correlation structure. The new approach, namely the
maximisation of the integrated partial likelihood, combined with a stochastic
estimation procedure allows also for a wide choice of distributions for the
random effects. First, we demonstrate the almost sure convergence of the
stochastic algorithm towards a critical point of the integrated partial
likelihood. Second, numerical convergence properties are evaluated by
simulation. Third, the advantage of using an unspecified baseline hazard is
demonstrated through application on cancer clinical trial data.
| stat |
A hierarchical expected improvement method for Bayesian optimization | The Expected Improvement (EI) method, proposed by Jones et al. (1998), is a
widely-used Bayesian optimization method, which makes use of a fitted Gaussian
process model for efficient black-box optimization. However, one key drawback
of EI is that it is overly greedy in exploiting the fitted Gaussian process
model for optimization, which results in suboptimal solutions even with large
sample sizes. To address this, we propose a new hierarchical EI (HEI)
framework, which makes use of a hierarchical Gaussian process model. HEI
preserves a closed-form acquisition function, and corrects the over-greediness
of EI by encouraging exploration of the optimization space. We then introduce
hyperparameter estimation methods which allow HEI to mimic a fully Bayesian
optimization procedure, while avoiding expensive Markov-chain Monte Carlo
sampling steps. We prove the global convergence of HEI over a broad function
space, and establish near-minimax convergence rates under certain prior
specifications. Numerical experiments show the improvement of HEI over existing
Bayesian optimization methods, for synthetic functions and a semiconductor
manufacturing optimization problem.
| stat |
Mixed Variable Bayesian Optimization with Frequency Modulated Kernels | The sample efficiency of Bayesian optimization(BO) is often boosted by
Gaussian Process(GP) surrogate models. However, on mixed variable spaces,
surrogate models other than GPs are prevalent, mainly due to the lack of
kernels which can model complex dependencies across different types of
variables. In this paper, we propose the frequency modulated (FM) kernel
flexibly modeling dependencies among different types of variables, so that BO
can enjoy the further improved sample efficiency. The FM kernel uses distances
on continuous variables to modulate the graph Fourier spectrum derived from
discrete variables. However, the frequency modulation does not always define a
kernel with the similarity measure behavior which returns higher values for
pairs of more similar points. Therefore, we specify and prove conditions for FM
kernels to be positive definite and to exhibit the similarity measure behavior.
In experiments, we demonstrate the improved sample efficiency of GP BO using FM
kernels (BO-FM).On synthetic problems and hyperparameter optimization problems,
BO-FM outperforms competitors consistently. Also, the importance of the
frequency modulation principle is empirically demonstrated on the same
problems. On joint optimization of neural architectures and SGD
hyperparameters, BO-FM outperforms competitors including Regularized
evolution(RE) and BOHB. Remarkably, BO-FM performs better even than RE and BOHB
using three times as many evaluations.
| stat |
Neural Network Based Parameter Estimation Method for the Pareto/NBD
Model | Whether stochastic or parametric, the Pareto/NBD model can only be utilized
for an in-sample prediction rather than an out-of-sample prediction. This
research thus provides a neural network based extension of the Pareto/NBD model
to estimate the out-of-sample parameters, which overrides the estimation burden
and the application dilemma of the Pareto/NBD approach. The empirical results
indicate that the Pareto/NBD model and neural network algorithms have similar
predictability for identifying inactive customers. Even with a strong trend
fitting on the customer count of each repeat purchase point, the Pareto/NBD
model underestimates repeat purchases at both the individual and aggregate
levels. Nonetheless, when embedding the likelihood function of the Pareto/NBD
model into the loss function, the proposed parameter estimation method shows
extraordinary predictability on repeat purchases at these two levels.
Furthermore, the proposed neural network based method is highly efficient and
resource-friendly and can be deployed in cloud computing to handle with big
data analysis.
| stat |
Random Features for Kernel Approximation: A Survey on Algorithms,
Theory, and Beyond | Random features is one of the most popular techniques to speed up kernel
methods in large-scale problems. Related works have been recognized by the
NeurIPS Test-of-Time award in 2017 and the ICML Best Paper Finalist in 2019.
The body of work on random features has grown rapidly, and hence it is
desirable to have a comprehensive overview on this topic explaining the
connections among various algorithms and theoretical results. In this survey,
we systematically review the work on random features from the past ten years.
First, the motivations, characteristics and contributions of representative
random features based algorithms are summarized according to their sampling
schemes, learning procedures, variance reduction properties and how they
exploit training data. Second, we review theoretical results that center around
the following key question: how many random features are needed to ensure a
high approximation quality or no loss in the empirical/expected risks of the
learned estimator. Third, we provide a comprehensive evaluation of popular
random features based algorithms on several large-scale benchmark datasets and
discuss their approximation quality and prediction performance for
classification. Last, we discuss the relationship between random features and
modern over-parameterized deep neural networks (DNNs), including the use of
high dimensional random features in the analysis of DNNs as well as the gaps
between current theoretical and empirical results. This survey may serve as a
gentle introduction to this topic, and as a users' guide for practitioners
interested in applying the representative algorithms and understanding
theoretical results under various technical assumptions. We hope that this
survey will facilitate discussion on the open problems in this topic, and more
importantly, shed light on future research directions.
| stat |
Bayesian Hierarchical Modeling on Covariance Valued Data | Analysis of structural and functional connectivity (FC) of human brains is of
pivotal importance for diagnosis of cognitive ability. The Human Connectome
Project (HCP) provides an excellent source of neural data across different
regions of interest (ROIs) of the living human brain. Individual specific data
were available from an existing analysis (Dai et al., 2017) in the form of time
varying covariance matrices representing the brain activity as the subjects
perform a specific task. As a preliminary objective of studying the
heterogeneity of brain connectomics across the population, we develop a
probabilistic model for a sample of covariance matrices using a scaled Wishart
distribution. We stress here that our data units are available in the form of
covariance matrices, and we use the Wishart distribution to create our
likelihood function rather than its more common usage as a prior on covariance
matrices. Based on empirical explorations suggesting the data matrices to have
low effective rank, we further model the center of the Wishart distribution
using an orthogonal factor model type decomposition. We encourage shrinkage
towards a low rank structure through a novel shrinkage prior and discuss
strategies to sample from the posterior distribution using a combination of
Gibbs and slice sampling. We extend our modeling framework to a dynamic setting
to detect change points. The efficacy of the approach is explored in various
simulation settings and exemplified on several case studies including our
motivating HCP data. We extend our modeling framework to a dynamic setting to
detect change points.
| stat |
Quantum Generalized Linear Models | Generalized linear models (GLM) are link function based statistical models.
Many supervised learning algorithms are extensions of GLMs and have link
functions built into the algorithm to model different outcome distributions.
There are two major drawbacks when using this approach in applications using
real world datasets. One is that none of the link functions available in the
popular packages is a good fit for the data. Second, it is computationally
inefficient and impractical to test all the possible distributions to find the
optimum one. In addition, many GLMs and their machine learning extensions
struggle on problems of overdispersion in Tweedie distributions. In this paper
we propose a quantum extension to GLM that overcomes these drawbacks. A quantum
gate with non-Gaussian transformation can be used to continuously deform the
outcome distribution from known results. In doing so, we eliminate the need for
a link function. Further, by using an algorithm that superposes all possible
distributions to collapse to fit a dataset, we optimize the model in a
computationally efficient way. We provide an initial proof-of-concept by
testing this approach on both a simulation of overdispersed data and then on a
benchmark dataset, which is quite overdispersed, and achieved state of the art
results. This is a game changer in several applied fields, such as part failure
modeling, medical research, actuarial science, finance and many other fields
where Tweedie regression and overdispersion are ubiquitous.
| stat |
A Spatial Concordance Correlation Coefficient with an Application to
Image Analysis | In this work we define a spatial concordance coefficient for second-order
stationary processes. This problem has been widely addressed in a non-spatial
context, but here we consider a coefficient that for a fixed spatial lag allows
one to compare two spatial sequences along a 45-degree line. The proposed
coefficient was explored for the bivariate Mat\'ern and Wendland covariance
functions. The asymptotic normality of a sample version of the spatial
concordance coefficient for an increasing domain sampling framework was
established for the Wendland covariance function. To work with large digital
images, we developed a local approach for estimating the concordance that uses
local spatial models on non-overlapping windows. Monte Carlo simulations were
used to gain additional insights into the asymptotic properties for finite
sample sizes. As an illustrative example, we applied this methodology to two
similar images of a deciduous forest canopy. The images were recorded with
different cameras but similar fields-of-view and within minutes of each other.
Our analysis showed that the local approach helped to explain a percentage of
the non-spatial concordance and to provided additional information about its
decay as a function of the spatial lag.
| stat |
The Role of Neural Network Activation Functions | A wide variety of activation functions have been proposed for neural
networks. The Rectified Linear Unit (ReLU) is especially popular today. There
are many practical reasons that motivate the use of the ReLU. This paper
provides new theoretical characterizations that support the use of the ReLU,
its variants such as the leaky ReLU, as well as other activation functions in
the case of univariate, single-hidden layer feedforward neural networks. Our
results also explain the importance of commonly used strategies in the design
and training of neural networks such as "weight decay" and "path-norm"
regularization, and provide a new justification for the use of "skip
connections" in network architectures. These new insights are obtained through
the lens of spline theory. In particular, we show how neural network training
problems are related to infinite-dimensional optimizations posed over Banach
spaces of functions whose solutions are well-known to be fractional and
polynomial splines, where the particular Banach space (which controls the order
of the spline) depends on the choice of activation function.
| stat |
Estimation of the Parameters of Vector Autoregressive (VAR) Time Series
Model with Symmetric Stable Noise | In this article, we propose the fractional lower order covariance method
(FLOC) for estimating the parameters of vector autoregressive process (VAR) of
order $p$, $p\geq 1$ with symmetric stable noise. Further, we show the
efficiency, accuracy, and simplicity of our methods through Monte-Carlo
simulation.
| stat |
Critical Window Variable Selection for Mixtures: Estimating the Impact
of Multiple Air Pollutants on Stillbirth | Understanding the role of time-varying pollution mixtures on human health is
critical as people are simultaneously exposed to multiple pollutants during
their lives. For vulnerable sub-populations who have well-defined exposure
periods (e.g., pregnant women), questions regarding critical windows of
exposure to these mixtures are important for mitigating harm. We extend
Critical Window Variable Selection (CWVS) to the multipollutant setting by
introducing CWVS for Mixtures (CWVSmix), a hierarchical Bayesian method that
combines smoothed variable selection and temporally correlated weight
parameters to (i) identify critical windows of exposure to mixtures of
time-varying pollutants, (ii) estimate the time-varying relative importance of
each individual pollutant and their first order interactions within the
mixture, and (iii) quantify the impact of the mixtures on health. Through
simulation, we show that CWVSmix offers the best balance of performance in each
of these categories in comparison to competing methods. Using these approaches,
we investigate the impact of exposure to multiple ambient air pollutants on the
risk of stillbirth in New Jersey, 2005-2014. We find consistent elevated risk
in gestational weeks 2, 16-17, and 20 for non-Hispanic Black mothers, with
pollution mixtures dominated by ammonium (weeks 2, 17, 20), nitrate (weeks 2,
17), nitrogen oxides (weeks 2, 16), PM2.5 (week 2), and sulfate (week 20). The
method is available in the R package CWVSmix.
| stat |
Multi-Fidelity High-Order Gaussian Processes for Physical Simulation | The key task of physical simulation is to solve partial differential
equations (PDEs) on discretized domains, which is known to be costly. In
particular, high-fidelity solutions are much more expensive than low-fidelity
ones. To reduce the cost, we consider novel Gaussian process (GP) models that
leverage simulation examples of different fidelities to predict
high-dimensional PDE solution outputs. Existing GP methods are either not
scalable to high-dimensional outputs or lack effective strategies to integrate
multi-fidelity examples. To address these issues, we propose Multi-Fidelity
High-Order Gaussian Process (MFHoGP) that can capture complex correlations both
between the outputs and between the fidelities to enhance solution estimation,
and scale to large numbers of outputs. Based on a novel nonlinear
coregionalization model, MFHoGP propagates bases throughout fidelities to fuse
information, and places a deep matrix GP prior over the basis weights to
capture the (nonlinear) relationships across the fidelities. To improve
inference efficiency and quality, we use bases decomposition to largely reduce
the model parameters, and layer-wise matrix Gaussian posteriors to capture the
posterior dependency and to simplify the computation. Our stochastic
variational learning algorithm successfully handles millions of outputs without
extra sparse approximations. We show the advantages of our method in several
typical applications.
| stat |
Double Robust Representation Learning for Counterfactual Prediction | Causal inference, or counterfactual prediction, is central to decision making
in healthcare, policy and social sciences. To de-bias causal estimators with
high-dimensional data in observational studies, recent advances suggest the
importance of combining machine learning models for both the propensity score
and the outcome function. We propose a novel scalable method to learn
double-robust representations for counterfactual predictions, leading to
consistent causal estimation if the model for either the propensity score or
the outcome, but not necessarily both, is correctly specified. Specifically, we
use the entropy balancing method to learn the weights that minimize the
Jensen-Shannon divergence of the representation between the treated and control
groups, based on which we make robust and efficient counterfactual predictions
for both individual and average treatment effects. We provide theoretical
justifications for the proposed method. The algorithm shows competitive
performance with the state-of-the-art on real world and synthetic data.
| stat |
Complexity, Statistical Risk, and Metric Entropy of Deep Nets Using
Total Path Variation | For any ReLU network there is a representation in which the sum of the
absolute values of the weights into each node is exactly $1$, and the input
layer variables are multiplied by a value $V$ coinciding with the total
variation of the path weights. Implications are given for Gaussian complexity,
Rademacher complexity, statistical risk, and metric entropy, all of which are
shown to be proportional to $V$. There is no dependence on the number of nodes
per layer, except for the number of inputs $d$. For estimation with
sub-Gaussian noise, the mean square generalization error bounds that can be
obtained are of order $V \sqrt{L + \log d}/\sqrt{n}$, where $L$ is the number
of layers and $n$ is the sample size.
| stat |
Individualized Multilayer Tensor Learning with An Application in Imaging
Analysis | This work is motivated by multimodality breast cancer imaging data, which is
quite challenging in that the signals of discrete tumor-associated
microvesicles (TMVs) are randomly distributed with heterogeneous patterns. This
imposes a significant challenge for conventional imaging regression and
dimension reduction models assuming a homogeneous feature structure. We develop
an innovative multilayer tensor learning method to incorporate heterogeneity to
a higher-order tensor decomposition and predict disease status effectively
through utilizing subject-wise imaging features and multimodality information.
Specifically, we construct a multilayer decomposition which leverages an
individualized imaging layer in addition to a modality-specific tensor
structure. One major advantage of our approach is that we are able to
efficiently capture the heterogeneous spatial features of signals that are not
characterized by a population structure as well as integrating multimodality
information simultaneously. To achieve scalable computing, we develop a new
bi-level block improvement algorithm. In theory, we investigate both the
algorithm convergence property, tensor signal recovery error bound and
asymptotic consistency for prediction model estimation. We also apply the
proposed method for simulated and human breast cancer imaging data. Numerical
results demonstrate that the proposed method outperforms other existing
competing methods.
| stat |
Spatial Tweedie exponential dispersion models | This paper proposes a general modeling framework that allows for uncertainty
quantification at the individual covariate level and spatial referencing,
operating withing a double generalized linear model (DGLM). DGLMs provide a
general modeling framework allowing dispersion to depend in a link-linear
fashion on chosen covariates. We focus on working with Tweedie exponential
dispersion models while considering DGLMs, the reason being their recent
wide-spread use for modeling mixed response types. Adopting a regularization
based approach, we suggest a class of flexible convex penalties derived from an
un-directed graph that facilitates estimation of the unobserved spatial effect.
Developments are concisely showcased by proposing a co-ordinate descent
algorithm that jointly explains variation from covariates in mean and
dispersion through estimation of respective model coefficients while estimating
the unobserved spatial effect. Simulations performed show that proposed
approach is superior to competitors like the ridge and un-penalized versions.
Finally, a real data application is considered while modeling insurance losses
arising from automobile collisions in the state of Connecticut, USA for the
year 2008.
| stat |
Exact Multivariate Two-Sample Density-Based Empirical Likelihood Ratio
Tests Applicable to Retrospective and Group Sequential Studies | Nonparametric tests for equality of multivariate distributions are frequently
desired in research. It is commonly required that test-procedures based on
relatively small samples of vectors accurately control the corresponding Type I
Error (TIE) rates. Often, in the multivariate testing, extensions of
null-distribution-free univariate methods, e.g., Kolmogorov-Smirnov and
Cramer-von Mises type schemes, are not exact, since their null distributions
depend on underlying data distributions. The present paper extends the
density-based empirical likelihood technique in order to nonparametrically
approximate the most powerful test for the multivariate two-sample (MTS)
problem, yielding an exact finite-sample test statistic. We rigorously
establish and apply one-to-one-mapping between the equality of vectors
distributions and the equality of distributions of relevant univariate linear
projections. In this framework, we prove an algorithm that simplifies the use
of projection pursuit, employing only a few of the infinitely many linear
combinations of observed vectors components. The displayed distribution-free
strategy is employed in retrospective and group sequential manners. The
asymptotic consistency of the proposed technique is shown. Monte Carlo studies
demonstrate that the proposed procedures exhibit extremely high and stable
power characteristics across a variety of settings. Supplementary materials for
this article are available online.
| stat |
Extending classical surrogate modelling to high-dimensions through
supervised dimensionality reduction: a data-driven approach | Thanks to their versatility, ease of deployment and high-performance,
surrogate models have become staple tools in the arsenal of uncertainty
quantification (UQ). From local interpolants to global spectral decompositions,
surrogates are characterised by their ability to efficiently emulate complex
computational models based on a small set of model runs used for training. An
inherent limitation of many surrogate models is their susceptibility to the
curse of dimensionality, which traditionally limits their applicability to a
maximum of $\mathcal{O}(10^2)$ input dimensions. We present a novel approach at
high-dimensional surrogate modelling that is model-, dimensionality reduction-
and surrogate model- agnostic (black box), and can enable the solution of high
dimensional (i.e. up to $\mathcal{O}(10^4)$) problems. After introducing the
general algorithm, we demonstrate its performance by combining Kriging and
polynomial chaos expansions surrogates and kernel principal component analysis.
In particular, we compare the generalisation performance that the resulting
surrogates achieve to the classical sequential application of dimensionality
reduction followed by surrogate modelling on several benchmark applications,
comprising an analytical function and two engineering applications of
increasing dimensionality and complexity.
| stat |
Uncertainty modelling and computational aspects of data association | A novel solution to the smoothing problem for multi-object dynamical systems
is proposed and evaluated. The systems of interest contain an unknown and
varying number of dynamical objects that are partially observed under noisy and
corrupted observations. An alternative representation of uncertainty is
considered in order to account for the lack of information about the different
aspects of this type of complex system. The corresponding statistical model can
be formulated as a hierarchical model consisting of conditionally-independent
hidden Markov models. This particular structure is leveraged to propose an
efficient method in the context of Markov chain Monte Carlo (MCMC) by relying
on an approximate solution to the corresponding filtering problem, in a similar
fashion to particle MCMC. This approach is shown to outperform existing
algorithms in a range of scenarios.
| stat |
Graphical continuous Lyapunov models | The linear Lyapunov equation of a covariance matrix parametrizes the
equilibrium covariance matrix of a stochastic process. This parametrization can
be interpreted as a new graphical model class, and we show how the model class
behaves under marginalization and introduce a method for structure learning via
$\ell_1$-penalized loss minimization. Our proposed method is demonstrated to
outperform alternative structure learning algorithms in a simulation study, and
we illustrate its application for protein phosphorylation network
reconstruction.
| stat |
The Building Data Genome Project 2, energy meter data from the ASHRAE
Great Energy Predictor III competition | This paper describes an open data set of 3,053 energy meters from 1,636
non-residential buildings with a range of two full years (2016 and 2017) at an
hourly frequency (17,544 measurements per meter resulting in approximately 53.6
million measurements). These meters were collected from 19 sites across North
America and Europe, with one or more meters per building measuring whole
building electrical, heating and cooling water, steam, and solar energy as well
as water and irrigation meters. Part of these data were used in the Great
Energy Predictor III (GEPIII) competition hosted by the ASHRAE organization in
October-December 2019. GEPIII was a machine learning competition for long-term
prediction with an application to measurement and verification. This paper
describes the process of data collection, cleaning, and convergence of
time-series meter data, the meta-data about the buildings, and complementary
weather data. This data set can be used for further prediction benchmarking and
prototyping as well as anomaly detection, energy analysis, and building type
classification.
| stat |
On estimation and inference in latent structure random graphs | We define a latent structure model (LSM) random graph as a random dot product
graph (RDPG) in which the latent position distribution incorporates both
probabilistic and geometric constraints, delineated by a family of underlying
distributions on some fixed Euclidean space, and a structural support
submanifold from which the latent positions for the graph are drawn. For a
one-dimensional latent structure model with known structural support, we show
how spectral estimates of the latent positions of an RDPG can be used for
efficient estimation of the paramaters of the LSM. We describe how to estimate
or learn the structural support in cases where it is unknown, with an
illustrative focus on graphs with latent positions along the Hardy-Weinberg
curve. Finally, we use the latent structure model formulation to test bilateral
homology in the Drosophila connectome.
| stat |
The Exact Asymptotic Form of Bayesian Generalization Error in Latent
Dirichlet Allocation | Latent Dirichlet allocation (LDA) obtains essential information from data by
using Bayesian inference. It is applied to knowledge discovery via dimension
reducing and clustering in many fields. However, its generalization error had
not been yet clarified since it is a singular statistical model where there is
no one-to-one mapping from parameters to probability distributions. In this
paper, we give the exact asymptotic form of its generalization error and
marginal likelihood, by theoretical analysis of its learning coefficient using
algebraic geometry. The theoretical result shows that the Bayesian
generalization error in LDA is expressed in terms of that in matrix
factorization and a penalty from the simplex restriction of LDA's parameter
region. A numerical experiment is consistent to the theoretical result.
| stat |
Spectral graph clustering via the Expectation-Solution algorithm | The stochastic blockmodel (SBM) models the connectivity within and between
disjoint subsets of nodes in networks. Prior work demonstrated that the rows of
an SBM's adjacency spectral embedding (ASE) and Laplacian spectral embedding
(LSE) both converge in law to Gaussian mixtures where the components are curved
exponential families. Maximum likelihood estimation via the
Expectation-Maximization (EM) algorithm for a full Gaussian mixture model (GMM)
can then perform the task of clustering graph nodes, albeit without appealing
to the components' curvature. Noting that EM is a special case of the
Expectation-Solution (ES) algorithm, we propose two ES algorithms that allow us
to take full advantage of these curved structures. After presenting the ES
algorithm for the general curved-Gaussian mixture, we develop those
corresponding to the ASE and LSE limiting distributions. Simulating from
artificial SBMs and a brain connectome SBM reveals that clustering graph nodes
via our ES algorithms improves upon that of EM for a full GMM.
| stat |
Covariate adjustment in subgroup analyses of randomized clinical trials:
A propensity score approach | Background: Subgroup analyses are frequently conducted in randomized clinical
trials to assess evidence of heterogeneous treatment effect across patient
subpopulations. Although randomization balances covariates within subgroups in
expectation, chance imbalance may be amplified in small subgroups and harm the
precision. Covariate adjustment in overall analysis of RCT is often conducted,
via either ANCOVA or propensity score weighting, but for subgroup analysis has
been rarely discussed. In this article, we develop propensity score weighting
methodology for covariate adjustment to improve the precision and power of
subgroup analyses in RCTs. Methods: We extend the propensity score weighting
methodology to subgroup analyses by fitting a logistic propensity model with
pre-specified covariate-subgroup interactions. We show that, by construction,
overlap weighting exactly balances the covariates with interaction terms in
each subgroup. Extensive simulations were performed to compare the operating
characteristics of unadjusted, different propensity score weighting and the
ANCOVA estimator. We apply these methods to the HF-ACTION trial to evaluate the
effect of exercise training on 6-minute walk test in pre-specified subgroups.
Results: Standard errors of the adjusted estimators are smaller than those of
the unadjusted estimator. The propensity score weighting estimator is as
efficient as ANCOVA, and is often more efficient when subgroup sample size is
small (e.g.<125), and/or when outcome model is misspecified. The weighting
estimators with full-interaction propensity model consistently outperform the
standard main-effect propensity model. Conclusion: Propensity score weighting
is a transparent and objective method to adjust chance imbalance of important
covariates in subgroup analyses of RCTs. It is crucial to include the full
covariate-subgroup interactions in the propensity score model.
| stat |
Gap-Measure Tests with Applications to Data Integrity Verification | In this paper we propose and examine gap statistics for assessing uniform
distribution hypotheses. We provide examples relevant to data integrity testing
for which max-gap statistics provide greater sensitivity than chi-square
($\chi^2$), thus allowing the new test to be used in place of or as a
complement to $\chi^2$ testing for purposes of distinguishing a larger class of
deviations from uniformity. We establish that the proposed max-gap test has the
same sequential and parallel computational complexity as $\chi^2$ and thus is
applicable for Big Data analytics and integrity verification.
| stat |
Optimized Realization of Bayesian Networks in Reduced Normal Form using
Latent Variable Model | Bayesian networks in their Factor Graph Reduced Normal Form (FGrn) are a
powerful paradigm for implementing inference graphs. Unfortunately, the
computational and memory costs of these networks may be considerable, even for
relatively small networks, and this is one of the main reasons why these
structures have often been underused in practice. In this work, through a
detailed algorithmic and structural analysis, various solutions for cost
reduction are proposed. An online version of the classic batch learning
algorithm is also analyzed, showing very similar results (in an unsupervised
context); which is essential even if multilevel structures are to be built. The
solutions proposed, together with the possible online learning algorithm, are
included in a C++ library that is quite efficient, especially if compared to
the direct use of the well-known sum-product and Maximum Likelihood (ML)
algorithms. The results are discussed with particular reference to a Latent
Variable Model (LVM) structure.
| stat |
Design and Analysis of Switchback Experiments | In switchback experiments, a firm sequentially exposes an experimental unit
to a random treatment, measures its response, and repeats the procedure for
several periods to determine which treatment leads to the best outcome.
Although practitioners have widely adopted this experimental design technique,
the development of its theoretical properties and the derivation of optimal
design procedures have been, to the best of our knowledge, elusive. In this
paper, we address these limitations by establishing the necessary results to
ensure that practitioners can apply this powerful class of experiments with
minimal assumptions. Our main result is the derivation of the optimal design of
switchback experiments under a range of different assumptions on the order of
carryover effect - that is, the length of time a treatment persists in
impacting the outcome. We cast the experimental design problem as a minimax
discrete robust optimization problem, identify the worst-case adversarial
strategy, establish structural results for the optimal design, and finally
solve the problem via a continuous relaxation. For the optimal design, we
derive two approaches for performing inference after running the experiment.
The first provides exact randomization based $p$-values and the second uses a
finite population central limit theorem to conduct conservative hypothesis
tests and build confidence intervals. We further provide theoretical results
for our inferential procedures when the order of the carryover effect is
misspecified. For firms that possess the capability to run multiple switchback
experiments, we also provide a data-driven strategy to identify the likely
order of carryover effect. To study the empirical properties of our results, we
conduct extensive simulations. We conclude the paper by providing some
practical suggestions.
| stat |
Modeling sign concordance of quantile regression residuals with multiple
outcomes | Quantile regression permits describing how quantiles of a scalar response
variable depend on a set of predictors. Because a unique definition of
multivariate quantiles is lacking, extending quantile regression to
multivariate responses is somewhat complicated. In this paper, we describe a
simple approach based on a two-step procedure: in the first step, quantile
regression is applied to each response separately; in the second step, the
joint distribution of the signs of the residuals is modeled through multinomial
regression. The described approach does not require a multidimensional
definition of quantiles, and can be used to capture important features of a
multivariate response and assess the effects of covariates on the correlation
structure. We apply the proposed method to analyze two different datasets.
| stat |
EWS-GCN: Edge Weight-Shared Graph Convolutional Network for
Transactional Banking Data | In this paper, we discuss how modern deep learning approaches can be applied
to the credit scoring of bank clients. We show that information about
connections between clients based on money transfers between them allows us to
significantly improve the quality of credit scoring compared to the approaches
using information about the target client solely. As a final solution, we
develop a new graph neural network model EWS-GCN that combines ideas of graph
convolutional and recurrent neural networks via attention mechanism. The
resulting model allows for robust training and efficient processing of
large-scale data. We also demonstrate that our model outperforms the
state-of-the-art graph neural networks achieving excellent results
| stat |
The Contribution Plot: Decomposition and Graphical Display of the RV
Coefficient, with Application to Genetic and Brain Imaging Biomarkers of
Alzheimer's Disease | Alzheimer's disease (AD) is a chronic neurodegenerative disease that causes
memory loss and decline in cognitive abilities. AD is the sixth leading cause
of death in the United States, affecting an estimated 5 million Americans. To
assess the association between multiple genetic variants and multiple
measurements of structural changes in the brain a recent study of AD used a
multivariate measure of linear dependence, the RV coefficient. The authors
decomposed the RV coefficient into contributions from individual variants and
displayed these contributions graphically. We investigate the properties of
such a `contribution plot' in terms of an underlying linear model, and discuss
estimation of the components of the plot when the correlation signal may be
sparse. The contribution plot is applied to simulated data and to genomic and
brain imaging data from the Alzheimer's Disease Neuroimaging Initiative.
| stat |
CircSpaceTime: an R package for spatial and spatio-temporal modeling of
Circular data | CircSpaceTime is the only R package currently available that implements
Bayesian models for spatial and spatio-temporal interpolation of circular data.
Such data are often found in applications where, among the many, wind
directions, animal movement directions, and wave directions are involved. To
analyze such data we need models for observations at locations s and times t,
as the so-called geostatistical models, providing structured dependence assumed
to decay in distance and time. The approach we take begins with Gaussian
processes defined for linear variables over space and time. Then, we use either
wrapping or projection to obtain processes for circular data. The models are
cast as hierarchical, with fitting and inference within a Bayesian framework.
Altogether, this package implements work developed by a series of papers; the
most relevant being Jona Lasinio, Gelfand, and Jona Lasinio (2012); Wang and
Gelfand (2014); Mastrantonio, Jona Lasinio, and Gelfand (2016). All procedures
are written using Rcpp. Estimates are obtained by MCMC allowing parallelized
multiple chains run. The implementation of the proposed models is considerably
improved on the simple routines adopted in the research papers. As original
running examples, for the spatial and spatio-temporal settings, we use wind
directions datasets over central Italy.
| stat |
Nonparametric semisupervised classification for signal detection in high
energy physics | Model-independent searches in particle physics aim at completing our
knowledge of the universe by looking for new possible particles not predicted
by the current theories. Such particles, referred to as signal, are expected to
behave as a deviation from the background, representing the known physics.
Information available on the background can be incorporated in the search, in
order to identify potential anomalies. From a statistical perspective, the
problem is recasted to a peculiar classification one where only partial
information is accessible. Therefore a semisupervised approach shall be
adopted, either by strengthening or by relaxing assumptions underlying
clustering or classification methods respectively. In this work, following the
first route, we semisupervise nonparametric clustering in order to identify a
possible signal. The main contribution consists in tuning a nonparametric
estimate of the density underlying the experimental data with the aid of the
available information on the physical theory. As a side contribution, a
variable selection procedure is presented. The whole procedure is tested on a
dataset mimicking proton-proton collisions performed within a particle
accelerator. While finding motivation in the field of particle physics, the
approach is applicable to various science domains, where similar problems of
anomaly detection arise.
| stat |
Identification of Dominant Features in Spatial Data | Dominant features of spatial data are connected structures or patterns that
emerge from location-based variation and manifest at specific scales or
resolutions. To identify dominant features, we propose a sequential application
of multiresolution decomposition and variogram function estimation.
Multiresolution decomposition separates data into additive components, and in
this way enables the recognition of their dominant features. A dedicated
multiresolution decomposition method is developed for arbitrary gridded spatial
data, where the underlying model includes a precision and spatial-weight matrix
to capture spatial correlation. The data are separated into their components by
smoothing on different scales, such that larger scales have longer spatial
correlation ranges. Moreover, our model can handle missing values, which is
often useful in applications. Variogram function estimation can be used to
describe properties in spatial data. Such functions are therefore estimated for
each component to determine its effective range, which assesses the
width-extent of the dominant feature. Finally, Bayesian analysis enables the
inference of identified dominant features and to judge whether these are
credibly different. The efficient implementation of the method relies mainly on
a sparse-matrix data structure and algorithms. By applying the method to
simulated data we demonstrate its applicability and theoretical soundness. In
disciplines that use spatial data, this method can lead to new insights, as we
exemplify by identifying the dominant features in a forest dataset. In that
application, the width-extents of the dominant features have an ecological
interpretation, namely the species interaction range, and their estimates
support the derivation of ecosystem properties such as biodiversity indices.
| stat |
Optimal acceptance sampling for modules F and F1 of the European
Measuring Instruments Directive | Acceptance sampling plans offered by ISO 2859-1 are far from optimal under
the conditions for statistical verification in modules F and F1 as prescribed
by Annex II of the Measuring Instruments Directive (MID) 2014/32/EU, resulting
in sample sizes that are larger than necessary. An optimised single-sampling
scheme is derived, both for large lots using the binomial distribution and for
finite-sized lots using the exact hypergeometric distribution, resulting in
smaller sample sizes that are economically more efficient while offering the
full statistical protection required by the MID.
| stat |
Double/Debiased Machine Learning for Logistic Partially Linear Model | We propose double/debiased machine learning approaches to infer (at the
parametric rate) the parametric component of a logistic partially linear model
with the binary response following a conditional logistic model of a low
dimensional linear parametric function of some key (exposure) covariates and a
nonparametric function adjusting for the confounding effect of other
covariates. We consider a Neyman orthogonal (doubly robust) score equation
consisting of two nuisance functions: nonparametric component in the logistic
model and conditional mean of the exposure on the other covariates and with the
response fixed. To estimate the nuisance models, we separately consider the use
of high dimensional (HD) sparse parametric models and more general (typically
nonparametric) machine learning (ML) methods. In the HD case, we derive certain
moment equations to calibrate the first-order bias of the nuisance models and
grant our method a model double robustness property in the sense that our
estimator achieves the desirable rate when at least one of the nuisance models
is correctly specified and both of them are ultra-sparse. In the ML case, the
non-linearity of the logit link makes it substantially harder than the
partially linear setting to use an arbitrary conditional mean learning
algorithm to estimate the nuisance component of the logistic model. We handle
this obstacle through a novel full model refitting procedure that is
easy-to-implement and facilitates the use of nonparametric ML algorithms in our
framework. Our ML estimator is rate doubly robust in the same sense as
Chernozhukov et al. (2018a). We evaluate our methods through simulation studies
and apply them in assessing the effect of emergency contraceptive (EC) pill on
early gestation foetal with a policy reform in Chile in 2008 (Bentancor and
Clarke, 2017).
| stat |
FATSO: A family of operators for variable selection in linear models | In linear models it is common to have situations where several regression
coefficients are zero. In these situations a common tool to perform regression
is a variable selection operator. One of the most common such operators is the
LASSO operator, which promotes point estimates which are zero. The LASSO
operator and similar approaches, however, give little in terms of easily
interpretable parameters to determine the degree of variable selectivity. In
this paper we propose a new family of selection operators which builds on the
geometry of LASSO but which yield an easily interpretable way to tune
selectivity. These operators correspond to Bayesian prior densities and hence
are suitable for Bayesian inference. We present some examples using simulated
and real data, with promising results.
| stat |
Bayesian Distance Weighted Discrimination | Distance weighted discrimination (DWD) is a linear discrimination method that
is particularly well-suited for classification tasks with high-dimensional
data. The DWD coefficients minimize an intuitive objective function, which can
solved very efficiently using state-of-the-art optimization techniques.
However, DWD has not yet been cast into a model-based framework for statistical
inference. In this article we show that DWD identifies the mode of a proper
Bayesian posterior distribution, that results from a particular link function
for the class probabilities and a shrinkage-inducing proper prior distribution
on the coefficients. We describe a relatively efficient Markov chain Monte
Carlo (MCMC) algorithm to simulate from the true posterior under this Bayesian
framework. We show that the posterior is asymptotically normal and derive the
mean and covariance matrix of its limiting distribution. Through several
simulation studies and an application to breast cancer genomics we demonstrate
how the Bayesian approach to DWD can be used to (1) compute well-calibrated
posterior class probabilities, (2) assess uncertainty in the DWD coefficients
and resulting sample scores, (3) improve power via semi-supervised analysis
when not all class labels are available, and (4) automatically determine a
penalty tuning parameter within the model-based framework. R code to perform
Bayesian DWD is available at https://github.com/lockEF/BayesianDWD .
| stat |
Assessing the effects of exposure to sulfuric acid aerosol on
respiratory function in adults | Sulfuric acid aerosol is suspected to be a major contributor to mortality and
morbidity associated with air pollution. The objective of the study is to
determine if exposure of human participants to anticipated levels of sulfuric
acid aerosol ($\sim 100\mu g/m^3 $) in the near future would have an adverse
effect on respiratory function. We used data from 28 adults exposed to sulfuric
acid for 4 hours in a controlled exposure chamber over a 3 day period with
repeated measures of pulmonary function (FEV1) recorded at 2-hour intervals.
Measurements were also recorded after 2 and 24 hours post exposure. We
formulated a linear mixed effect model for FEV1 with fixed effects (day of
treatment, hour, day-hour interaction, and smoking status), a random intercept
and an AR1 covariance structure to estimate the effect of aerosol exposure on
FEV1. We further assessed whether smoking status modified the exposure effects
and compared the analysis to the method used by Kerr et al.,1981. The findings
of the study show that the effect of day 3 exposure is negatively associated
with lung function (coefficient ($\beta$), -0.08; 95% CI, -0.16 to -0.01). A
weak negative association is also observed with increasing hours of exposure
($\beta$, -0.01; 95% CI, -0.03 to 0.00). Among the smokers, we found a
significant negative association with hours of exposure ($\beta$, -0.02; 95%
CI, -0.03 to -0.00), day 3 exposure ($\beta$, -0.11; 95% CI, -0.14 to -0.02)
and a borderline adverse effect for day 2 treatment ($\beta$, -0.06; 95% CI,
-0.14 to 0.03) whilst no significant association was observed for nonsmokers.
In conclusion, anticipated deposits of sulfuric acid aerosol in the near would
adversely affect respiratory function. The effect observed in smokers is
significantly more adverse than in nonsmokers.
| stat |
Learning Direct and Inverse Transmission Matrices | Linear problems appear in a variety of disciplines and their application for
the transmission matrix recovery is one of the most stimulating challenges in
biomedical imaging. Its knowledge turns any random media into an optical tool
that can focus or transmit an image through disorder. Here, converting an
input-output problem into a statistical mechanical formulation, we investigate
how inference protocols can learn the transmission couplings by
pseudolikelihood maximization. Bridging linear regression and thermodynamics
let us propose an innovative framework to pursue the solution of the
scattering-riddle.
| stat |
Exploratory Mediation Analysis with Many Potential Mediators | Social and behavioral scientists are increasingly employing technologies such
as fMRI, smartphones, and gene sequencing, which yield 'high-dimensional'
datasets with more columns than rows. There is increasing interest, but little
substantive theory, in the role the variables in these data play in known
processes. This necessitates exploratory mediation analysis, for which
structural equation modeling is the benchmark method. However, this method
cannot perform mediation analysis with more variables than observations. One
option is to run a series of univariate mediation models, which incorrectly
assumes independence of the mediators. Another option is regularization, but
the available implementations may lead to high false positive rates. In this
paper, we develop a hybrid approach which uses components of both filter and
regularization: the 'Coordinate-wise Mediation Filter'. It performs filtering
conditional on the other selected mediators. We show through simulation that it
improves performance over existing methods. Finally, we provide an empirical
example, showing how our method may be used for epigenetic research.
| stat |
Bayesian model selection approach for colored graphical Gaussian models | We consider a class of colored graphical Gaussian models obtained by placing
symmetry constraints on the precision matrix in a Bayesian framework. The prior
distribution on the precision matrix is the colored $G$-Wishart prior which is
the Diaconis-Ylvisaker conjugate prior. In this paper, we develop a
computationally efficient model search algorithm which combines linear
regression with a double reversible jump Markov chain Monte Carlo (MCMC)
method. The latter is to estimate the Bayes factors expressed as the ratio of
posterior probabilities of two competing models. We also establish the
asymptotic consistency property of the model selection procedure based on the
Bayes factors. Our procedure avoids an exhaustive search which is
computationally impossible. Our method is illustrated with simulations and a
real-world application with a protein signalling data set.
| stat |
A Manifold Proximal Linear Method for Sparse Spectral Clustering with
Application to Single-Cell RNA Sequencing Data Analysis | Spectral clustering is one of the fundamental unsupervised learning methods
widely used in data analysis. Sparse spectral clustering (SSC) imposes sparsity
to the spectral clustering and it improves the interpretability of the model.
This paper considers a widely adopted model for SSC, which can be formulated as
an optimization problem over the Stiefel manifold with nonsmooth and nonconvex
objective. Such an optimization problem is very challenging to solve. Existing
methods usually solve its convex relaxation or need to smooth its nonsmooth
part using certain smoothing techniques. In this paper, we propose a manifold
proximal linear method (ManPL) that solves the original SSC formulation. We
also extend the algorithm to solve the multiple-kernel SSC problems, for which
an alternating ManPL algorithm is proposed. Convergence and iteration
complexity results of the proposed methods are established. We demonstrate the
advantage of our proposed methods over existing methods via the single-cell RNA
sequencing data analysis.
| stat |
Combining Generative and Discriminative Models for Hybrid Inference | A graphical model is a structured representation of the data generating
process. The traditional method to reason over random variables is to perform
inference in this graphical model. However, in many cases the generating
process is only a poor approximation of the much more complex true data
generating process, leading to suboptimal estimation. The subtleties of the
generative process are however captured in the data itself and we can `learn to
infer', that is, learn a direct mapping from observations to explanatory latent
variables. In this work we propose a hybrid model that combines graphical
inference with a learned inverse model, which we structure as in a graph neural
network, while the iterative algorithm as a whole is formulated as a recurrent
neural network. By using cross-validation we can automatically balance the
amount of work performed by graphical inference versus learned inference. We
apply our ideas to the Kalman filter, a Gaussian hidden Markov model for time
sequences, and show, among other things, that our model can estimate the
trajectory of a noisy chaotic Lorenz Attractor much more accurately than either
the learned or graphical inference run in isolation.
| stat |
Stochastic Generalized Lotka-Volterra Model with An Application to
Learning Microbial Community Structures | Inferring microbial community structure based on temporal metagenomics data
is an important goal in microbiome studies. The deterministic generalized
Lotka-Volterra differential (GLV) equations have been used to model the
dynamics of microbial data. However, these approaches fail to take random
environmental fluctuations into account, which may negatively impact the
estimates. We propose a new stochastic GLV (SGLV) differential equation model,
where the random perturbations of Brownian motion in the model can naturally
account for the external environmental effects on the microbial community. We
establish new conditions and show various mathematical properties of the
solutions including general existence and uniqueness, stationary distribution,
and ergodicity. We further develop approximate maximum likelihood estimators
based on discrete observations and systematically investigate the consistency
and asymptotic normality of the proposed estimators. Our method is demonstrated
through simulation studies and an application to the well-known "moving
picture" temporal microbial dataset.
| stat |
Gradient Boosting to Boost the Efficiency of Hydraulic Fracturing | In this paper, we present a data-driven model for forecasting the production
increase after hydraulic fracturing (HF). We use data from fracturing jobs
performed at one of the Siberian oilfields. The data includes features,
characterizing the jobs, and geological information. To predict an oil rate
after the fracturing machine learning (ML) technique was applied. We compared
the ML-based prediction to a prediction based on the experience of reservoir
and production engineers responsible for the HF-job planning. We discuss the
potential for further development of ML techniques for predicting changes in
oil rate after HF.
| stat |
Time series model selection with a meta-learning approach; evidence from
a pool of forecasting algorithms | One of the challenging questions in time series forecasting is how to find
the best algorithm. In recent years, a recommender system scheme has been
developed for time series analysis using a meta-learning approach. This system
selects the best forecasting method with consideration of the time series
characteristics. In this paper, we propose a novel approach to focusing on some
of the unanswered questions resulting from the use of meta-learning in time
series forecasting. Therefore, three main gaps in previous works are addressed
including, analyzing various subsets of top forecasters as inputs for
meta-learners; evaluating the effect of forecasting error measures; and
assessing the role of the dimensionality of the feature space on the
forecasting errors of meta-learners. All of these objectives are achieved with
the help of a diverse state-of-the-art pool of forecasters and meta-learners.
For this purpose, first, a pool of forecasting algorithms is implemented on the
NN5 competition dataset and ranked based on the two error measures. Then, six
machine-learning classifiers known as meta-learners, are trained on the
extracted features of the time series in order to assign the most suitable
forecasting method for the various subsets of the pool of forecasters.
Furthermore, two-dimensionality reduction methods are implemented in order to
investigate the role of feature space dimension on the performance of
meta-learners. In general, it was found that meta-learners were able to defeat
all of the individual benchmark forecasters; this performance was improved even
after applying the feature selection method.
| stat |
Estimating adult death rates from sibling histories: A network approach | Hundreds of millions of people live in countries that do not have complete
death registration systems, meaning that most deaths are not recorded and
critical quantities like life expectancy cannot be directly measured. The
sibling survival method is a leading approach to estimating adult mortality in
the absence of death registration. The idea is to ask a survey respondent to
enumerate her siblings and to report about their survival status. In many
countries and time periods, sibling survival data are the only
nationally-representative source of information about adult mortality. Although
a huge amount of sibling survival data has been collected, important
methodological questions about the method remain unresolved. To help make
progress on this issue, we propose re-framing the sibling survival method as a
network sampling problem. This approach enables us to formally derive
statistical estimators for sibling survival data. Our derivation clarifies the
precise conditions that sibling history estimates rely upon; it leads to
internal consistency checks that can help assess data and reporting quality;
and it reveals important quantities that could potentially be measured to relax
assumptions in the future. We introduce the R package siblingsurvival, which
implements the methods we describe.
| stat |
Learning Bounds for Risk-sensitive Learning | In risk-sensitive learning, one aims to find a hypothesis that minimizes a
risk-averse (or risk-seeking) measure of loss, instead of the standard expected
loss. In this paper, we propose to study the generalization properties of
risk-sensitive learning schemes whose optimand is described via optimized
certainty equivalents (OCE): our general scheme can handle various known risks,
e.g., the entropic risk, mean-variance, and conditional value-at-risk, as
special cases. We provide two learning bounds on the performance of empirical
OCE minimizer. The first result gives an OCE guarantee based on the Rademacher
average of the hypothesis space, which generalizes and improves existing
results on the expected loss and the conditional value-at-risk. The second
result, based on a novel variance-based characterization of OCE, gives an
expected loss guarantee with a suppressed dependence on the smoothness of the
selected OCE. Finally, we demonstrate the practical implications of the
proposed bounds via exploratory experiments on neural networks.
| stat |
Conditional Versus Adversarial Euler-based Generators For Time Series | We introduce new generative models for time series based on Euler
discretization that do not require any pre-stationarization procedure.
Specifically, we develop two GAN based methods, relying on the adaptation of
Wasserstein GANs (Arjovsky et al., 2017) and DVD GANs (Clark et al., 2019b) to
time series. Alternatively, we consider a conditional Euler Generator (CEGEN)
minimizing a distance between the induced conditional densities. In the context
of It\^o processes, we theoretically validate this approach and demonstrate
using the Bures metric that reaching a low loss level provides accurate
estimations for both the drift and the volatility terms of the underlying
process. Tests on simple models show how the Euler discretization and the use
of Wasserstein distance allow the proposed GANs and (more considerably) CEGEN
to outperform state-of-the-art Time Series GAN generation( Yoon et al., 2019b)
on time structure metrics. In higher dimensions we observe that CEGEN manages
to get the correct covariance structures. Finally we illustrate how our model
can be combined to a Monte Carlo simulator in a low data context by using a
transfer learning technique
| stat |
Independent Subspace Analysis for Unsupervised Learning of Disentangled
Representations | Recently there has been an increased interest in unsupervised learning of
disentangled representations using the Variational Autoencoder (VAE) framework.
Most of the existing work has focused largely on modifying the variational cost
function to achieve this goal. We first show that these modifications, e.g.
beta-VAE, simplify the tendency of variational inference to underfit causing
pathological over-pruning and over-orthogonalization of learned components.
Second we propose a complementary approach: to modify the probabilistic model
with a structured latent prior. This prior allows to discover latent variable
representations that are structured into a hierarchy of independent vector
spaces. The proposed prior has three major advantages: First, in contrast to
the standard VAE normal prior the proposed prior is not rotationally invariant.
This resolves the problem of unidentifiability of the standard VAE normal
prior. Second, we demonstrate that the proposed prior encourages a disentangled
latent representation which facilitates learning of disentangled
representations. Third, extensive quantitative experiments demonstrate that the
prior significantly mitigates the trade-off between reconstruction loss and
disentanglement over the state of the art.
| stat |
Sharp instruments for classifying compliers and generalizing causal
effects | It is well-known that, without restricting treatment effect heterogeneity,
instrumental variable (IV) methods only identify "local" effects among
compliers, i.e., those subjects who take treatment only when encouraged by the
IV. Local effects are controversial since they seem to only apply to an
unidentified subgroup; this has led many to denounce these effects as having
little policy relevance. However, we show that such pessimism is not always
warranted: it is possible in some cases to accurately predict who compliers
are, and obtain tight bounds on more generalizable effects in identifiable
subgroups. We propose methods for doing so and study their estimation error and
asymptotic properties, showing that these tasks can in theory be accomplished
even with very weak IVs. We go on to introduce a new measure of IV quality
called "sharpness", which reflects the variation in compliance explained by
covariates, and captures how well one can identify compliers and obtain tight
bounds on identifiable subgroup effects. We develop an estimator of sharpness,
and show that it is asymptotically efficient under weak conditions. Finally we
explore finite-sample properties via simulation, and apply the methods to study
canvassing effects on voter turnout. We propose that sharpness should be
presented alongside strength to assess IV quality.
| stat |
Median bias reduction in cumulative link models | This paper presents a novel estimation approach for cumulative link models,
based on median bias reduction as developed in Kenne Pagui et al. (2017). The
median bias reduced estimator is obtained as solution of an estimating equation
based on an adjustment of the score. It allows to obtain higher-order median
centering of maximum likelihood estimates without requiring their finiteness.
Moreover, the estimator is equivariant under componentwise monotone
reparameterizations and the method is effective in preventing boundary
estimates. We evaluate the properties of the median bias reduced estimator
through simulation studies and compare it with the two main competitors, the
maximum likelihood and the mean bias reduced (Firth, 1993) estimators. Finally,
we show an application where the proposed estimator is able to solve the
boundary estimates problem.
| stat |
A Note on Estimating Optimal Dynamic Treatment Strategies Under Resource
Constraints Using Dynamic Marginal Structural Models | Existing strategies for determining the optimal treatment or monitoring
strategy typically assume unlimited access to resources. However, when a health
system has resource constraints, such as limited funds, access to medication,
or monitoring capabilities, medical decisions must balance impacts on both
individual and population health outcomes. That is, decisions should account
for competition between individuals in resource usage. One simple solution is
to estimate the (counterfactual) resource usage under the possible
interventions and choose the optimal strategy for which resource usage is
within acceptable limits. We propose a method to identify the optimal dynamic
intervention strategy that leads to the best expected health outcome accounting
for a health system's resource constraints. We then apply this method to
determine the optimal dynamic monitoring strategy for people living with HIV
when resource limits on monitoring exist using observational data from the
HIV-CAUSAL Collaboration.
| stat |
Variational approximations of empirical Bayes posteriors in
high-dimensional linear models | In high-dimensions, the prior tails can have a significant effect on both
posterior computation and asymptotic concentration rates. To achieve optimal
rates while keeping the posterior computations relatively simple, an empirical
Bayes approach has recently been proposed, featuring thin-tailed conjugate
priors with data-driven centers. While conjugate priors ease some of the
computational burden, Markov chain Monte Carlo methods are still needed, which
can be expensive when dimension is high. In this paper, we develop a
variational approximation to the empirical Bayes posterior that is fast to
compute and retains the optimal concentration rate properties of the original.
In simulations, our method is shown to have superior performance compared to
existing variational approximations in the literature across a wide range of
high-dimensional settings.
| stat |
Benign Overfitting in Linear Regression | The phenomenon of benign overfitting is one of the key mysteries uncovered by
deep learning methodology: deep neural networks seem to predict well, even with
a perfect fit to noisy training data. Motivated by this phenomenon, we consider
when a perfect fit to training data in linear regression is compatible with
accurate prediction. We give a characterization of linear regression problems
for which the minimum norm interpolating prediction rule has near-optimal
prediction accuracy. The characterization is in terms of two notions of the
effective rank of the data covariance. It shows that overparameterization is
essential for benign overfitting in this setting: the number of directions in
parameter space that are unimportant for prediction must significantly exceed
the sample size. By studying examples of data covariance properties that this
characterization shows are required for benign overfitting, we find an
important role for finite-dimensional data: the accuracy of the minimum norm
interpolating prediction rule approaches the best possible accuracy for a much
narrower range of properties of the data distribution when the data lies in an
infinite dimensional space versus when the data lies in a finite dimensional
space whose dimension grows faster than the sample size.
| stat |
Generalization error in high-dimensional perceptrons: Approaching Bayes
error with convex optimization | We consider a commonly studied supervised classification of a synthetic
dataset whose labels are generated by feeding a one-layer neural network with
random iid inputs. We study the generalization performances of standard
classifiers in the high-dimensional regime where $\alpha=n/d$ is kept finite in
the limit of a high dimension $d$ and number of samples $n$. Our contribution
is three-fold: First, we prove a formula for the generalization error achieved
by $\ell_2$ regularized classifiers that minimize a convex loss. This formula
was first obtained by the heuristic replica method of statistical physics.
Secondly, focussing on commonly used loss functions and optimizing the $\ell_2$
regularization strength, we observe that while ridge regression performance is
poor, logistic and hinge regression are surprisingly able to approach the
Bayes-optimal generalization error extremely closely. As $\alpha \to \infty$
they lead to Bayes-optimal rates, a fact that does not follow from predictions
of margin-based generalization error bounds. Third, we design an optimal loss
and regularizer that provably leads to Bayes-optimal generalization error.
| stat |
A Multi-Scale Tensor Network Architecture for Classification and
Regression | We present an algorithm for supervised learning using tensor networks,
employing a step of preprocessing the data by coarse-graining through a
sequence of wavelet transformations. We represent these transformations as a
set of tensor network layers identical to those in a multi-scale entanglement
renormalization ansatz (MERA) tensor network, and perform supervised learning
and regression tasks through a model based on a matrix product state (MPS)
tensor network acting on the coarse-grained data. Because the entire model
consists of tensor contractions (apart from the initial non-linear feature
map), we can adaptively fine-grain the optimized MPS model backwards through
the layers with essentially no loss in performance. The MPS itself is trained
using an adaptive algorithm based on the density matrix renormalization group
(DMRG) algorithm. We test our methods by performing a classification task on
audio data and a regression task on temperature time-series data, studying the
dependence of training accuracy on the number of coarse-graining layers and
showing how fine-graining through the network may be used to initialize models
with access to finer-scale features.
| stat |
Breakdown points of penalized and hybrid M-estimators of covariance | We introduce a class of hybrid M-estimators of multivariate scatter which,
analogous to the popular spatial sign covariance matrix (SSCM), possess high
breakdown points. We also show that the SSCM can be viewed as an extreme member
of this class. Unlike the SSCM, but like the regular M-estimators of scatter,
this new class of estimators takes into account the shape of the contours of
the data cloud for downweighting observations.
| stat |
Estimation of the Parameters of Symmetric Stable ARMA and ARMA-GARCH
Models | In this article, we first propose the modified Hannan-Rissanen Method for
estimating the parameters of the autoregressive moving average (ARMA) process
with symmetric stable noise and symmetric stable generalized autoregressive
conditional heteroskedastic (GARCH) noise. Next, we propose the modified
empirical characteristic function method for the estimation of GARCH parameters
with symmetric stable noise. Further, we show the efficiency, accuracy, and
simplicity of our methods through Monte-Carlo simulation. Finally, we apply our
proposed methods to model financial data.
| stat |
Image fusion using symmetric skip autoencodervia an Adversarial
Regulariser | It is a challenging task to extract the best of both worlds by combining the
spatial characteristics of a visible image and the spectral content of an
infrared image. In this work, we propose a spatially constrained adversarial
autoencoder that extracts deep features from the infrared and visible images to
obtain a more exhaustive and global representation. In this paper, we propose a
residual autoencoder architecture, regularised by a residual adversarial
network, to generate a more realistic fused image. The residual module serves
as primary building for the encoder, decoder and adversarial network, as an add
on the symmetric skip connections perform the functionality of embedding the
spatial characteristics directly from the initial layers of encoder structure
to the decoder part of the network. The spectral information in the infrared
image is incorporated by adding the feature maps over several layers in the
encoder part of the fusion structure, which makes inference on both the visual
and infrared images separately. In order to efficiently optimize the parameters
of the network, we propose an adversarial regulariser network which would
perform supervised learning on the fused image and the original visual image.
| stat |
Horseshoe-based Bayesian nonparametric estimation of effective
population size trajectories | Phylodynamics is an area of population genetics that uses genetic sequence
data to estimate past population dynamics. Modern state-of-the-art Bayesian
nonparametric methods for recovering population size trajectories of unknown
form use either change-point models or Gaussian process priors. Change-point
models suffer from computational issues when the number of change-points is
unknown and needs to be estimated. Gaussian process-based methods lack local
adaptivity and cannot accurately recover trajectories that exhibit features
such as abrupt changes in trend or varying levels of smoothness. We propose a
novel, locally-adaptive approach to Bayesian nonparametric phylodynamic
inference that has the flexibility to accommodate a large class of functional
behaviors. Local adaptivity results from modeling the log-transformed effective
population size a priori as a horseshoe Markov random field, a recently
proposed statistical model that blends together the best properties of the
change-point and Gaussian process modeling paradigms. We use simulated data to
assess model performance, and find that our proposed method results in reduced
bias and increased precision when compared to contemporary methods. We also use
our models to reconstruct past changes in genetic diversity of human hepatitis
C virus in Egypt and to estimate population size changes of ancient and modern
steppe bison. These analyses show that our new method captures features of the
population size trajectories that were missed by the state-of-the-art methods.
| stat |
Common Growth Patterns for Regional Social Networks: a Point Process
Approach | Although recent research on social networks emphasizes microscopic dynamics
such as retweets and social connectivity of an individual user, we focus on
macroscopic growth dynamics of social network link formation. Rather than
focusing on one particular dataset, we find invariant behavior in regional
social networks that are geographically concentrated. Empirical findings
suggest that the startup phase of a regional network can be modeled by a
self-exciting point process. After the startup phase ends, the growth of the
links can be modeled by a non-homogeneous Poisson process with constant rate
across the day but varying rates from day to day, plus a nightly inactive
period when local users are expected to be asleep. Conclusions are drawn based
on analyzing four different datasets, three of which are regional and a
non-regional one is included for contrast.
| stat |
ICU Delirium Prediction Models: A Systematic Review | Purpose: Summarize ICU delirium prediction models published within the past
five years. Methods: Electronic searches were conducted in April 2019 using
PubMed, Embase, Cochrane Central, Web of Science, and CINAHL to identify peer
reviewed studies published in English during the past five years that
specifically addressed the development, validation, or recalibration of
delirium prediction models in adult ICU populations. Data were extracted using
CHARMS checklist elements for systematic reviews of prediction studies,
including the following characteristics: study design, participant descriptions
and recruitment methods, predicted outcomes, a priori candidate predictors,
sample size, model development, model performance, study results,
interpretation of those results, and whether the study included missing data.
Results: Twenty studies featuring 26 distinct prediction models were included.
Model performance varied greatly, as assessed by AUROC (0.68-0.94), specificity
(56.5%-92.5%), and sensitivity (59%-90.9%). Most models used data collected
from a single time point or window to predict the occurrence of delirium at any
point during hospital or ICU admission, and lacked mechanisms for providing
pragmatic, actionable predictions to clinicians. Conclusions: Although most ICU
delirium prediction models have relatively good performance, they have limited
applicability to clinical practice. Most models were static, making predictions
based on data collected at a single time-point, failing to account for
fluctuating conditions during ICU admission. Further research is needed to
create clinically relevant dynamic delirium prediction models that can adapt to
changes in individual patient physiology over time and deliver actionable
predictions to clinicians.
| stat |
Functional time series prediction under partial observation of the
future curve | This paper tackles one of the most fundamental goals in functional time
series analysis which is to provide reliable predictions for functions.
Existing functional time series methods seek to predict a complete future
functional observation based on a set of observed complete trajectories. The
problem of interest discussed here is how to advance prediction methodology to
cases where partial information on the next trajectory is available, with the
aim of improving prediction accuracy. To solve this problem, we propose a new
method "partial functional prediction (PFP)". The proposed method combines
"next-interval" prediction and fully functional regression prediction, so that
the partially observed part of the trajectory can aid in producing a better
prediction for the unobserved part of the future curve. In PFP, we include
automatic selection criterion for tuning parameters based on minimizing the
prediction error. Simulations indicate that the proposed method can outperform
existing methods with respect to mean-square prediction error and its practical
utility is illustrated in an analysis of environmental and traffic flow data.
| stat |
Early Anomaly Detection and Location in Distribution Network: A
Data-Driven Approach | The measurement data collected from the supervisory control and data
acquisition (SCADA) system installed in distribution network can reflect the
operational state of the network effectively. In this paper, a random matrix
theory (RMT) based approach is developed for early anomaly detection and
localization by using the data. For every feeder in the distribution network, a
corresponding data matrix is formed. Based on the Marchenko-Pastur Law for the
empirical spectral analysis of covariance `signal+noise' matrix, the linear
eigenvalue statistics are introduced to indicate the anomaly, and the outliers
and their corresponding eigenvectors are analyzed for locating the anomaly. As
for the low observability feeders in the distribution network, an increasing
data dimension algorithm is designed for the formulated low-dimensional
matrices being more accurately analyzed. The developed approach can detect and
localize the anomaly at an early stage, and it is robust to random disturbance
and measurement error. Cases on Matpower simulation data and real SCADA data
corroborate the feasibility of the approach.
| stat |
Modeling of Missing Dynamical Systems: Deriving Parametric Models using
a Nonparametric Framework | In this paper, we consider modeling missing dynamics with a nonparametric
non-Markovian model, constructed using the theory of kernel embedding of
conditional distributions on appropriate Reproducing Kernel Hilbert Spaces
(RKHS), equipped with orthonormal basis functions. Depending on the choice of
the basis functions, the resulting closure model from this nonparametric
modeling formulation is in the form of parametric model. This suggests that the
success of various parametric modeling approaches that were proposed in various
domains of applications can be understood through the RKHS representations.
When the missing dynamical terms evolve faster than the relevant observable of
interest, the proposed approach is consistent with the effective dynamics
derived from the classical averaging theory. In the linear Gaussian case
without the time-scale gap, we will show that the proposed non-Markovian model
with a very long memory yields an accurate estimation of the nontrivial
autocovariance function for the relevant variable of the full dynamics.
Supporting numerical results on instructive nonlinear dynamics show that the
proposed approach is able to replicate high-dimensional missing dynamical terms
on problems with and without the separation of temporal scales.
| stat |
Disentangled Sticky Hierarchical Dirichlet Process Hidden Markov Model | The Hierarchical Dirichlet Process Hidden Markov Model (HDP-HMM) has been
used widely as a natural Bayesian nonparametric extension of the classical
Hidden Markov Model for learning from sequential and time-series data. A sticky
extension of the HDP-HMM has been proposed to strengthen the self-persistence
probability in the HDP-HMM. However, the sticky HDP-HMM entangles the strength
of the self-persistence prior and transition prior together, limiting its
expressiveness. Here, we propose a more general model: the disentangled sticky
HDP-HMM (DS-HDP-HMM). We develop novel Gibbs sampling algorithms for efficient
inference in this model. We show that the disentangled sticky HDP-HMM
outperforms the sticky HDP-HMM and HDP-HMM on both synthetic and real data, and
apply the new approach to analyze neural data and segment behavioral video
data.
| stat |
Differentiating the multipoint Expected Improvement for optimal batch
design | This work deals with parallel optimization of expensive objective functions
which are modeled as sample realizations of Gaussian processes. The study is
formalized as a Bayesian optimization problem, or continuous multi-armed bandit
problem, where a batch of q > 0 arms is pulled in parallel at each iteration.
Several algorithms have been developed for choosing batches by trading off
exploitation and exploration. As of today, the maximum Expected Improvement
(EI) and Upper Confidence Bound (UCB) selection rules appear as the most
prominent approaches for batch selection. Here, we build upon recent work on
the multipoint Expected Improvement criterion, for which an analytic expansion
relying on Tallis' formula was recently established. The computational burden
of this selection rule being still an issue in application, we derive a
closed-form expression for the gradient of the multipoint Expected Improvement,
which aims at facilitating its maximization using gradient-based ascent
algorithms. Substantial computational savings are shown in application. In
addition, our algorithms are tested numerically and compared to
state-of-the-art UCB-based batch-sequential algorithms. Combining starting
designs relying on UCB with gradient-based EI local optimization finally
appears as a sound option for batch design in distributed Gaussian Process
optimization.
| stat |
Bayesian Methods for the Analysis of Early-Phase Oncology Basket Trials
with Information Borrowing across Cancer Types | Research in oncology has changed the focus from histological properties of
tumors in a specific organ to a specific genomic aberration potentially shared
by multiple cancer types. This motivates the basket trial, which assesses the
efficacy of treatment simultaneously on multiple cancer types that have a
common aberration. Although the assumption of homogeneous treatment effects
seems reasonable given the shared aberration, in reality, the treatment effect
may vary by cancer type, and potentially only a subgroup of the cancer types
respond to the treatment. Various approaches have been proposed to increase the
trial power by borrowing information across cancer types, which, however, tend
to inflate the type I error rate. In this paper, we review some representative
Bayesian information borrowing methods for the analysis of early-phase basket
trials. We then propose a novel method called the Bayesian hierarchical model
with a correlated prior (CBHM), which conducts more flexible borrowing across
cancer types according to sample similarity. We did simulation studies to
compare CBHM with independent analysis and three information borrowing
approaches: the conventional Bayesian hierarchical model, the EXNEX approach
and Liu's two-stage approach. Simulation results show that all information
borrowing approaches substantially improve the power of independent analysis if
a large proportion of the cancer types truly respond to the treatment. Our
proposed CBHM approach shows an advantage over the existing information
borrowing approaches, with a power similar to that of EXNEX or Liu's approach,
but the potential to provide substantially better control of type I error rate.
| stat |
Inference in High-Dimensional Linear Measurement Error Models | For a high-dimensional linear model with a finite number of covariates
measured with error, we study statistical inference on the parameters
associated with the error-prone covariates, and propose a new corrected
decorrelated score test and the corresponding one-step estimator. We further
establish asymptotic properties of the newly proposed test statistic and the
one-step estimator. Under local alternatives, we show that the limiting
distribution of our corrected decorrelated score test statistic is non-central
normal. The finite-sample performance of the proposed inference procedure is
examined through simulation studies. We further illustrate the proposed
procedure via an empirical analysis of a real data example.
| stat |
Bayesian Surrogate Analysis and Uncertainty Propagation with Explicit
Surrogate Uncertainties and Implicit Spatio-temporal Correlations | We introduce Bayesian Probability Theory to investigate uncertainty
propagation based on meta-models. We approach the problem from the perspective
of data analysis, with a given (however almost-arbitrary) input probability
distribution and a given "training" set of computer simulations. While proven
mathematically to be the unique consistent probability calculus, the subject of
this paper is not to demonstrate beauty but usefulness. We explicitly list all
propositions and lay open the general structure of any uncertainty propagation
based on meta-models. The former allows rigorous treatment at any stage, while
the latter allows us to quantify the interaction of the surrogate uncertainties
with the usual parameter uncertainties. Additionally, we show a simple way to
implicitly include spatio-temporal correlations. We then apply the framework
jointly to a family of generalized linear meta-model that implicitly includes
Polynomial Chaos Expansions as a special case. While we assume a Gaussian
surrogate-uncertainty, we do not assume a scale for the surrogate uncertainty
to be known, i.e. a Student-t. We end up with semi-analytic formulas for
surrogate uncertainties and uncertainty propagation
| stat |
Compressed Dictionary Learning | In this paper we show that the computational complexity of the Iterative
Thresholding and K-residual-Means (ITKrM) algorithm for dictionary learning can
be significantly reduced by using dimensionality-reduction techniques based on
the Johnson-Lindenstrauss lemma. The dimensionality reduction is efficiently
carried out with the fast Fourier transform. We introduce the Iterative
compressed-Thresholding and K-Means (IcTKM) algorithm for fast dictionary
learning and study its convergence properties. We show that IcTKM can locally
recover an incoherent, overcomplete generating dictionary of $K$ atoms from
training signals of sparsity level $S$ with high probability. Fast dictionary
learning is achieved by embedding the training data and the dictionary into $m
< d$ dimensions, and recovery is shown to be locally stable with an embedding
dimension which scales as low as $m = O(S \log^4 S \log^3 K)$. The compression
effectively shatters the data dimension bottleneck in the computational cost of
ITKrM, reducing it by a factor $O(m/d)$. Our theoretical results are
complemented with numerical simulations which demonstrate that IcTKM is a
powerful, low-cost algorithm for learning dictionaries from high-dimensional
data sets.
| stat |
Spectral Top-Down Recovery of Latent Tree Models | Modeling the distribution of high dimensional data by a latent tree graphical
model is a common approach in multiple scientific domains. A common task is to
infer the underlying tree structure given only observations of the terminal
nodes. Many algorithms for tree recovery are computationally intensive, which
limits their applicability to trees of moderate size. For large trees, a common
approach, termed divide-and-conquer, is to recover the tree structure in two
steps. First, recover the structure separately for multiple randomly selected
subsets of the terminal nodes. Second, merge the resulting subtrees to form a
full tree. Here, we develop Spectral Top-Down Recovery (STDR), a
divide-and-conquer approach for inference of large latent tree models. Unlike
previous methods, STDR's partitioning step is non-random. Instead, it is based
on the Fiedler vector of a suitable Laplacian matrix related to the observed
nodes. We prove that under certain conditions this partitioning is consistent
with the tree structure. This, in turn leads to a significantly simpler merging
procedure of the small subtrees. We prove that STDR is statistically
consistent, and bound the number of samples required to accurately recover the
tree with high probability. Using simulated data from several common tree
models in phylogenetics, we demonstrate that STDR has a significant advantage
in terms of runtime, with improved or similar accuracy.
| stat |
New Computational and Statistical Aspects of Regularized Regression with
Application to Rare Feature Selection and Aggregation | Prior knowledge on properties of a target model often come as discrete or
combinatorial descriptions. This work provides a unified computational
framework for defining norms that promote such structures. More specifically,
we develop associated tools for optimization involving such norms given only
the orthogonal projection oracle onto the non-convex set of desired models. As
an example, we study a norm, which we term the doubly-sparse norm, for
promoting vectors with few nonzero entries taking only a few distinct values.
We further discuss how the K-means algorithm can serve as the underlying
projection oracle in this case and how it can be efficiently represented as a
quadratically constrained quadratic program. Our motivation for the study of
this norm is regularized regression in the presence of rare features which
poses a challenge to various methods within high-dimensional statistics, and in
machine learning in general. The proposed estimation procedure is designed to
perform automatic feature selection and aggregation for which we develop
statistical bounds. The bounds are general and offer a statistical framework
for norm-based regularization. The bounds rely on novel geometric quantities on
which we attempt to elaborate as well.
| stat |
Statistical Inference for the Population Landscape via Moment Adjusted
Stochastic Gradients | Modern statistical inference tasks often require iterative optimization
methods to compute the solution. Convergence analysis from an optimization
viewpoint only informs us how well the solution is approximated numerically but
overlooks the sampling nature of the data. In contrast, recognizing the
randomness in the data, statisticians are keen to provide uncertainty
quantification, or confidence, for the solution obtained using iterative
optimization methods. This paper makes progress along this direction by
introducing the moment-adjusted stochastic gradient descents, a new stochastic
optimization method for statistical inference. We establish non-asymptotic
theory that characterizes the statistical distribution for certain iterative
methods with optimization guarantees. On the statistical front, the theory
allows for model mis-specification, with very mild conditions on the data. For
optimization, the theory is flexible for both convex and non-convex cases.
Remarkably, the moment-adjusting idea motivated from "error standardization" in
statistics achieves a similar effect as acceleration in first-order
optimization methods used to fit generalized linear models. We also demonstrate
this acceleration effect in the non-convex setting through numerical
experiments.
| stat |
Distributionally Robust Reinforcement Learning | Real-world applications require RL algorithms to act safely. During learning
process, it is likely that the agent executes sub-optimal actions that may lead
to unsafe/poor states of the system. Exploration is particularly brittle in
high-dimensional state/action space due to increased number of low-performing
actions. In this work, we consider risk-averse exploration in approximate RL
setting. To ensure safety during learning, we propose the distributionally
robust policy iteration scheme that provides lower bound guarantee on
state-values. Our approach induces a dynamic level of risk to prevent poor
decisions and yet preserves the convergence to the optimal policy. Our
formulation results in a efficient algorithm that accounts for a simple
re-weighting of policy actions in the standard policy iteration scheme. We
extend our approach to continuous state/action space and present a practical
algorithm, distributionally robust soft actor-critic, that implements a
different exploration strategy: it acts conservatively at short-term and it
explores optimistically in a long-run. We provide promising experimental
results on continuous control tasks.
| stat |
A Bayesian hierarchical model to estimate land surface phenology
parameters with harmonized Landsat 8 and Sentinel-2 images | We develop a Bayesian Land Surface Phenology (LSP) model and examine its
performance using Enhanced Vegetation Index (EVI) observations derived from the
Harmonized Landsat Sentinel-2 (HLS) dataset. Building on previous work, we
propose a double logistic function that, once couched within a Bayesian model,
yields posterior distributions for all LSP parameters. We assess the efficacy
of the Normal, Truncated Normal, and Beta likelihoods to deliver robust LSP
parameter estimates. Two case studies are presented and used to explore aspects
of the proposed model. The first, conducted over forested pixels within a HLS
tile, explores choice of likelihood and space-time varying HLS data
availability for long-term average LSP parameter point and uncertainty
estimation. The second, conducted on a small area of interest within the HLS
tile on an annual time-step, further examines the impact of sample size and
choice of likelihood on LSP parameter estimates. Results indicate that while
the Truncated Normal and Beta likelihoods are theoretically preferable when the
vegetation index is bounded, all three likelihoods performed similarly when the
number of index observations is sufficiently large and values are not near the
index bounds. Both case studies demonstrate how pixel-level LSP parameter
posterior distributions can be used to propagate uncertainty through subsequent
analysis. As a companion to this article, we provide an open-source \R package
\pkg{rsBayes} and supplementary data and code used to reproduce the analysis
results. The proposed model specification and software implementation delivers
computationally efficient, statistically robust, and inferentially rich LSP
parameter posterior distributions at the pixel-level across massive raster time
series datasets.
| stat |
Network cross-validation by edge sampling | While many statistical models and methods are now available for network
analysis, resampling network data remains a challenging problem.
Cross-validation is a useful general tool for model selection and parameter
tuning, but is not directly applicable to networks since splitting network
nodes into groups requires deleting edges and destroys some of the network
structure. Here we propose a new network resampling strategy based on splitting
node pairs rather than nodes applicable to cross-validation for a wide range of
network model selection tasks. We provide a theoretical justification for our
method in a general setting and examples of how our method can be used in
specific network model selection and parameter tuning tasks. Numerical results
on simulated networks and on a citation network of statisticians show that this
cross-validation approach works well for model selection.
| stat |
Joint Mean-Vector and Var-Matrix estimation for Locally Stationary
VAR(1) processes | During the last two decades, locally stationary processes have been widely
studied in the time series literature. In this paper we consider the
locally-stationary vector-auto-regression model of order one, or LS-VAR(1), and
estimate its parameters by weighted least squares. The LS-VAR(1) we consider
allows for a smoothly time-varying non-diagonal VAR matrix, as well as for a
smoothly time-varying non-zero mean. The weighting scheme is based on kernel
smoothers. The time-varying mean and the time-varying VAR matrix are estimated
jointly, and the definition of the local-linear weighting matrix is provided in
closed-from. The quality of the estimated curves is illustrated through
simulation results.
| stat |
The Size of a $t$-Digest | A $t$-digest is a compact data structure that allows estimates of quantiles
which increased accuracy near $q = 0$ or $q=1$. This is done by clustering
samples from $\mathbb R$ subject to a constraint that the number of points
associated with any particular centroid is constrained so that the so-called
$k$-size of the centroid is always $\le 1$. The $k$-size is defined using a
scale function that maps quantile $q$ to index $k$. This paper provides bounds
on the sizes of $t$-digests created using any of four known scale functions.
| stat |
Bayesian hierarchical stacking | Stacking is a widely used model averaging technique that yields
asymptotically optimal prediction among all linear averages. We show that
stacking is most effective when the model predictive performance is
heterogeneous in inputs, so that we can further improve the stacked mixture
with a hierarchical model. With the input-varying yet partially-pooled model
weights, hierarchical stacking improves average and conditional predictions.
Our Bayesian formulation includes constant-weight (complete-pooling) stacking
as a special case. We generalize to incorporate discrete and continuous inputs,
other structured priors, and time-series and longitudinal data. We demonstrate
on several applied problems.
| stat |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.