title
stringlengths 6
244
| abstract
stringlengths 19
5.09k
| label
stringclasses 10
values |
|---|---|---|
Relationship Between Changing Malaria Burden and Low Birth Weight in
Sub-Saharan Africa | According to the World Health Organization (WHO), in 2018, an estimated 228
million malaria cases occurred worldwide with most cases occurring in
sub-Saharan Africa. Scale up of vector control tools coupled with increased
access to diagnosis and effective treatment has resulted in a large decline in
malaria prevalence in some areas, but other areas have seen little change.
Although interventional studies demonstrate that preventing malaria during
pregnancy can reduce the low birth weight (i.e., child's birth weight $<$ 2,500
grams) rate, it remains unknown whether natural changes in parasite
transmission and malaria burden can improve birth outcomes. In this work, we
conducted an observational study of the effect of changing malaria burden on
low birth weight using data from 18,112 births in 19 countries in sub-Saharan
African countries. A malaria prevalence decline from a high rate (Plasmodium
falciparum parasite rate in children aged 2-up-to-10 (i.e., $PfPR_{2-10}$)
$>$0.4) to a low rate ($PfPR_{2-10}$ $<$0.2) is estimated to reduce the rate of
low birth weight by 1.48 percentage point (95% confidence interval: 3.70
percentage point reduction, 0.74 percentage point increase), which is a 17%
reduction in the low birth weight rate compared to the average (8.6%) in our
study population with observed birth weight records (1.48/8.6 $\approx$ 17%).
When focusing on first pregnancies, a malaria prevalence decline from a high
rate to a low rate is estimated to reduce the low birth weight rate by more
than for all births, 3.73 percentage point (95% confidence interval: 9.11
percentage point reduction, 1.64 percentage point increase).
| stat |
Consistent Batch Normalization for Weighted Loss in Imbalanced-Data
Environment | In this study, classification problems based on feedforward neural networks
in a data-imbalanced environment are considered. Learning from an imbalanced
dataset is one of the most important practical problems in the field of machine
learning. A weighted loss function (WLF) based on a cost-sensitive approach is
a well-known and effective method for imbalanced datasets. A combination of WLF
and batch normalization (BN) is considered in this study. BN is considered as a
powerful standard technique in the recent developments in deep learning. A
simple combination of both methods leads to a size-inconsistency problem due to
a mismatch between the interpretations of the effective size of the dataset in
both methods. A simple modification to BN, called weighted BN (WBN), is
proposed to correct the size mismatch. The idea of WBN is simple and natural.
The proposed method in a data-imbalanced environment is validated using
numerical experiments.
| stat |
Analysis of the Formation of the Structure of Social Networks using
Latent Space Models for Ranked Dynamic Networks | The formation of social networks and the evolution of their structures have
been of interest to researchers for many decades. We wish to answer questions
about network stability, group formation and popularity effects. We propose a
latent space model for ranked dynamic networks that can be used to intuitively
frame and answer these questions. The well known data collected by Newcomb in
the 1950's is very well suited to analyze the formation of a social network. We
applied our model to this data in order to investigate the network stability,
what groupings emerge and when they emerge, and how individual popularity is
associated with individual stability.
| stat |
Effect of Interim Adaptations in Group Sequential Designs | This manuscript investigates unconditional and conditional-on-stopping
maximum likelihood estimators (MLEs), information measures and information loss
associated with conditioning in group sequential designs (GSDs). The
possibility of early stopping brings truncation to the distributional form of
MLEs; sequentially, GSD decisions eliminate some events from the sample space.
Multiple testing induces mixtures on the adapted sample space. Distributions of
MLEs are mixtures of truncated distributions. Test statistics that are
asymptotically normal without GSD, have asymptotic distributions, under GSD,
that are non-normal mixtures of truncated normal distributions under local
alternatives; under fixed alternatives, asymptotic distributions of test
statistics are degenerate. Estimation of various statistical quantities such as
information, information fractions, and confidence intervals should account for
the effect of planned adaptations. Calculation of adapted information fractions
requires substantial computational effort. Therefore, a new GSD is proposed in
which stage-specific sample sizes are fully determined by desired operational
characteristics, and calculation of information fractions is not needed.
| stat |
Adaptive Sampling to Reduce Disparate Performance | Existing methods for reducing disparate performance of a classifier across
different demographic groups assume that one has access to a large data set,
thereby focusing on the algorithmic aspect of optimizing overall performance
subject to additional constraints. However, poor data collection and imbalanced
data sets can severely affect the quality of these methods. In this work, we
consider a setting where data collection and optimization are performed
simultaneously. In such a scenario, a natural strategy to mitigate the
performance difference of the classifier is to provide additional training data
drawn from the demographic groups that are worse off. In this paper, we propose
to consistently follow this strategy throughout the whole training process and
to guide the resulting classifier towards equal performance on the different
groups by adaptively sampling each data point from the group that is currently
disadvantaged. We provide a rigorous theoretical analysis of our approach in a
simplified one-dimensional setting and an extensive experimental evaluation on
numerous real-world data sets, including a case study on the data collected
during the Flint water crisis.
| stat |
Considering discrepancy when calibrating a mechanistic electrophysiology
model | Uncertainty quantification (UQ) is a vital step in using mathematical models
and simulations to take decisions. The field of cardiac simulation has begun to
explore and adopt UQ methods to characterise uncertainty in model inputs and
how that propagates through to outputs or predictions. In this perspective
piece we draw attention to an important and under-addressed source of
uncertainty in our predictions -- that of uncertainty in the model structure or
the equations themselves. The difference between imperfect models and reality
is termed model discrepancy, and we are often uncertain as to the size and
consequences of this discrepancy. Here we provide two examples of the
consequences of discrepancy when calibrating models at the ion channel and
action potential scales. Furthermore, we attempt to account for this
discrepancy when calibrating and validating an ion channel model using
different methods, based on modelling the discrepancy using Gaussian processes
(GPs) and autoregressive-moving-average (ARMA) models, then highlight the
advantages and shortcomings of each approach. Finally, suggestions and lines of
enquiry for future work are provided.
| stat |
On a Japanese Subjective Well-Being Indicator Based on Twitter data | This study presents for the first time the SWB-J index, a subjective
well-being indicator for Japan based on Twitter data. The index is composed by
eight dimensions of subjective well-being and is estimated relying on Twitter
data by using human supervised sentiment analysis. The index is then compared
with the analogous SWB-I index for Italy, in order to verify possible analogies
and cultural differences. Further, through structural equation models, a causal
assumption is tested to see whether the economic and health conditions of the
country influence the well-being latent variable and how this latent dimension
affects the SWB-J and SWB-I indicators. It turns out that, as expected, the
economic and health welfare is only one aspect of the multidimensional
well-being that is captured by the Twitter-based indicator.
| stat |
ERStruct: An Eigenvalue Ratio Approach to Inferring Population Structure
from Sequencing Data | Inference of population structure from genetic data plays an important role
in population and medical genetics studies. The traditional EIGENSTRAT method
has been widely used for computing and selecting top principal components that
capture population structure information (Price et al., 2006). With the
advancement and decreasing cost of sequencing technology, whole-genome
sequencing data provide much richer information about the underlying population
structures. However, the EIGENSTRAT method was originally developed for
analyzing array-based genotype data and thus may not perform well on sequencing
data for two reasons. First, the number of genetic variants $p$ is much larger
than the sample size $n$ in sequencing data such that the sample-to-marker
ratio $n/p$ is nearly zero, violating the assumption of the Tracy-Widom test
used in the EIGENSTRAT method. Second, the EIGENSTRAT method might not be able
to handle the linkage disequilibrium (LD) well in sequencing data. To resolve
those two critical issues, we propose a new statistical method called ERStruct
to estimate the number of sub-populations based on sequencing data. We propose
to use the ratio of successive eigenvalues as a more robust testing statistic,
and then we approximate the null distribution of our proposed test statistic
using modern random matrix theory. Simulation studies found that our proposed
ERStruct method has improved performance compared to the traditional
Tracy-Widom test on sequencing data. We further illustrate our ERStruct method
using the sequencing data set from the 1000 Genomes Project. We also
implemented our ERStruct in a MATLAB toolbox which is now publicly available on
github: https://github.com/bglvly/ERStruct.
| stat |
A new segmentation method for the homogenisation of GNSS-derived IWV
time-series | Homogenization is an important and crucial step to improve the usage of
observational data for climate analysis. This work is motivated by the analysis
of long series of GNSS Integrated Water Vapour (IWV) data which have not yet
been used in this context. This paper proposes a novel segmentation method that
integrates a periodic bias and a heterogeneous, monthly varying, variance. The
method consists in estimating first the variance using a robust estimator and
then estimating the segmentation and periodic bias iteratively. This strategy
allows for the use of the dynamic programming algorithm that remains the most
efficient exact algorithm to estimate the change-point positions. The
statistical performance of the method is assessed through numerical
experiments. An application to a real data set of 120 global GNSS stations is
presented. The method is implemented in the R package GNSSseg that will be
available on the CRAN.
| stat |
A new multivariate meta-analysis model for many variates and few studies | Studies often estimate associations between an outcome and multiple variates.
For example, studies of diagnostic test accuracy estimate sensitivity and
specificity, and studies of predictive and prognostic factors typically
estimate associations for multiple factors. Meta-analysis is a family of
statistical methods for synthesizing estimates across multiple studies.
Multivariate models exist that account for within-study correlations and
between-study heterogeneity. The number of parameters that must be estimated in
existing models is quadratic in the number of variates (e.g., risk factors).
This means they may not be usable if data are sparse with many variates and few
studies. We propose a new model that addresses this problem by approximating a
variance-covariance matrix that models within-study correlation and
between-study heterogeneity in a low-dimensional space using random projection.
The number of parameters that must be estimated in this model scales linearly
in the number of variates and quadratically in the dimension of the
approximating space, making estimation more tractable. We performed a
simulation study to compare coverage, bias, and precision of estimates made
using the proposed model to those from univariate meta-analyses. We demonstrate
the method using data from an ongoing systematic review on predictors of pain
and function after total knee arthroplasty. Finally, we suggest a decision tool
to help analysts choose among available models.
| stat |
Functional data analysis: An application to COVID-19 data in the United
States | The COVID-19 pandemic so far has caused huge negative impacts on different
areas all over the world, and the United States (US) is one of the most
affected countries. In this paper, we use methods from the functional data
analysis to look into the COVID-19 data in the US. We explore the modes of
variation of the data through a functional principal component analysis (FPCA),
and study the canonical correlation between confirmed and death cases. In
addition, we run a cluster analysis at the state level so as to investigate the
relation between geographical locations and the clustering structure. Lastly,
we consider a functional time series model fitted to the cumulative confirmed
cases in the US, and make forecasts based on the dynamic FPCA. Both point and
interval forecasts are provided, and the methods for assessing the accuracy of
the forecasts are also included.
| stat |
Projected Estimation for Large-dimensional Matrix Factor Models | In this study, we propose a projection estimation method for
large-dimensional matrix factor models with cross-sectionally spiked
eigenvalues. By projecting the observation matrix onto the row or column factor
space, we simplify factor analysis for matrix series to that for a
lower-dimensional tensor. This method also reduces the magnitudes of the
idiosyncratic error components, thereby increasing the signal-to-noise ratio,
because the projection matrix linearly filters the idiosyncratic error matrix.
We theoretically prove that the projected estimators of the factor loading
matrices achieve faster convergence rates than existing estimators under
similar conditions. Asymptotic distributions of the projected estimators are
also presented. A novel iterative procedure is given to specify the pair of row
and column factor numbers. Extensive numerical studies verify the empirical
performance of the projection method. Two real examples in finance and
macroeconomics reveal factor patterns across rows and columns, which coincides
with financial, economic, or geographical interpretations.
| stat |
Statistical matching of non-Gaussian data | The statistical matching problem is a data integration problem with
structured missing data. The general form involves the analysis of multiple
datasets that only have a strict subset of variables jointly observed across
all datasets. The simplest version involves two datasets, labelled A and B,
with three variables of interest $X, Y$ and $Z$. Variables $X$ and $Y$ are
observed in dataset A and variables $X$ and $Z$ are observed in dataset $B$.
Statistical inference is complicated by the absence of joint $(Y, Z)$
observations. Parametric modelling can be challenging due to identifiability
issues and the difficulty of parameter estimation. We develop computationally
feasible procedures for the statistical matching of non-Gaussian data using
suitable data augmentation schemes and identifiability constraints.
Nearest-neighbour imputation is a common alternative technique due to its ease
of use and generality. Nearest-neighbour matching is based on a conditional
independence assumption that may be inappropriate for non-Gaussian data. The
violation of the conditional independence assumption can lead to improper
imputations. We compare model based approaches to nearest-neighbour imputation
on a number of flow cytometry datasets and find that the model based approach
can address some of the weaknesses of the nonparametric nearest-neighbour
technique.
| stat |
Sequential Bayesian experimental design for estimation of extreme-event
probability in stochastic dynamical systems | We consider a dynamical system with two sources of uncertainties: (1)
parameterized input with a known probability distribution and (2) stochastic
input-to-response (ItR) function with heteroscedastic randomness. Our purpose
is to efficiently quantify the extreme response probability when the ItR
function is expensive to evaluate. The problem setup arises often in physics
and engineering problems, with randomness in ItR coming from either intrinsic
uncertainties (say, as a solution to a stochastic equation) or additional
(critical) uncertainties that are not incorporated in the input parameter
space. To reduce the required sampling numbers, we develop a sequential
Bayesian experimental design method leveraging the variational heteroscedastic
Gaussian process regression (VHGPR) to account for the stochastic ItR, along
with a new criterion to select the next-best samples sequentially. The validity
of our new method is first tested in two synthetic problems with the stochastic
ItR functions defined artificially. Finally, we demonstrate the application of
our method to an engineering problem of estimating the extreme ship motion
probability in ensemble of wave groups, where the uncertainty in ItR naturally
originates from the uncertain initial condition of ship motion in each wave
group.
| stat |
A one-way ANOVA test for functional data with graphical interpretation | A new functional ANOVA test, with a graphical interpretation of the result,
is presented. The test is an extension of the global envelope test introduced
by Myllymaki et al. (2017, Global envelope tests for spatial processes, J. R.
Statist. Soc. B 79, 381--404, doi: 10.1111/rssb.12172). The graphical
interpretation is realized by a global envelope which is drawn jointly for all
samples of functions. If a mean function computed from the empirical data is
out of the given envelope, the null hypothesis is rejected with the
predetermined significance level $\alpha$. The advantages of the proposed
one-way functional ANOVA are that it identifies the domains of the functions
which are responsible for the potential rejection. We introduce two versions of
this test: the first gives a graphical interpretation of the test results in
the original space of the functions and the second immediately offers a
post-hoc test by identifying the significant pair-wise differences between
groups. The proposed tests rely on discretization of the functions, therefore
the tests are also applicable in the multidimensional ANOVA problem. In the
empirical part of the article, we demonstrate the use of the method by
analyzing fiscal decentralization in European countries. The aim of the
empirical analysis is to capture differences between the levels of government
expenditure decentralization ratio among different groups of European
countries. The idea behind, based on the existing literature, is
straightforward: countries with a longer European integration history are
supposed to decentralize more of their government expenditure. We use the
government expenditure centralization ratios of 29 European Union and EFTA
countries in period from 1995 to 2016 sorted into three groups according to the
presumed level of European economic and political integration.
| stat |
Sparse recovery by reduced variance stochastic approximation | In this paper, we discuss application of iterative Stochastic Optimization
routines to the problem of sparse signal recovery from noisy observation. Using
Stochastic Mirror Descent algorithm as a building block, we develop a
multistage procedure for recovery of sparse solutions to Stochastic
Optimization problem under assumption of smoothness and quadratic minoration on
the expected objective. An interesting feature of the proposed algorithm is its
linear convergence of the approximate solution during the preliminary phase of
the routine when the component of stochastic error in the gradient observation
which is due to bad initial approximation of the optimal solution is larger
than the "ideal" asymptotic error component owing to observation noise "at the
optimal solution." We also show how one can straightforwardly enhance
reliability of the corresponding solution by using Median-of-Means like
techniques. We illustrate the performance of the proposed algorithms in
application to classical problems of recovery of sparse and low rank signals in
linear regression framework. We show, under rather weak assumption on the
regressor and noise distributions, how they lead to parameter estimates which
obey (up to factors which are logarithmic in problem dimension and confidence
level) the best known to us accuracy bounds.
| stat |
A Nonparametric Bayesian Item Response Modeling Approach for Clustering
Items and Individuals Simultaneously | Item response theory (IRT) is a popular modeling paradigm for measuring
subject latent traits and item properties according to discrete responses in
tests or questionnaires. There are very limited discussions on heterogeneity
pattern detection for both items and individuals. In this paper, we introduce a
nonparametric Bayesian approach for clustering items and individuals
simultaneously under the Rasch model. Specifically, our proposed method is
based on the mixture of finite mixtures (MFM) model. MFM obtains the number of
clusters and the clustering configurations for both items and individuals
simultaneously. The performance of parameters estimation and parameters
clustering under the MFM Rasch model is evaluated by simulation studies, and a
real date set is applied to illustrate the MFM Rasch modeling.
| stat |
Bayesian Testing for Exogenous Partition Structures in Stochastic Block
Models | Network data often exhibit block structures characterized by clusters of
nodes with similar patterns of edge formation. When such relational data are
complemented by additional information on exogenous node partitions, these
sources of knowledge are typically included in the model to supervise the
cluster assignment mechanism or to improve inference on edge probabilities.
Although these solutions are routinely implemented, there is a lack of formal
approaches to test if a given external node partition is in line with the
endogenous clustering structure encoding stochastic equivalence patterns among
the nodes in the network. To fill this gap, we develop a formal Bayesian
testing procedure which relies on the calculation of the Bayes factor between a
stochastic block model with known grouping structure defined by the exogenous
node partition and an infinite relational model that allows the endogenous
clustering configurations to be unknown, random and fully revealed by the
block-connectivity patterns in the network. A simple Markov chain Monte Carlo
method for computing the Bayes factor and quantifying uncertainty in the
endogenous groups is proposed. This routine is evaluated in simulations and in
an application to study exogenous equivalence structures in brain networks of
Alzheimer's patients.
| stat |
Learning Disentangled Representations via Mutual Information Estimation | In this paper, we investigate the problem of learning disentangled
representations. Given a pair of images sharing some attributes, we aim to
create a low-dimensional representation which is split into two parts: a shared
representation that captures the common information between the images and an
exclusive representation that contains the specific information of each image.
To address this issue, we propose a model based on mutual information
estimation without relying on image reconstruction or image generation. Mutual
information maximization is performed to capture the attributes of data in the
shared and exclusive representations while we minimize the mutual information
between the shared and exclusive representation to enforce representation
disentanglement. We show that these representations are useful to perform
downstream tasks such as image classification and image retrieval based on the
shared or exclusive component. Moreover, classification results show that our
model outperforms the state-of-the-art model based on VAE/GAN approaches in
representation disentanglement.
| stat |
Accurate $p$-Value Calculation for Generalized Fisher's Combination
Tests Under Dependence | Combining dependent tests of significance has broad applications but the
$p$-value calculation is challenging. Current moment-matching methods (e.g.,
Brown's approximation) for Fisher's combination test tend to significantly
inflate the type I error rate at the level less than 0.05. It could lead to
significant false discoveries in big data analyses. This paper provides several
more accurate and computationally efficient $p$-value calculation methods for a
general family of Fisher type statistics, referred as the GFisher. The GFisher
covers Fisher's combination, Good's statistic, Lancaster's statistic, weighted
Z-score combination, etc. It allows a flexible weighting scheme, as well as an
omnibus procedure that automatically adapts proper weights and degrees of
freedom to a given data. The new $p$-value calculation methods are based on
novel ideas of moment-ratio matching and joint-distribution surrogating.
Systematic simulations show that they are accurate under multivariate Gaussian,
and robust under the generalized linear model and the multivariate
$t$-distribution, down to at least $10^{-6}$ level. We illustrate the
usefulness of the GFisher and the new $p$-value calculation methods in
analyzing both simulated and real data of gene-based SNP-set association
studies in genetics. Relevant computation has been implemented into R package
$GFisher$.
| stat |
Meta-Learning by Adjusting Priors Based on Extended PAC-Bayes Theory | In meta-learning an agent extracts knowledge from observed tasks, aiming to
facilitate learning of novel future tasks. Under the assumption that future
tasks are 'related' to previous tasks, the accumulated knowledge should be
learned in a way which captures the common structure across learned tasks,
while allowing the learner sufficient flexibility to adapt to novel aspects of
new tasks. We present a framework for meta-learning that is based on
generalization error bounds, allowing us to extend various PAC-Bayes bounds to
meta-learning. Learning takes place through the construction of a distribution
over hypotheses based on the observed tasks, and its utilization for learning a
new task. Thus, prior knowledge is incorporated through setting an
experience-dependent prior for novel tasks. We develop a gradient-based
algorithm which minimizes an objective function derived from the bounds and
demonstrate its effectiveness numerically with deep neural networks. In
addition to establishing the improved performance available through
meta-learning, we demonstrate the intuitive way by which prior information is
manifested at different levels of the network.
| stat |
Extracting Trips from Multi-Sourced Data for Mobility Pattern Analysis:
An App-Based Data Example | Passively-generated data, such as GPS data and cellular data, bring
tremendous opportunities for human mobility analysis and transportation
applications. Since their primary purposes are often non-transportation
related, the passively-generated data need to be processed to extract trips.
Most existing trip extraction methods rely on data that are generated via a
single positioning technology such as GPS or triangulation through cellular
towers (thereby called single-sourced data), and methods to extract trips from
data generated via multiple positioning technologies (or, multi-sourced data)
are absent. And yet, multi-sourced data are now increasingly common. Generated
using multiple technologies (e.g., GPS, cellular network- and WiFi-based),
multi-sourced data contain high variances in their temporal and spatial
properties. In this study, we propose a 'Divide, Conquer and Integrate' (DCI)
framework to extract trips from multi-sourced data. We evaluate the proposed
framework by applying it to an app-based data, which is multi-sourced and has
high variances in both location accuracy and observation interval (i.e. time
interval between two consecutive observations). On a manually labeled sample of
the app-based data, the framework outperforms the state-of-the-art SVM model
that is designed for GPS data. The effectiveness of the framework is also
illustrated by consistent mobility patterns obtained from the app-based data
and an externally collected household travel survey data for the same region
and the same period.
| stat |
Discretely Relaxing Continuous Variables for tractable Variational
Inference | We explore a new research direction in Bayesian variational inference with
discrete latent variable priors where we exploit Kronecker matrix algebra for
efficient and exact computations of the evidence lower bound (ELBO). The
proposed "DIRECT" approach has several advantages over its predecessors; (i) it
can exactly compute ELBO gradients (i.e. unbiased, zero-variance gradient
estimates), eliminating the need for high-variance stochastic gradient
estimators and enabling the use of quasi-Newton optimization methods; (ii) its
training complexity is independent of the number of training points, permitting
inference on large datasets; and (iii) its posterior samples consist of sparse
and low-precision quantized integers which permit fast inference on hardware
limited devices. In addition, our DIRECT models can exactly compute statistical
moments of the parameterized predictive posterior without relying on Monte
Carlo sampling. The DIRECT approach is not practical for all likelihoods,
however, we identify a popular model structure which is practical, and
demonstrate accurate inference using latent variables discretized as extremely
low-precision 4-bit quantized integers. While the ELBO computations considered
in the numerical studies require over $10^{2352}$ log-likelihood evaluations,
we train on datasets with over two-million points in just seconds.
| stat |
Treatment effect estimation with Multilevel Regression and
Poststratification | Multilevel regression and poststratification (MRP) is a flexible modeling
technique that has been used in a broad range of small-area estimation
problems. Traditionally, MRP studies have been focused on non-causal settings,
where estimating a single population value using a nonrepresentative sample was
of primary interest. In this manuscript, MRP-style estimators will be evaluated
in an experimental causal inference setting. We simulate a large-scale
randomized control trial with a stratified cluster sampling design, and compare
traditional and nonparametric treatment effect estimation methods with MRP
methodology. Using MRP-style estimators, treatment effect estimates for areas
as small as 1.3$\%$ of the population have lower bias and variance than
standard causal inference methods, even in the presence of treatment effect
heterogeneity. The design of our simulation studies also requires us to build
upon a MRP variant that allows for non-census covariates to be incorporated
into poststratification.
| stat |
A Bayesian approach to study synergistic interaction effects in in-vitro
drug combination experiments | In cancer translational research, increasing effort is devoted to the study
of the combined effect of two drugs when they are administered simultaneously.
In this paper, we introduce a new approach to estimate the part of the effect
of the two drugs due to the interaction of the compounds, i.e. which is due to
synergistic or antagonistic effects of the two drugs, compared to a reference
value representing the condition when the combined compounds do not interact,
called zero-interaction. We describe an in-vitro cell viability experiment as a
random experiment, by interpreting cell viability as the probability of a cell
in the experiment to be viable after treatment, and including information
related to different exposure conditions. We propose a flexible Bayesian spline
regression framework for modelling the viability surface of two drugs combined
as a function of the concentrations. Since the proposed approach is based on a
statistical model, it allows to include replicates of the experiments, to
evaluate the uncertainty of the estimates, and to perform prediction. We test
the model fit and prediction performance on a simulation study, and on an
ovarian cancer cell dataset. Posterior estimates of the zero-interaction level
and of the synergy term, obtained via adaptive MCMC algorithms, are used to
compute interpretable measures of efficacy of the combined experiment,
including relative volume under the surface (rVUS) measures to summarise the
zero-interaction and synergy terms and a bi-variate alternative to the
well-known EC50 measure.
| stat |
Missing at Random or Not: A Semiparametric Testing Approach | Practical problems with missing data are common, and statistical methods have
been developed concerning the validity and/or efficiency of statistical
procedures. On a central focus, there have been longstanding interests on the
mechanism governing data missingness, and correctly deciding the appropriate
mechanism is crucially relevant for conducting proper practical investigations.
The conventional notions include the three common potential classes -- missing
completely at random, missing at random, and missing not at random. In this
paper, we present a new hypothesis testing approach for deciding between
missing at random and missing not at random. Since the potential alternatives
of missing at random are broad, we focus our investigation on a general class
of models with instrumental variables for data missing not at random. Our
setting is broadly applicable, thanks to that the model concerning the missing
data is nonparametric, requiring no explicit model specification for the data
missingness. The foundational idea is to develop appropriate discrepancy
measures between estimators whose properties significantly differ only when
missing at random does not hold. We show that our new hypothesis testing
approach achieves an objective data oriented choice between missing at random
or not. We demonstrate the feasibility, validity, and efficacy of the new test
by theoretical analysis, simulation studies, and a real data analysis.
| stat |
Robust Recursive Partitioning for Heterogeneous Treatment Effects with
Uncertainty Quantification | Subgroup analysis of treatment effects plays an important role in
applications from medicine to public policy to recommender systems. It allows
physicians (for example) to identify groups of patients for whom a given drug
or treatment is likely to be effective and groups of patients for which it is
not. Most of the current methods of subgroup analysis begin with a particular
algorithm for estimating individualized treatment effects (ITE) and identify
subgroups by maximizing the difference across subgroups of the average
treatment effect in each subgroup. These approaches have several weaknesses:
they rely on a particular algorithm for estimating ITE, they ignore
(in)homogeneity within identified subgroups, and they do not produce good
confidence estimates. This paper develops a new method for subgroup analysis,
R2P, that addresses all these weaknesses. R2P uses an arbitrary, exogenously
prescribed algorithm for estimating ITE and quantifies the uncertainty of the
ITE estimation, using a construction that is more robust than other methods.
Experiments using synthetic and semi-synthetic datasets (based on real data)
demonstrate that R2P constructs partitions that are simultaneously more
homogeneous within groups and more heterogeneous across groups than the
partitions produced by other methods. Moreover, because R2P can employ any ITE
estimator, it also produces much narrower confidence intervals with a
prescribed coverage guarantee than other methods.
| stat |
Interactive identification of individuals with positive treatment effect
while controlling false discoveries | Out of the participants in a randomized experiment with anticipated
heterogeneous treatment effects, is it possible to identify which ones have a
positive treatment effect, even though each has only taken either treatment or
control but not both? While subgroup analysis has received attention, claims
about individual participants are more challenging. We frame the problem in
terms of multiple hypothesis testing: we think of each individual as a null
hypothesis (the potential outcomes are equal, for example) and aim to identify
individuals for whom the null is false (the treatment potential outcome
stochastically dominates the control, for example). We develop a novel
algorithm that identifies such a subset, with nonasymptotic control of the
false discovery rate (FDR). Our algorithm allows for interaction -- a human
data scientist (or a computer program acting on the human's behalf) may
adaptively guide the algorithm in a data-dependent manner to gain high
identification power. We also propose several extensions: (a) relaxing the null
to nonpositive effects, (b) moving from unpaired to paired samples, and (c)
subgroup identification. We demonstrate via numerical experiments and
theoretical analysis that the proposed method has valid FDR control in finite
samples and reasonably high identification power.
| stat |
Selective machine learning of doubly robust functionals | While model selection is a well-studied topic in parametric and nonparametric
regression or density estimation, selection of possibly high-dimensional
nuisance parameters in semiparametric problems is far less developed. In this
paper, we propose a selective machine learning framework for making inferences
about a finite-dimensional functional defined on a semiparametric model, when
the latter admits a doubly robust estimating function and several candidate
machine learning algorithms are available for estimating the nuisance
parameters. We introduce two new selection criteria for bias reduction in
estimating the functional of interest, each based on a novel definition of
pseudo-risk for the functional that embodies the double robustness property and
thus is used to select the pair of learners that is nearest to fulfilling this
property. We establish an oracle property for a multi-fold cross-validation
version of the new selection criteria which states that our empirical criteria
perform nearly as well as an oracle with a priori knowledge of the pseudo-risk
for each pair of candidate learners. We also describe a smooth approximation to
the selection criteria which allows for valid post-selection inference.
Finally, we apply the approach to model selection of a semiparametric estimator
of average treatment effect given an ensemble of candidate machine learners to
account for confounding in an observational study.
| stat |
A mixed model approach to drought prediction using artificial neural
networks: Case of an operational drought monitoring environment | Droughts, with their increasing frequency of occurrence, continue to
negatively affect livelihoods and elements at risk. For example, the 2011 in
drought in east Africa has caused massive losses document to have cost the
Kenyan economy over $12bn. With the foregoing, the demand for ex-ante drought
monitoring systems is ever-increasing. The study uses 10 precipitation and
vegetation variables that are lagged over 1, 2 and 3-month time-steps to
predict drought situations. In the model space search for the most predictive
artificial neural network (ANN) model, as opposed to the traditional greedy
search for the most predictive variables, we use the General Additive Model
(GAM) approach. Together with a set of assumptions, we thereby reduce the
cardinality of the space of models. Even though we build a total of 102 GAM
models, only 21 have R2 greater than 0.7 and are thus subjected to the ANN
process. The ANN process itself uses the brute-force approach that
automatically partitions the training data into 10 sub-samples, builds the ANN
models in these samples and evaluates their performance using multiple metrics.
The results show the superiority of 1-month lag of the variables as compared to
longer time lags of 2 and 3 months. The champion ANN model recorded an R2 of
0.78 in model testing using the out-of-sample data. This illustrates its
ability to be a good predictor of drought situations 1-month ahead.
Investigated as a classifier, the champion has a modest accuracy of 66% and a
multi-class area under the ROC curve (AUROC) of 89.99%
| stat |
Adversarial Examples Are Not Bugs, They Are Features | Adversarial examples have attracted significant attention in machine
learning, but the reasons for their existence and pervasiveness remain unclear.
We demonstrate that adversarial examples can be directly attributed to the
presence of non-robust features: features derived from patterns in the data
distribution that are highly predictive, yet brittle and incomprehensible to
humans. After capturing these features within a theoretical framework, we
establish their widespread existence in standard datasets. Finally, we present
a simple setting where we can rigorously tie the phenomena we observe in
practice to a misalignment between the (human-specified) notion of robustness
and the inherent geometry of the data.
| stat |
Wearables and location tracking technologies for mental-state sensing in
outdoor environments | Advances in commercial wearable devices are increasingly facilitating the
collection and analysis of everyday physiological data. This paper discusses
the theoretical and practical aspects of using such ambulatory devices for the
detection of episodic changes in physiological signals as a marker for mental
state in outdoor environments. A pilot study was conducted to evaluate the
feasibility of utilizing commercial wearables in combination with location
tracking technologies. The study measured physiological signals for 15
participants, including heart rate, heart-rate variability, and skin
conductance. Participants' signals were recorded during an outdoor walk that
was tracked using a GPS logger. The walk was designed to pass through various
types of environments including green, blue, and urban spaces as well as a more
stressful road crossing. The data that was obtained was used to demonstrate how
biosensors information can be contextualized and enriched using location
information. Significant episodic changes in physiological signals under
real-world conditions were detectable in the stressful road crossing, but not
in the other types of environments. The article concludes that despite
challenges and limitations of current off-the-shelf wearables, the utilization
of these devices offers novel opportunities for evaluating episodic changes in
physiological signals as a marker for mental state during everyday activities
including in outdoor environments.
| stat |
Confounding Feature Acquisition for Causal Effect Estimation | Reliable treatment effect estimation from observational data depends on the
availability of all confounding information. While much work has targeted
treatment effect estimation from observational data, there is relatively little
work in the setting of confounding variable missingness, where collecting more
information on confounders is often costly or time-consuming. In this work, we
frame this challenge as a problem of feature acquisition of confounding
features for causal inference. Our goal is to prioritize acquiring values for a
fixed and known subset of missing confounders in samples that lead to efficient
average treatment effect estimation. We propose two acquisition strategies
based on i) covariate balancing (CB), and ii) reducing statistical estimation
error on observed factual outcome error (OE). We compare CB and OE on five
common causal effect estimation methods, and demonstrate improved sample
efficiency of OE over baseline methods under various settings. We also provide
visualizations for further analysis on the difference between our proposed
methods.
| stat |
End-to-End Probabilistic Inference for Nonstationary Audio Analysis | A typical audio signal processing pipeline includes multiple disjoint
analysis stages, including calculation of a time-frequency representation
followed by spectrogram-based feature analysis. We show how time-frequency
analysis and nonnegative matrix factorisation can be jointly formulated as a
spectral mixture Gaussian process model with nonstationary priors over the
amplitude variance parameters. Further, we formulate this nonlinear model's
state space representation, making it amenable to infinite-horizon Gaussian
process regression with approximate inference via expectation propagation,
which scales linearly in the number of time steps and quadratically in the
state dimensionality. By doing so, we are able to process audio signals with
hundreds of thousands of data points. We demonstrate, on various tasks with
empirical data, how this inference scheme outperforms more standard techniques
that rely on extended Kalman filtering.
| stat |
Heterogeneous Multi-output Gaussian Process Prediction | We present a novel extension of multi-output Gaussian processes for handling
heterogeneous outputs. We assume that each output has its own likelihood
function and use a vector-valued Gaussian process prior to jointly model the
parameters in all likelihoods as latent functions. Our multi-output Gaussian
process uses a covariance function with a linear model of coregionalisation
form. Assuming conditional independence across the underlying latent functions
together with an inducing variable framework, we are able to obtain tractable
variational bounds amenable to stochastic variational inference. We illustrate
the performance of the model on synthetic data and two real datasets: a human
behavioral study and a demographic high-dimensional dataset.
| stat |
Interference, Bias, and Variance in Two-Sided Marketplace
Experimentation: Guidance for Platforms | Two-sided marketplace platforms often run experiments to test the effect of
an intervention before launching it platform-wide. A typical approach is to
randomize individuals into the treatment group, which receives the
intervention, and the control group, which does not. The platform then compares
the performance in the two groups to estimate the effect if the intervention
were launched to everyone. We focus on two common experiment types, where the
platform randomizes individuals either on the supply side or on the demand
side. The resulting estimates of the treatment effect in these experiments are
typically biased: because individuals in the market compete with each other,
individuals in the treatment group affect those in the control group and vice
versa, creating interference. We develop a simple tractable market model to
study bias and variance in these experiments with interference. We focus on two
choices available to the platform: (1) Which side of the platform should it
randomize on (supply or demand)? (2) What proportion of individuals should be
allocated to treatment? We find that both choices affect the bias and variance
of the resulting estimators but in different ways. The bias-optimal choice of
experiment type depends on the relative amounts of supply and demand in the
market, and we discuss how a platform can use market data to select the
experiment type. Importantly, we find in many circumstances, choosing the
bias-optimal experiment type has little effect on variance. On the other hand,
the choice of treatment proportion can induce a bias-variance tradeoff, where
the bias-minimizing proportion increases variance. We discuss how a platform
can navigate this tradeoff and best choose the treatment proportion, using a
combination of modeling as well as contextual knowledge about the market, the
risk of the intervention, and reasonable effect sizes of the intervention.
| stat |
Community Detection for Hypergraph Networks via Regularized Tensor Power
Iteration | To date, social network analysis has been largely focused on pairwise
interactions. The study of higher-order interactions, via a hypergraph network,
brings in new insights. We study community detection in a hypergraph network. A
popular approach is to project the hypergraph to a graph and then apply
community detection methods for graph networks, but we show that this approach
may cause unwanted information loss. We propose a new method for community
detection that operates directly on the hypergraph. At the heart of our method
is a regularized higher-order orthogonal iteration (reg-HOOI) algorithm that
computes an approximate low-rank decomposition of the network adjacency tensor.
Compared with existing tensor decomposition methods such as HOSVD and vanilla
HOOI, reg-HOOI yields better performance, especially when the hypergraph is
sparse. Given the output of tensor decomposition, we then generalize the
community detection method SCORE (Jin, 2015) from graph networks to hypergraph
networks. We call our new method Tensor-SCORE.
In theory, we introduce a degree-corrected block model for hypergraphs
(hDCBM), and show that Tensor-SCORE yields consistent community detection for a
wide range of network sparsity and degree heterogeneity. As a byproduct, we
derive the rates of convergence on estimating the principal subspace by
reg-HOOI, with different initializations, including the two new initialization
methods we propose, a diagonal-removed HOSVD and a randomized graph projection.
We apply our method to several real hypergraph networks which yields
encouraging results. It suggests that exploring higher-order interactions
provides additional information not seen in graph representations.
| stat |
Spatio-Temporal RBF Neural Networks | Herein, we propose a spatio-temporal extension of RBFNN for nonlinear system
identification problem. The proposed algorithm employs the concept of
time-space orthogonality and separately models the dynamics and nonlinear
complexities of the system. The proposed RBF architecture is explored for the
estimation of a highly nonlinear system and results are compared with the
standard architecture for both the conventional and fractional gradient
decent-based learning rules. The spatio-temporal RBF is shown to perform better
than the standard and fractional RBFNNs by achieving fast convergence and
significantly reduced estimation error.
| stat |
From the power law to extreme value mixture distributions | The power law is useful in describing count phenomena such as network degrees
and word frequencies. With a single parameter, it captures the main feature
that the frequencies are linear on the log-log scale. Nevertheless, there have
been criticisms of the power law, and various approaches have been proposed to
resolve issues such as selecting the required threshold and quantifying the
uncertainty around it, and to test hypotheses on whether the data could have
come from the power law. As extreme value theory generalises the (continuous)
power law, it is natural to consider the former as a solution to these problems
around the latter. In this paper, we propose two extreme value mixture
distributions, in one of which the power law is incorporated, without the need
of pre-specifying the threshold. The proposed distributions are shown to fit
the data well, quantify the threshold uncertainty in a natural way, and
satisfactorily answer whether the power law is useful enough.
| stat |
Recovering individual-level spatial inference from aggregated binary
data | Binary regression models are commonly used in disciplines such as
epidemiology and ecology to determine how spatial covariates influence
individuals. In many studies, binary data are shared in a spatially aggregated
form to protect privacy. For example, rather than reporting the location and
result for each individual that was tested for a disease, researchers may
report that a disease was detected or not detected within geopolitical units.
Often, the spatial aggregation process obscures the values of response
variables, spatial covariates, and locations of each individual, which makes
recovering individual-level inference difficult. We show that applying a series
of transformations, including a change of support, to a bivariate point process
model allows researchers to recover individual-level inference for spatial
covariates from spatially aggregated binary data. The series of transformations
preserves the convenient interpretation of desirable binary regression models
that are commonly applied to individual-level data. Using a simulation
experiment, we compare the performance of our proposed method under varying
types of spatial aggregation against the performance of standard approaches
using the original individual-level data. We illustrate our method by modeling
individual-level probability of infection using a data set that has been
aggregated to protect an at-risk and endangered species of bats. Our simulation
experiment and data illustration demonstrate the utility of the proposed method
when access to original non-aggregated data is impractical or prohibited.
| stat |
A Unified Method for Improved Inference in Random-effects Meta-analysis | Random-effects meta-analyses have been widely applied in evidence synthesis
for various types of medical studies. However, standard inference methods (e.g.
restricted maximum likelihood estimation) usually underestimate statistical
errors and possibly provide highly overconfident results under realistic
situations; for instance, coverage probabilities of confidence intervals can be
substantially below the nominal level. The main reason is that these inference
methods rely on large sample approximations even though the number of
synthesized studies is usually small or moderate in practice. In this article
we solve this problem using a unified inference method based on Monte Carlo
conditioning for broad application to random-effects meta-analysis. The
developed method provides improved confidence intervals with coverage
probabilities that are closer to the nominal level than standard methods. As
specific applications, we provide new inference procedures for three types of
meta-analysis: conventional univariate meta-analysis for pairwise treatment
comparisons, meta-analysis of diagnostic test accuracy, and multiple treatment
comparisons via network meta-analysis. We also illustrate the practical
effectiveness of these methods via real data applications and simulation
studies.
| stat |
An Analysis of the Adaptation Speed of Causal Models | Consider a collection of datasets generated by unknown interventions on an
unknown structural causal model $G$. Recently, Bengio et al. (2020) conjectured
that among all candidate models, $G$ is the fastest to adapt from one dataset
to another, along with promising experiments. Indeed, intuitively $G$ has less
mechanisms to adapt, but this justification is incomplete. Our contribution is
a more thorough analysis of this hypothesis. We investigate the adaptation
speed of cause-effect SCMs. Using convergence rates from stochastic
optimization, we justify that a relevant proxy for adaptation speed is distance
in parameter space after intervention. Applying this proxy to categorical and
normal cause-effect models, we show two results. When the intervention is on
the cause variable, the SCM with the correct causal direction is advantaged by
a large factor. When the intervention is on the effect variable, we
characterize the relative adaptation speed. Surprisingly, we find situations
where the anticausal model is advantaged, falsifying the initial hypothesis.
Code to reproduce experiments is available at
https://github.com/remilepriol/causal-adaptation-speed
| stat |
Adjusting for Unmeasured Confounding in Marginal Structural Models with
Propensity-Score Fixed Effects | Marginal structural models are a popular tool for investigating the effects
of time-varying treatments, but they require an assumption of no unobserved
confounders between the treatment and outcome. With observational data, this
assumption may be difficult to maintain, and in studies with panel data, many
researchers use fixed effects models to purge the data of time-constant
unmeasured confounding. Unfortunately, traditional linear fixed effects models
are not suitable for estimating the effects of time-varying treatments, since
they can only estimate lagged effects under implausible assumptions. To resolve
this tension, we a propose a novel inverse probability of treatment weighting
estimator with propensity-score fixed effects to adjust for time-constant
unmeasured confounding in marginal structural models of fixed-length treatment
histories. We show that these estimators are consistent and asymptotically
normal when the number of units and time periods grow at a similar rate. Unlike
traditional fixed effect models, this approach works even when the outcome is
only measured at a single point in time as is common in marginal structural
models. We apply these methods to estimating the effect of negative advertising
on the electoral success of candidates for statewide offices in the United
States.
| stat |
Na\"ive regression requires weaker assumptions than factor models to
adjust for multiple cause confounding | The empirical practice of using factor models to adjust for shared,
unobserved confounders, $\mathbf{Z}$, in observational settings with multiple
treatments, $\mathbf{A}$, is widespread in fields including genetics, networks,
medicine, and politics. Wang and Blei (2019, WB) formalizes these procedures
and develops the "deconfounder," a causal inference method using factor models
of $\mathbf{A}$ to estimate "substitute confounders," $\hat{\mathbf{Z}}$, then
estimating treatment effects by regressing the outcome, $\mathbf{Y}$, on part
of $\mathbf{A}$ while adjusting for $\hat{\mathbf{Z}}$. WB claim the
deconfounder is unbiased when there are no single-cause confounders and
$\hat{\mathbf{Z}}$ is "pinpointed." We clarify pinpointing requires each
confounder to affect infinitely many treatments. We prove under these
assumptions, a na\"ive semiparametric regression of $\mathbf{Y}$ on
$\mathbf{A}$ is asymptotically unbiased. Deconfounder variants nesting this
regression are therefore also asymptotically unbiased, but variants using
$\hat{\mathbf{Z}}$ and subsets of causes require further untestable
assumptions. We replicate every deconfounder analysis with available data and
find it fails to consistently outperform na\"ive regression. In practice, the
deconfounder produces implausible estimates in WB's case study to movie
earnings: estimates suggest comic author Stan Lee's cameo appearances causally
contributed \$15.5 billion, most of Marvel movie revenue. We conclude neither
approach is a viable substitute for careful research design in real-world
applications.
| stat |
A Semi-Automatic Method for History Matching using Sequential Monte
Carlo | The aim of the history matching method is to locate non-implausible regions
of the parameter space of complex deterministic or stochastic models by
matching model outputs with data. It does this via a series of waves where at
each wave an emulator is fitted to a small number of training samples. An
implausibility measure is defined which takes into account the closeness of
simulated and observed outputs as well as emulator uncertainty. As the waves
progress, the emulator becomes more accurate so that training samples are more
concentrated on promising regions of the space and poorer parts of the space
are rejected with more confidence. Whilst history matching has proved to be
useful, existing implementations are not fully automated and some ad-hoc
choices are made during the process, which involves user intervention and is
time consuming. This occurs especially when the non-implausible region becomes
small and it is difficult to sample this space uniformly to generate new
training points. In this article we develop a sequential Monte Carlo (SMC)
algorithm for implementing history matching that is semi-automated. Our novel
SMC approach reveals that the history matching method yields a non-implausible
region that can be multi-modal, highly irregular and very difficult to sample
uniformly. Our SMC approach offers a much more reliable sampling of the
non-implausible space, which requires additional computation compared to other
approaches used in the literature.
| stat |
Inference on high-dimensional implicit dynamic models using a guided
intermediate resampling filter | We propose a method for inference on moderately high-dimensional, nonlinear,
non-Gaussian, partially observed Markov process models for which the transition
density is not analytically tractable. Markov processes with intractable
transition densities arise in models defined implicitly by simulation
algorithms. Widely used particle filter methods are applicable to nonlinear,
non-Gaussian models but suffer from the curse of dimensionality. Improved
scalability is provided by ensemble Kalman filter methods, but these are
inappropriate for highly nonlinear and non-Gaussian models. We propose a
particle filter method having improved practical and theoretical scalability
with respect to the model dimension. This method is applicable to implicitly
defined models having analytically intractable transition densities. Our method
is developed based on the assumption that the latent process is defined in
continuous time and that a simulator of this latent process is available. In
this method, particles are propagated at intermediate time intervals between
observations and are resampled based on a forecast likelihood of future
observations. We combine this particle filter with parameter estimation
methodology to enable likelihood-based inference for highly nonlinear
spatiotemporal systems. We demonstrate our methodology on a stochastic Lorenz
96 model and a model for the population dynamics of infectious diseases in a
network of linked regions.
| stat |
A Hybrid Two-layer Feature Selection Method Using GeneticAlgorithm and
Elastic Net | Feature selection, as a critical pre-processing step for machine learning,
aims at determining representative predictors from a high-dimensional feature
space dataset to improve the prediction accuracy. However, the increase in
feature space dimensionality, comparing to the number of observations, poses a
severe challenge to many existing feature selection methods with respect to
computational efficiency and prediction performance. This paper presents a new
hybrid two-layer feature selection approach that combines a wrapper and an
embedded method in constructing an appropriate subset of predictors. In the
first layer of the proposed method, the Genetic Algorithm(GA) has been adopted
as a wrapper to search for the optimal subset of predictors, which aims to
reduce the number of predictors and the prediction error. As one of the
meta-heuristic approaches, GA is selected due to its computational efficiency;
however, GAs do not guarantee the optimality. To address this issue, a second
layer is added to the proposed method to eliminate any remaining
redundant/irrelevant predictors to improve the prediction accuracy. Elastic
Net(EN) has been selected as the embedded method in the second layer because of
its flexibility in adjusting the penalty terms in regularization process and
time efficiency. This hybrid two-layer approach has been applied on a Maize
genetic dataset from NAM population, which consists of multiple subsets of
datasets with different ratio of the number of predictors to the number of
observations. The numerical results confirm the superiority of the proposed
model.
| stat |
Double spike Dirichlet priors for structured weighting | Assigning weights to a large pool of objects is a fundamental task in a wide
variety of applications. In this article, we introduce a concept of structured
high-dimensional probability simplexes, whose most components are zero or near
zero and the remaining ones are close to each other. Such structure is well
motivated by 1) high-dimensional weights that are common in modern
applications, and 2) ubiquitous examples in which equal weights -- despite
their simplicity -- often achieve favorable or even state-of-the-art predictive
performances. This particular structure, however, presents unique challenges
both computationally and statistically. To address these challenges, we propose
a new class of double spike Dirichlet priors to shrink a probability simplex to
one with the desired structure. When applied to ensemble learning, such priors
lead to a Bayesian method for structured high-dimensional ensembles that is
useful for forecast combination and improving random forests, while enabling
uncertainty quantification. We design efficient Markov chain Monte Carlo
algorithms for easy implementation. Posterior contraction rates are established
to provide theoretical support. We demonstrate the wide applicability and
competitive performance of the proposed methods through simulations and two
real data applications using the European Central Bank Survey of Professional
Forecasters dataset and a UCI dataset.
| stat |
Invertible Gaussian Reparameterization: Revisiting the Gumbel-Softmax | The Gumbel-Softmax is a continuous distribution over the simplex that is
often used as a relaxation of discrete distributions. Because it can be readily
interpreted and easily reparameterized, it enjoys widespread use. We propose a
modular and more flexible family of reparameterizable distributions where
Gaussian noise is transformed into a one-hot approximation through an
invertible function. This invertible function is composed of a modified softmax
and can incorporate diverse transformations that serve different specific
purposes. For example, the stick-breaking procedure allows us to extend the
reparameterization trick to distributions with countably infinite support, thus
enabling the use of our distribution along nonparametric models, or normalizing
flows let us increase the flexibility of the distribution. Our construction
enjoys theoretical advantages over the Gumbel-Softmax, such as closed form KL,
and significantly outperforms it in a variety of experiments. Our code is
available at https://github.com/cunningham-lab/igr.
| stat |
Adaptive Network Sparsification with Dependent Variational
Beta-Bernoulli Dropout | While variational dropout approaches have been shown to be effective for
network sparsification, they are still suboptimal in the sense that they set
the dropout rate for each neuron without consideration of the input data. With
such input-independent dropout, each neuron is evolved to be generic across
inputs, which makes it difficult to sparsify networks without accuracy loss. To
overcome this limitation, we propose adaptive variational dropout whose
probabilities are drawn from sparsity-inducing beta Bernoulli prior. It allows
each neuron to be evolved either to be generic or specific for certain inputs,
or dropped altogether. Such input-adaptive sparsity-inducing dropout allows the
resulting network to tolerate larger degree of sparsity without losing its
expressive power by removing redundancies among features. We validate our
dependent variational beta-Bernoulli dropout on multiple public datasets, on
which it obtains significantly more compact networks than baseline methods,
with consistent accuracy improvements over the base networks.
| stat |
Effects of interventions and optimal strategies in the stochastic system
approach to causality | We consider the problem of defining the effect of an intervention on a
time-varying risk factor or treatment for a disease or a physiological marker;
we develop here the latter case. So, the system considered is $(Y,A,C)$, where
$Y=(Y_t)$, is the marker process of interest, $A=A_t$ the treatment. A
realistic case is that the treatment can be changed only at discrete times. In
an observational study the treatment attribution law is unknown; however, the
physical law can be estimated without knowing the treatment attribution law,
provided a well-specified model is available. An intervention is specified by
the treatment attribution law, which is thus known. Simple interventions will
simply randomize the attribution of the treatment; interventions that take into
account the past history will be called "strategies". The effect of
interventions can be defined by a risk function $R^{\intr}=\Ee_{\intr}[L(\bar
Y_{t_J}, \bar A_{t_{J}},C)]$, where $L(\bar Y_{t_J}, \bar A_{t_{J}},C)$ is a
loss function, and contrasts between risk functions for different strategies
can be formed. Once we can compute effects for any strategy, we can search for
optimal or sub-optimal strategies; in particular we can find optimal parametric
strategies. We present several ways for designing strategies. As an
illustration, we consider the choice of a strategy for containing the HIV load
below a certain level while limiting the treatment burden. A simulation study
demonstrates the possibility of finding optimal parametric strategies.
| stat |
Hypotheses Testing from Complex Survey Data Using Bootstrap Weights: A
Unified Approach | Standard statistical methods that do not take proper account of the
complexity of survey design can lead to erroneous inferences when applied to
survey data due to unequal selection probabilities, clustering, and other
design features. In particular, the actual type I error rates of tests of
hypotheses using standard methods can be much bigger than the nominal
significance level. Methods that take account of survey design features in
testing hypotheses have been proposed, including Wald tests and quasi-score
tests that involve the estimated covariance matrices of parameter estimates. In
this paper, we present a unified approach to hypothesis testing that does not
require computing the covariance matrices by constructing bootstrap
approximations to weighted likelihood ratio statistics and weighted quasi-score
statistics and establish the asymptotic validity of the proposed bootstrap
tests. In addition, we also consider hypothesis testing from categorical data
and present a bootstrap procedure for testing simple goodness of fit and
independence in a two-way table. In the simulation studies, the type I error
rates of the proposed approach are much closer to their nominal significance
level compared with the naive likelihood-ratio test and quasi-score test. An
application to data from an educational survey under a logistic regression
model is also presented.
| stat |
Direct Evolutionary Optimization of Variational Autoencoders With Binary
Latents | Discrete latent variables are considered important for real world data, which
has motivated research on Variational Autoencoders (VAEs) with discrete
latents. However, standard VAE-training is not possible in this case, which has
motivated different strategies to manipulate discrete distributions in order to
train discrete VAEs similarly to conventional ones. Here we ask if it is also
possible to keep the discrete nature of the latents fully intact by applying a
direct discrete optimization for the encoding model. The approach is
consequently strongly diverting from standard VAE-training by sidestepping
sampling approximation, reparameterization trick and amortization. Discrete
optimization is realized in a variational setting using truncated posteriors in
conjunction with evolutionary algorithms. For VAEs with binary latents, we (A)
show how such a discrete variational method ties into gradient ascent for
network weights, and (B) how the decoder is used to select latent states for
training. Conventional amortized training is more efficient and applicable to
large neural networks. However, using smaller networks, we here find direct
discrete optimization to be efficiently scalable to hundreds of latents. More
importantly, we find the effectiveness of direct optimization to be highly
competitive in `zero-shot' learning. In contrast to large supervised networks,
the here investigated VAEs can, e.g., denoise a single image without previous
training on clean data and/or training on large image datasets. More generally,
the studied approach shows that training of VAEs is indeed possible without
sampling-based approximation and reparameterization, which may be interesting
for the analysis of VAE-training in general. For `zero-shot' settings a direct
optimization, furthermore, makes VAEs competitive where they have previously
been outperformed by non-generative approaches.
| stat |
Polarizing Front Ends for Robust CNNs | The vulnerability of deep neural networks to small, adversarially designed
perturbations can be attributed to their "excessive linearity." In this paper,
we propose a bottom-up strategy for attenuating adversarial perturbations using
a nonlinear front end which polarizes and quantizes the data. We observe that
ideal polarization can be utilized to completely eliminate perturbations,
develop algorithms to learn approximately polarizing bases for data, and
investigate the effectiveness of the proposed strategy on the MNIST and Fashion
MNIST datasets.
| stat |
Denoising Score-Matching for Uncertainty Quantification in Inverse
Problems | Deep neural networks have proven extremely efficient at solving a wide
rangeof inverse problems, but most often the uncertainty on the solution they
provideis hard to quantify. In this work, we propose a generic Bayesian
framework forsolving inverse problems, in which we limit the use of deep neural
networks tolearning a prior distribution on the signals to recover. We adopt
recent denoisingscore matching techniques to learn this prior from data, and
subsequently use it aspart of an annealed Hamiltonian Monte-Carlo scheme to
sample the full posteriorof image inverse problems. We apply this framework to
Magnetic ResonanceImage (MRI) reconstruction and illustrate how this approach
not only yields highquality reconstructions but can also be used to assess the
uncertainty on particularfeatures of a reconstructed image.
| stat |
Size of Interventional Markov Equivalence Classes in Random DAG Models | Directed acyclic graph (DAG) models are popular for capturing causal
relationships. From observational and interventional data, a DAG model can only
be determined up to its \emph{interventional Markov equivalence class} (I-MEC).
We investigate the size of MECs for random DAG models generated by uniformly
sampling and ordering an Erd\H{o}s-R\'{e}nyi graph. For constant density, we
show that the expected $\log$ observational MEC size asymptotically (in the
number of vertices) approaches a constant. We characterize I-MEC size in a
similar fashion in the above settings with high precision. We show that the
asymptotic expected number of interventions required to fully identify a DAG is
a constant. These results are obtained by exploiting Meek rules and coupling
arguments to provide sharp upper and lower bounds on the asymptotic quantities,
which are then calculated numerically up to high precision. Our results have
important consequences for experimental design of interventions and the
development of algorithms for causal inference.
| stat |
ForecastTB An R Package as a Test-Bench for Time Series Forecasting
Application of Wind Speed and Solar Radiation Modeling | This paper introduces an R package ForecastTB that can be used to compare the
accuracy of different forecasting methods as related to the characteristics of
a time series dataset. The ForecastTB is a plug-and-play structured module, and
several forecasting methods can be included with simple instructions. The
proposed test-bench is not limited to the default forecasting and error metric
functions, and users are able to append, remove, or choose the desired methods
as per requirements. Besides, several plotting functions and statistical
performance metrics are provided to visualize the comparative performance and
accuracy of different forecasting methods. Furthermore, this paper presents
real application examples with natural time series datasets (i.e., wind speed
and solar radiation) to exhibit the features of the ForecastTB package to
evaluate forecasting comparison analysis as affected by the characteristics of
a dataset. Modeling results indicated the applicability and robustness of the
proposed R package ForecastTB for time series forecasting.
| stat |
Statistical Inference for distributions with one Poisson conditional | It will be recalled that the classical bivariate normal distributions have
normal marginals and normal conditionals. It is natural to ask whether a
similar phenomenon can be encountered involving Poisson marginals and
conditionals. Reference to Arnold, Castillo and Sarabia's (1999) book on
conditionally specified models will confirm that Poisson marginals will be
encountered, together with both conditionals being of the Poisson form, only in
the case in which the variables are independent. Instead, in the present
article we will be focusing on bivariate distributions with one marginal and
the other family of conditionals being of the Poisson form. Such distributions
are called Pseudo-Poisson distributions. We discuss distributional features of
such models, explore inferential aspects and include an example of applications
of the Pseudo-Poisson model to sets of over-dispersed data.
| stat |
Stochastic Normalizing Flows | We introduce stochastic normalizing flows, an extension of continuous
normalizing flows for maximum likelihood estimation and variational inference
(VI) using stochastic differential equations (SDEs). Using the theory of rough
paths, the underlying Brownian motion is treated as a latent variable and
approximated, enabling efficient training of neural SDEs as random neural
ordinary differential equations. These SDEs can be used for constructing
efficient Markov chains to sample from the underlying distribution of a given
dataset. Furthermore, by considering families of targeted SDEs with prescribed
stationary distribution, we can apply VI to the optimization of hyperparameters
in stochastic MCMC.
| stat |
Decoder-free Robustness Disentanglement without (Additional) Supervision | Adversarial Training (AT) is proposed to alleviate the adversarial
vulnerability of machine learning models by extracting only robust features
from the input, which, however, inevitably leads to severe accuracy reduction
as it discards the non-robust yet useful features. This motivates us to
preserve both robust and non-robust features and separate them with
disentangled representation learning. Our proposed Adversarial Asymmetric
Training (AAT) algorithm can reliably disentangle robust and non-robust
representations without additional supervision on robustness. Empirical results
show our method does not only successfully preserve accuracy by combining two
representations, but also achieve much better disentanglement than previous
work.
| stat |
The Convex Information Bottleneck Lagrangian | The information bottleneck (IB) problem tackles the issue of obtaining
relevant compressed representations $T$ of some random variable $X$ for the
task of predicting $Y$. It is defined as a constrained optimization problem
which maximizes the information the representation has about the task,
$I(T;Y)$, while ensuring that a certain level of compression $r$ is achieved
(i.e., $ I(X;T) \leq r$). For practical reasons, the problem is usually solved
by maximizing the IB Lagrangian (i.e., $\mathcal{L}_{\text{IB}}(T;\beta) =
I(T;Y) - \beta I(X;T)$) for many values of $\beta \in [0,1]$. Then, the curve
of maximal $I(T;Y)$ for a given $I(X;T)$ is drawn and a representation with the
desired predictability and compression is selected. It is known when $Y$ is a
deterministic function of $X$, the IB curve cannot be explored and another
Lagrangian has been proposed to tackle this problem: the squared IB Lagrangian:
$\mathcal{L}_{\text{sq-IB}}(T;\beta_{\text{sq}})=I(T;Y)-\beta_{\text{sq}}I(X;T)^2$.
In this paper, we (i) present a general family of Lagrangians which allow for
the exploration of the IB curve in all scenarios; (ii) provide the exact
one-to-one mapping between the Lagrange multiplier and the desired compression
rate $r$ for known IB curve shapes; and (iii) show we can approximately obtain
a specific compression level with the convex IB Lagrangian for both known and
unknown IB curve shapes. This eliminates the burden of solving the optimization
problem for many values of the Lagrange multiplier. That is, we prove that we
can solve the original constrained problem with a single optimization.
| stat |
BooST: Boosting Smooth Trees for Partial Effect Estimation in Nonlinear
Regressions | In this paper, we introduce a new machine learning (ML) model for nonlinear
regression called the Boosted Smooth Transition Regression Trees (BooST), which
is a combination of boosting algorithms with smooth transition regression
trees. The main advantage of the BooST model is the estimation of the
derivatives (partial effects) of very general nonlinear models. Therefore, the
model can provide more interpretation about the mapping between the covariates
and the dependent variable than other tree-based models, such as Random
Forests. We present several examples with both simulated and real data.
| stat |
Mixture Models and Networks -- Overview of Stochastic Blockmodelling | Mixture models are probabilistic models aimed at uncovering and representing
latent subgroups within a population. In the realm of network data analysis,
the latent subgroups of nodes are typically identified by their connectivity
behaviour, with nodes behaving similarly belonging to the same community. In
this context, mixture modelling is pursued through stochastic blockmodelling.
We consider stochastic blockmodels and some of their variants and extensions
from a mixture modelling perspective. We also survey some of the main classes
of estimation methods available, and propose an alternative approach. In
addition to the discussion of inferential properties and estimating procedures,
we focus on the application of the models to several real-world network
datasets, showcasing the advantages and pitfalls of different approaches.
| stat |
Preserving the distribution function in surveys in case of imputation
for zero inflated data | Item non-response in surveys is usually handled by single imputation, whose
main objective is to reduce the non-response bias. Imputation methods need to
be adapted to the study variable. For instance, in business surveys, the
interest variables often contain a large number of zeros. Motivated by a
mixture regression model, we propose two imputation procedures for such data
and study their statistical properties. We show that these procedures preserve
the distribution function if the imputation model is well specified. The
results of a simulation study illustrate the good performance of the proposed
methods in terms of bias and mean square error.
| stat |
Letter to the Editor | Galarza, Lachos and Bandyopadhyay (2017) have recently proposed a method of
estimating linear quantile mixed models (Geraci and Bottai, 2014) based on a
Monte Carlo EM algorithm. They assert that their procedure represents an
improvement over the numerical quadrature and non-smooth optimization approach
implemented by Geraci (2014). The objective of this note is to demonstrate that
this claim is incorrect. We also point out several inaccuracies and
shortcomings in their paper which affect other results and conclusions that can
be drawn.
| stat |
Monte Carlo Approximation of Bayes Factors via Mixing with Surrogate
Distributions | By mixing the target posterior distribution with a surrogate distribution, of
which the normalizing constant is tractable, we propose a method for estimating
the marginal likelihood using the Wang-Landau algorithm. We show that a faster
convergence of the proposed method can be achieved via the momentum
acceleration. Two implementation strategies are detailed: (i) facilitating
global jumps between the posterior and surrogate distributions via the
Multiple-try Metropolis; (ii) constructing the surrogate via the variational
approximation. When a surrogate is difficult to come by, we describe a new
jumping mechanism for general reversible jump Markov chain Monte Carlo
algorithms, which combines the Multiple-try Metropolis and a directional
sampling algorithm. We illustrate the proposed methods on several statistical
models, including the Log-Gaussian Cox process, the Bayesian Lasso, the
logistic regression, and the g-prior Bayesian variable selection.
| stat |
Causal Discovery with General Non-Linear Relationships Using Non-Linear
ICA | We consider the problem of inferring causal relationships between two or more
passively observed variables. While the problem of such causal discovery has
been extensively studied especially in the bivariate setting, the majority of
current methods assume a linear causal relationship, and the few methods which
consider non-linear dependencies usually make the assumption of additive noise.
Here, we propose a framework through which we can perform causal discovery in
the presence of general non-linear relationships. The proposed method is based
on recent progress in non-linear independent component analysis and exploits
the non-stationarity of observations in order to recover the underlying sources
or latent disturbances. We show rigorously that in the case of bivariate causal
discovery, such non-linear ICA can be used to infer the causal direction via a
series of independence tests. We further propose an alternative measure of
causal direction based on asymptotic approximations to the likelihood ratio, as
well as an extension to multivariate causal discovery. We demonstrate the
capabilities of the proposed method via a series of simulation studies and
conclude with an application to neuroimaging data.
| stat |
Learning a Generator Model from Terminal Bus Data | In this work we investigate approaches to reconstruct generator models from
measurements available at the generator terminal bus using machine learning
(ML) techniques. The goal is to develop an emulator which is trained online and
is capable of fast predictive computations. The training is illustrated on
synthetic data generated based on available open-source dynamical generator
model. Two ML techniques were developed and tested: (a) standard vector
auto-regressive (VAR) model; and (b) novel customized long short-term memory
(LSTM) deep learning model. Trade-offs in reconstruction ability between
computationally light but linear AR model and powerful but computationally
demanding LSTM model are established and analyzed.
| stat |
Prediction of Time-to-terminal Event (TTTE) in a Class of Joint Dynamic
Models | In different areas of research, multiple recurrent competing risks (RCR) are
often observed on the same observational unit. For instance, different types of
cancer relapses are observed on the same patient and several types of component
failures are observed in the same reliability system. When a terminal event
(TE) such as death is also observed on the same unit, since the RCRs are
generally informative about death, we develop joint dynamic models that
simultaneously model the RCRs and the TE. A key interest of such joint dynamic
modeling is to predict time-to-terminal event (TTTE) for new units that have
not experienced the TE by the end of monitoring period. In this paper, we
propose a simulation approach to predict TTTE which arises from a class of
joint dynamic models of RCRs and TE. The proposed approach can be applied to
problems in precision medicine and potentially many other settings. The
simulation method makes personalized predictions of TTTE and provides an
empirical predictive distribution of TTTE. Predictions of the RCR occurrences
beyond a possibly random monitoring time and leading up to the TE occurrence
are also produced. The approach is dynamic in that each simulated occurrence of
RCR increases the amount of knowledge we obtain on an observational unit which
informs the simulation of TTTE. We demonstrate the approach on a synthetic
dataset and evaluate predictive accuracy of the prediction method through
5-fold cross-validation using empirical Brier score.
| stat |
Using Undersampling with Ensemble Learning to Identify Factors
Contributing to Preterm Birth | In this paper, we propose Ensemble Learning models to identify factors
contributing to preterm birth. Our work leverages a rich dataset collected by a
NIEHS P42 Center that is trying to identify the dominant factors responsible
for the high rate of premature births in northern Puerto Rico. We investigate
analytical models addressing two major challenges present in the dataset: 1)
the significant amount of incomplete data in the dataset, and 2) class
imbalance in the dataset. First, we leverage and compare two types of missing
data imputation methods: 1) mean-based and 2) similarity-based, increasing the
completeness of this dataset. Second, we propose a feature selection and
evaluation model based on using undersampling with Ensemble Learning to address
class imbalance present in the dataset. We leverage and compare multiple
Ensemble Feature selection methods, including Complete Linear Aggregation
(CLA), Weighted Mean Aggregation (WMA), Feature Occurrence Frequency (OFA), and
Classification Accuracy Based Aggregation (CAA). To further address missing
data present in each feature, we propose two novel methods: 1) Missing Data
Rate and Accuracy Based Aggregation (MAA), and 2) Entropy and Accuracy Based
Aggregation (EAA). Both proposed models balance the degree of data variance
introduced by the missing data handling during the feature selection process
while maintaining model performance. Our results show a 42\% improvement in
sensitivity versus fallout over previous state-of-the-art methods.
| stat |
Computing Maximum Likelihood Estimates for Gaussian Graphical Models
with Macaulay2 | We introduce the package GraphicalModelsMLE for computing the maximum
likelihood estimator (MLE) of a Gaussian graphical model in the computer
algebra system Macaulay2. The package allows to compute for the class of
loopless mixed graphs. Additional functionality allows to explore the
underlying algebraic structure of the model, such as its ML degree and the
ideal of score equations.
| stat |
Time-evolving psychological processes over repeated decisions | Many psychological experiments have participants repeat a simple task. This
repetition is often necessary in order to gain the statistical precision
required to answer questions about quantitative theories of the psychological
processes underlying performance. In such experiments, time-on-task can have
important and sizable effects on performance, changing the psychological
processes under investigation in interesting ways. These changes are often
ignored, and the underlying model is treated as static. We apply modern
statistical approaches to extend a static model of decision-making to account
for changes with time-on-task. Using data from three highly-cited experiments,
we show that there are changes in performance with time-on-task, and that these
changes vary substantially over individuals - both in magnitude and direction.
Model-based analysis reveals how different cognitive processes contribute to
the observed changes. We find strong evidence in favor of a first order
autoregressive process governing the time-based evolution of individual
subjects' model parameters. The central idea of our approach can be applied
quite generally to quantitative psychological theories, beyond the model that
we investigate and the experimental data that we use.
| stat |
Markov Chain Monte Carlo Methods, a survey with some frequent
misunderstandings | In this chapter, we review some of the most standard MCMC tools used in
Bayesian computation, along with vignettes on standard misunderstandings of
these approaches taken from Q \&~A's on the forum Cross-validated answered by
the first author.
| stat |
Communication-Efficient Local Decentralized SGD Methods | Recently, the technique of local updates is a powerful tool in centralized
settings to improve communication efficiency via periodical communication. For
decentralized settings, it is still unclear how to efficiently combine local
updates and decentralized communication. In this work, we propose an algorithm
named as LD-SGD, which incorporates arbitrary update schemes that alternate
between multiple Local updates and multiple Decentralized SGDs, and provide an
analytical framework for LD-SGD. Under the framework, we present a sufficient
condition to guarantee the convergence. We show that LD-SGD converges to a
critical point for a wide range of update schemes when the objective is
non-convex and the training data are non-identically independent distributed.
Moreover, our framework brings many insights into the design of update schemes
for decentralized optimization. As examples, we specify two update schemes and
show how they help improve communication efficiency. Specifically, the first
scheme alternates the number of local and global update steps. From our
analysis, the ratio of the number of local updates to that of decentralized SGD
trades off communication and computation. The second scheme is to periodically
shrink the length of local updates. We show that the decaying strategy helps
improve communication efficiency both theoretically and empirically.
| stat |
Assessing Dynamic Effects on a Bayesian Matrix-Variate Dynamic Linear
Model: an Application to fMRI Data Analysis | In this work, we propose a modeling procedure for fMRI data analysis using a
Bayesian Matrix-Variate Dynamic Linear Model (MVDLM). With this type of model,
less complex than the more traditional temporal-spatial models, we are able to
take into account the temporal and -- at least locally -- the spatial
structures that are usually present in this type of data. Despite employing a
voxel-wise approach, every voxel in the brain is jointly modeled with its
nearest neighbors, which are defined through a euclidian metric. MVDLM's have
been widely used in applications where the interest lies in to perform
predictions and/or analysis of covariance structures among time series. In this
context, our interest is rather to assess the dynamic effects which are related
to voxel activation. In order to do so, we develop three algorithms to simulate
online-trajectories related to the state parameter and with those curves or
simulated trajectories we compute a Monte Carlo evidence for voxel activation.
Through two practical examples and two different types of assessments, we show
that our method can be viewed for the practitioners as a reliable tool for fMRI
data analysis. Despite all the examples and analysis are illustrated just for a
single subject analysis, we also describe how more general group analysis can
be implemented.
| stat |
Dynamic Bayesian Neural Networks | We define an evolving in time Bayesian neural network called a Hidden Markov
neural network. The weights of a feed-forward neural network are modelled with
the hidden states of a Hidden Markov model, whose observed process is given by
the available data. A filtering algorithm is used to learn a variational
approximation to the evolving in time posterior over the weights. Training is
pursued through a sequential version of Bayes by Backprop Blundell et al. 2015,
which is enriched with a stronger regularization technique called variational
DropConnect. The experiments test variational DropConnect on MNIST and display
the performance of Hidden Markov neural networks on time series.
| stat |
Neural Manifold Ordinary Differential Equations | To better conform to data geometry, recent deep generative modelling
techniques adapt Euclidean constructions to non-Euclidean spaces. In this
paper, we study normalizing flows on manifolds. Previous work has developed
flow models for specific cases; however, these advancements hand craft layers
on a manifold-by-manifold basis, restricting generality and inducing cumbersome
design constraints. We overcome these issues by introducing Neural Manifold
Ordinary Differential Equations, a manifold generalization of Neural ODEs,
which enables the construction of Manifold Continuous Normalizing Flows
(MCNFs). MCNFs require only local geometry (therefore generalizing to arbitrary
manifolds) and compute probabilities with continuous change of variables
(allowing for a simple and expressive flow construction). We find that
leveraging continuous manifold dynamics produces a marked improvement for both
density estimation and downstream tasks.
| stat |
Stochastic Recursive Momentum for Policy Gradient Methods | In this paper, we propose a novel algorithm named STOchastic Recursive
Momentum for Policy Gradient (STORM-PG), which operates a SARAH-type stochastic
recursive variance-reduced policy gradient in an exponential moving average
fashion. STORM-PG enjoys a provably sharp $O(1/\epsilon^3)$ sample complexity
bound for STORM-PG, matching the best-known convergence rate for policy
gradient algorithm. In the mean time, STORM-PG avoids the alternations between
large batches and small batches which persists in comparable variance-reduced
policy gradient methods, allowing considerably simpler parameter tuning.
Numerical experiments depicts the superiority of our algorithm over comparative
policy gradient algorithms.
| stat |
Night sleep duration trajectories and associated factors among preschool
children from the EDEN cohort | Objective. Sleep duration may vary inter-individually and intra-individually
over time. We aimed at both identifying night-sleep duration (NSD) trajectories
among preschoolers and studying associated factors. Methods. NSD were collected
within the French birth-cohort study EDEN at ages 2, 3 and 5-6 years through
parental questionnaires, and were used to model NSD trajectories among 1205
children. Familial socioeconomic factors, maternal sociodemographic, health and
lifestyle characteristics as well as child health, lifestyle, and sleep
characteristics at birth and/or at age 2 years were investigated in association
with NSD using multinomial logistic regressions. Results. Five distinct NSD
trajectories were identified: short (SS, <10h, 4.9%), medium-low (MLS, <11h,
47.8%), medium-high (MHS, $\approx$11h30, 37.2%), long (LS, $\ge$11h30, 4.5%)
and changing (CS, i.e. $\ge$11h30 then <11h, 5.6%) NSD trajectories.
Multivariable analyses showed in particular that, compared to the MHS
trajectory, factors associated with increased risk for belonging to SS
trajectory were male gender, first child, maternal age and working status,
night-waking, parental presence when falling asleep, television-viewing
duration and both the `Processed and fast foods' and the `Baby food' dietary
patterns at age 2 years. Factors positively associated with the CS trajectory
were maternal smoking, bottle-feeding at night and the `Processed and fast
foods' dietary pattern at age 2 years whereas child's activity and emotionality
scores at age 1 year were negatively associated. Conclusion. We identified
distinct NSD trajectories among preschoolers and associated early life factors.
Some of them may reflect less healthy lifestyle, providing cues for early
multi-behavioral prevention interventions
| stat |
Bayesian spatially varying coefficient models in the spBayes R package | This paper describes and illustrates new functionality for fitting spatially
varying coefficients models in the spBayes (version 0.4-2) R package. The new
spSVC function uses a computationally efficient Markov chain Monte Carlo
algorithm and extends current spBayes functions, that fit only space-varying
intercept regression models, to fit independent or multivariate Gaussian
process random effects for any set of columns in the regression design matrix.
Newly added OpenMP parallelization options for spSVC are discussed and
illustrated, as well as helper functions for joint and point-wise prediction
and model fit diagnostics. The utility of the proposed models is illustrated
using a PM10 analysis over central Europe.
| stat |
Analyzing second order stochasticity of neural spiking under
stimuli-bundle exposure | Conventional analysis of neuroscience data involves computing average neural
activity over a group of trials and/or a period of time. This approach may be
particularly problematic when assessing the response patterns of neurons to
more than one simultaneously presented stimulus. In such cases, the brain must
represent each individual component of the stimuli bundle, but
trial-and-time-pooled averaging methods are fundamentally unequipped to address
the means by which multi-item representation occurs. We introduce and
investigate a novel statistical analysis framework that relates the firing
pattern of a single cell, exposed to a stimuli bundle, to the ensemble of its
firing patterns under each constituent stimulus. Existing statistical tools
focus on what may be called "first order stochasticity" in trial-to-trial
variation in the form of unstructured noise around a fixed firing rate curve
associated with a given stimulus. Our analysis is based upon the theoretical
premise that exposure to a stimuli bundle induces additional stochasticity in
the cell's response pattern, in the form of a stochastically varying
recombination of its single stimulus firing rate curves. We discuss challenges
to statistical estimation of such "second order stochasticity" and address them
with a novel dynamic admixture Poisson process (DAPP) model. DAPP is a
hierarchical point process model that decomposes second order stochasticity
into a Gaussian stochastic process and a random vector of interpretable
features, and, facilitates borrowing of information on the latter across
repeated trials through latent clustering. We present empirical evidence of the
utility of the DAPP analysis with synthetic and real neural recordings.
| stat |
Randomized Hamiltonian Monte Carlo as Scaling Limit of the Bouncy
Particle Sampler and Dimension-Free Convergence Rates | The Bouncy Particle Sampler is a Markov chain Monte Carlo method based on a
nonreversible piecewise deterministic Markov process. In this scheme, a
particle explores the state space of interest by evolving according to a linear
dynamics which is altered by bouncing on the hyperplane tangent to the gradient
of the negative log-target density at the arrival times of an inhomogeneous
Poisson Process (PP) and by randomly perturbing its velocity at the arrival
times of an homogeneous PP. Under regularity conditions, we show here that the
process corresponding to the first component of the particle and its
corresponding velocity converges weakly towards a Randomized Hamiltonian Monte
Carlo (RHMC) process as the dimension of the ambient space goes to infinity.
RHMC is another piecewise deterministic non-reversible Markov process where a
Hamiltonian dynamics is altered at the arrival times of a homogeneous PP by
randomly perturbing the momentum component. We then establish dimension-free
convergence rates for RHMC for strongly log-concave targets with bounded
Hessians using coupling ideas and hypocoercivity techniques.
| stat |
Towards Robust Evaluations of Continual Learning | Experiments used in current continual learning research do not faithfully
assess fundamental challenges of learning continually. Instead of assessing
performance on challenging and representative experiment designs, recent
research has focused on increased dataset difficulty, while still using flawed
experiment set-ups. We examine standard evaluations and show why these
evaluations make some continual learning approaches look better than they are.
We introduce desiderata for continual learning evaluations and explain why
their absence creates misleading comparisons. Based on our desiderata we then
propose new experiment designs which we demonstrate with various continual
learning approaches and datasets. Our analysis calls for a reprioritization of
research effort by the community.
| stat |
The Generalized Complex Kernel Least-Mean-Square Algorithm | We propose a novel adaptive kernel based regression method for complex-valued
signals: the generalized complex-valued kernel least-mean-square (gCKLMS). We
borrow from the new results on widely linear reproducing kernel Hilbert space
(WL-RKHS) for nonlinear regression and complex-valued signals, recently
proposed by the authors. This paper shows that in the adaptive version of the
kernel regression for complex-valued signals we need to include another kernel
term, the so-called pseudo-kernel. This new solution is endowed with better
representation capabilities in complex-valued fields, since it can efficiently
decouple the learning of the real and the imaginary part. Also, we review
previous realizations of the complex KLMS algorithm and its augmented version
to prove that they can be rewritten as particular cases of the gCKLMS.
Furthermore, important conclusions on the kernels design are drawn that help to
greatly improve the convergence of the algorithms. In the experiments, we
revisit the nonlinear channel equalization problem to highlight the better
convergence of the gCKLMS compared to previous solutions. Also, the flexibility
of the proposed generalized approach is tested in a second experiment with
non-independent real and imaginary parts. The results illustrate the
significant performance improvements of the gCKLMS approach when the
complex-valued signals have different properties for the real and imaginary
parts.
| stat |
Modelling basketball players' performance and interactions between
teammates with a regime switching approach | Basketball players' performance measurement is of critical importance for a
broad spectrum of decisions related to training and game strategy. Despite this
recognized central role, the main part of the studies on this topic focus on
performance level measurement, neglecting other important characteristics, such
as variability. In this paper, shooting performance variability is modeled with
a Markov Switching dynamic, assuming the existence of two alternating
performance regimes. Then, the relationships between each player's variability
and the lineup composition is modeled as an ARIMA process with covariates and
described with network analysis tools, in order to extrapolate positive and
negative interactions between teammates
| stat |
A Double Penalty Model for Interpretability | Modern statistical learning techniques have often emphasized prediction
performance over interpretability, giving rise to "black box" models that may
be difficult to understand, and to generalize to other settings. We
conceptually divide a prediction model into interpretable and non-interpretable
portions, as a means to produce models that are highly interpretable with
little loss in performance. Implementation of the model is achieved by
considering separability of the interpretable and non-interpretable portions,
along with a doubly penalized procedure for model fitting. We specify
conditions under which convergence of model estimation can be achieved via
cyclic coordinate ascent, and the consistency of model estimation holds. We
apply the methods to datasets for microbiome host trait prediction and a
diabetes trait, and discuss practical tradeoff diagnostics to select models
with high interpretability.
| stat |
Network topology change-point detection from graph signals with prior
spectral signatures | We consider the problem of sequential graph topology change-point detection
from graph signals. We assume that signals on the nodes of the graph are
regularized by the underlying graph structure via a graph filtering model,
which we then leverage to distill the graph topology change-point detection
problem to a subspace detection problem. We demonstrate how prior information
on the spectral signature of the post-change graph can be incorporated to
implicitly denoise the observed sequential data, thus leading to a natural
CUSUM-based algorithm for change-point detection. Numerical experiments
illustrate the performance of our proposed approach, particularly underscoring
the benefits of (potentially noisy) prior information.
| stat |
A Kernel to Exploit Informative Missingness in Multivariate Time Series
from EHRs | A large fraction of the electronic health records (EHRs) consists of clinical
measurements collected over time, such as lab tests and vital signs, which
provide important information about a patient's health status. These sequences
of clinical measurements are naturally represented as time series,
characterized by multiple variables and large amounts of missing data, which
complicate the analysis. In this work, we propose a novel kernel which is
capable of exploiting both the information from the observed values as well the
information hidden in the missing patterns in multivariate time series (MTS)
originating e.g. from EHRs. The kernel, called TCK$_{IM}$, is designed using an
ensemble learning strategy in which the base models are novel mixed mode
Bayesian mixture models which can effectively exploit informative missingness
without having to resort to imputation methods. Moreover, the ensemble approach
ensures robustness to hyperparameters and therefore TCK$_{IM}$ is particularly
well suited if there is a lack of labels - a known challenge in medical
applications. Experiments on three real-world clinical datasets demonstrate the
effectiveness of the proposed kernel.
| stat |
Approximate Bayesian Computation in controlled branching processes: the
role of summary statistics | Controlled branching processes are stochastic growth population models in
which the number of individuals with reproductive capacity in each generation
is controlled by a random control function. The purpose of this work is to
examine the Approximate Bayesian Computation (ABC) methods and to propose
appropriate summary statistics for them in the context of these processes. This
methodology enables to approximate the posterior distribution of the parameters
of interest satisfactorily without explicit likelihood calculations and under a
minimal set of assumptions. In particular, the tolerance rejection algorithm,
the sequential Monte Carlo ABC algorithm, and a post-sampling correction method
based on local-linear regression are provided. The accuracy of the proposed
methods are illustrated and compared with a "likelihood free" Markov chain
Monte Carlo technique by the way of a simulated example developed with the
statistical software R.
| stat |
Automatically matching topographical measurements of cartridge cases
using a record linkage framework | Firing a gun leaves marks on cartridge cases which purportedly uniquely
identify the gun. Firearms examiners typically use a visual examination to
evaluate if two cartridge cases were fired from the same gun, and this is a
subjective process that has come under scrutiny. Matching can be done in a more
reproducible manner using automated algorithms. In this paper, we develop
methodology to compare topographical measurements of cartridge cases. We
demonstrate the use of a record linkage framework in this context. We compare
performance using topographical measurements to older reflectance microscopy
images, investigating the extent to which the former produce more accurate
comparisons. Using a diverse collection of images of over 1,100 cartridge
cases, we find that overall performance is generally improved using
topographical data. Some subsets of the data achieve almost perfect predictive
performance in terms of precision and recall, while some produce extremely poor
performance. Further work needs to be done to assess if examiners face similar
difficulties on certain gun and ammunition combinations. For automatic methods,
a fuller investigation into their fairness and robustness is necessary before
they can be deployed in practice.
| stat |
An Approximate Quasi-Likelihood Approach for Error-Prone Failure Time
Outcomes and Exposures | Measurement error arises commonly in clinical research settings that rely on
data from electronic health records or large observational cohorts. In
particular, self-reported outcomes are typical in cohort studies for chronic
diseases such as diabetes in order to avoid the burden of expensive diagnostic
tests. Dietary intake, which is also commonly collected by self-report and
subject to measurement error, is a major factor linked to diabetes and other
chronic diseases. These errors can bias exposure-disease associations that
ultimately can mislead clinical decision-making. We have extended an existing
semiparametric likelihood-based method for handling error-prone, discrete
failure time outcomes to also address covariate error. We conduct an extensive
numerical study to compare the proposed method to the naive approach that
ignores measurement error in terms of bias and efficiency in the estimation of
the regression parameter of interest. In all settings considered, the proposed
method showed minimal bias and maintained coverage probability, thus
outperforming the naive analysis which showed extreme bias and low coverage.
This method is applied to data from the Women's Health Initiative to assess the
association between energy and protein intake and the risk of incident diabetes
mellitus. Our results show that correcting for errors in both the self-reported
outcome and dietary exposures leads to considerably different hazard ratio
estimates than those from analyses that ignore measurement error, which
demonstrates the importance of correcting for both outcome and covariate error.
Computational details and R code for implementing the proposed method are
presented in Section S1 of the Supplementary Materials.
| stat |
Modeling Random Directions in 2D Simplex Data | We propose models and algorithms for learning about random directions in
two-dimensional simplex data, and apply our methods to the study of income
level proportions and their changes over time in a geostatistical area. There
are several notable challenges in the analysis of simplex-valued data: the
measurements must respect the simplex constraint and the changes exhibit
spatiotemporal smoothness while allowing for possible heterogeneous behaviors.
To that end, we propose Bayesian models that rely on and expand upon building
blocks in circular and spatial statistics by exploiting suitable transformation
based on the polar coordinates for circular data. Our models also account for
spatial correlation across locations in the simplex and the heterogeneous
patterns via mixture modeling. We describe some properties of the models and
model fitting via MCMC techniques. Our models and methods are illustrated via a
thorough simulation study, and applied to an analysis of movements and trends
of income categories using the Home Mortgage Disclosure Act data.
| stat |
Hierarchical Clustering for Smart Meter Electricity Loads based on
Quantile Autocovariances | In order to improve the efficiency and sustainability of electricity systems,
most countries worldwide are deploying advanced metering infrastructures, and
in particular household smart meters, in the residential sector. This
technology is able to record electricity load time series at a very high
frequency rates, information that can be exploited to develop new clustering
models to group individual households by similar consumptions patterns. To this
end, in this work we propose three hierarchical clustering methodologies that
allow capturing different characteristics of the time series. These are based
on a set of "dissimilarity" measures computed over different features: quantile
auto-covariances, and simple and partial autocorrelations. The main advantage
is that they allow summarizing each time series in a few representative
features so that they are computationally efficient, robust against outliers,
easy to automatize, and scalable to hundreds of thousands of smart meters
series. We evaluate the performance of each clustering model in a real-world
smart meter dataset with thousands of half-hourly time series. The results show
how the obtained clusters identify relevant consumption behaviors of households
and capture part of their geo-demographic segmentation. Moreover, we apply a
supervised classification procedure to explore which features are more relevant
to define each cluster.
| stat |
Predicting pregnancy using large-scale data from a women's health
tracking mobile application | Predicting pregnancy has been a fundamental problem in women's health for
more than 50 years. Previous datasets have been collected via carefully curated
medical studies, but the recent growth of women's health tracking mobile apps
offers potential for reaching a much broader population. However, the
feasibility of predicting pregnancy from mobile health tracking data is
unclear. Here we develop four models -- a logistic regression model, and 3 LSTM
models -- to predict a woman's probability of becoming pregnant using data from
a women's health tracking app, Clue by BioWink GmbH. Evaluating our models on a
dataset of 79 million logs from 65,276 women with ground truth pregnancy test
data, we show that our predicted pregnancy probabilities meaningfully stratify
women: women in the top 10% of predicted probabilities have a 89% chance of
becoming pregnant over 6 menstrual cycles, as compared to a 27% chance for
women in the bottom 10%. We develop a technique for extracting interpretable
time trends from our deep learning models, and show these trends are consistent
with previous fertility research. Our findings illustrate the potential that
women's health tracking data offers for predicting pregnancy on a broader
population; we conclude by discussing the steps needed to fulfill this
potential.
| stat |
A Bayesian Application in Judicial Decisions | This paper presents a new tool to support the decision concerning moral
damage indemnity values of the judiciary of Rio Grande do Sul, Brazil. A
Bayesian approach is given, in order to allow the assignment of the
magistrate's opinion about such indemnity amounts, based on historical values.
The solution is delivered in free software using public data, in order to
permit future audits.
| stat |
Assessing the Global Wind Atlas and local measurements for bias
correction of wind power generation simulated from MERRA-2 in Brazil | NASAs MERRA-2 reanalysis is a widely used dataset in renewable energy
resource modelling. The Global Wind Atlas (GWA) has been used to bias-correct
MERRA-2 data before. There is, however, a lack of an analysis of the
performance of MERRA-2 with bias correction from GWA on different spatial
levels - and for regions outside of Europe, China or the United States. This
study therefore evaluates different methods for wind power simulation on four
spatial resolution levels from wind park to national level in Brazil. In
particular, spatial interpolation methods and spatial as well as spatiotemporal
wind speed bias correction using local wind speed measurements and mean wind
speeds from the GWA are assessed. By validating the resulting timeseries
against observed generation it is assessed at which spatial levels the
different methods improve results - and whether global information derived from
the GWA can compete with locally measured wind speed data as a source of bias
correction. Results show that (i) bias correction with the GWA improves results
on state, sub-system, and national-level, but not on wind park level, that (ii)
the GWA improves results comparably to local measurements, and that (iii)
complex spatial interpolation methods do not contribute in improving quality of
the simulation.
| stat |
From Predictions to Prescriptions in Multistage Optimization Problems | In this paper, we introduce a framework for solving finite-horizon multistage
optimization problems under uncertainty in the presence of auxiliary data. We
assume the joint distribution of the uncertain quantities is unknown, but noisy
observations, along with observations of auxiliary covariates, are available.
We utilize effective predictive methods from machine learning (ML), including
$k$-nearest neighbors regression ($k$NN), classification and regression trees
(CART), and random forests (RF), to develop specific methods that are
applicable to a wide variety of problems. We demonstrate that our solution
methods are asymptotically optimal under mild conditions. Additionally, we
establish finite sample guarantees for the optimality of our method with $k$NN
weight functions. Finally, we demonstrate the practicality of our approach with
computational examples. We see a significant decrease in cost by taking into
account the auxiliary data in the multistage setting.
| stat |
Implementing Approximate Bayesian Inference using Adaptive Quadrature:
the aghq Package | I introduce the aghq package for implementing approximate Bayesian inference
using Adaptive Gauss-Hermite Quadrature. I describe the method and software,
and illustrate its use in several challenging low- and high-dimensional
examples considered by Bilodeau et. al. (2021) and others. Specifically, I show
how the aghq package is used as a basis for implementing more complicated
approximate Bayesian inference methods with two difficult applications in
non-Gaussian geostatistical modelling. I also show how the package can be used
to make fully Bayesian inferences in models currently fit using frequentist
inference by leveraging code from other packages, with an application to a
zero-inflated, overdispersed Poisson regression fit using the glmmTMB package.
| stat |
Graphon estimation via nearest neighbor algorithm and 2D fused lasso
denoising | We propose a class of methods for graphon estimation based on exploiting
connections with nonparametric regression. The idea is to construct an ordering
of the nodes in the network, similar in spirit to Chan and Airoldi (2014).
However, rather than only considering orderings based on the empirical degree
as in Chan and Airoldi (2014), we use the nearest neighbor algorithm which is
an approximating solution to the traveling salesman problem. This in turn can
handle general distances $\hat{d}$ between the nodes, something that allows us
to incorporate rich information of the network. Once an ordering is
constructed, we formulate a 2D grid graph denoising problem that we solve
through fused lasso regularization. For particular choices of the metric
$\hat{d}$, we show that the corresponding two-step estimator can attain
competitive rates when the true model is the stochastic block model, and when
the underlying graphon is piecewise H\"{o}lder or it has bounded variation.
| stat |
Forecasting day-ahead electricity prices: A review of state-of-the-art
algorithms, best practices and an open-access benchmark | While the field of electricity price forecasting has benefited from plenty of
contributions in the last two decades, it arguably lacks a rigorous approach to
evaluating new predictive algorithms. The latter are often compared using
unique, not publicly available datasets and across too short and limited to one
market test samples. The proposed new methods are rarely benchmarked against
well established and well performing simpler models, the accuracy metrics are
sometimes inadequate and testing the significance of differences in predictive
performance is seldom conducted. Consequently, it is not clear which methods
perform well nor what are the best practices when forecasting electricity
prices. In this paper, we tackle these issues by performing a literature survey
of state-of-the-art models, comparing state-of-the-art statistical and deep
learning methods across multiple years and markets, and by putting forward a
set of best practices. In addition, we make available the considered datasets,
forecasts of the state-of-the-art models, and a specifically designed python
toolbox, so that new algorithms can be rigorously evaluated in future studies.
| stat |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.