title
stringlengths 6
244
| abstract
stringlengths 19
5.09k
| label
stringclasses 10
values |
|---|---|---|
Optimal Bayesian Smoothing of Functional Observations over a Large Graph | In modern contexts, some types of data are observed in high-resolution,
essentially continuously in time. Such data units are best described as taking
values in a space of functions. Subject units carrying the observations may
have intrinsic relations among themselves, and are best described by the nodes
of a large graph. It is often sensible to think that the underlying signals in
these functional observations vary smoothly over the graph, in that neighboring
nodes have similar underlying signals. This qualitative information allows
borrowing of strength over neighboring nodes and consequently leads to more
accurate inference. In this paper, we consider a model with Gaussian functional
observations and adopt a Bayesian approach to smoothing over the nodes of the
graph. We characterize the minimax rate of estimation in terms of the
regularity of the signals and their variation across nodes quantified in terms
of the graph Laplacian. We show that an appropriate prior constructed from the
graph Laplacian can attain the minimax bound, while using a mixture prior, the
minimax rate up to a logarithmic factor can be attained simultaneously for all
possible values of functional and graphical smoothness. We also show that in
the fixed smoothness setting, an optimal sized credible region has arbitrarily
high frequentist coverage. A simulation experiment demonstrates that the method
performs better than potential competing methods like the random forest. The
method is also applied to a dataset on daily temperatures measured at several
weather stations in the US state of North Carolina.
| stat |
Max-Linear Regression by Scalable and Guaranteed Convex Programming | We consider the multivariate max-linear regression problem where the model
parameters
$\boldsymbol{\beta}_{1},\dotsc,\boldsymbol{\beta}_{k}\in\mathbb{R}^{p}$ need to
be estimated from $n$ independent samples of the (noisy) observations $y =
\max_{1\leq j \leq k} \boldsymbol{\beta}_{j}^{\mathsf{T}} \boldsymbol{x} +
\mathrm{noise}$. The max-linear model vastly generalizes the conventional
linear model, and it can approximate any convex function to an arbitrary
accuracy when the number of linear models $k$ is large enough. However, the
inherent nonlinearity of the max-linear model renders the estimation of the
regression parameters computationally challenging. Particularly, no estimator
based on convex programming is known in the literature. We formulate and
analyze a scalable convex program as the estimator for the max-linear
regression problem. Under the standard Gaussian observation setting, we present
a non-asymptotic performance guarantee showing that the convex program recovers
the parameters with high probability. When the $k$ linear components are
equally likely to achieve the maximum, our result shows that a sufficient
number of observations scales as $k^{2}p$ up to a logarithmic factor. This
significantly improves on the analogous prior result based on alternating
minimization (Ghosh et al., 2019). Finally, through a set of Monte Carlo
simulations, we illustrate that our theoretical result is consistent with
empirical behavior, and the convex estimator for max-linear regression is as
competitive as the alternating minimization algorithm in practice.
| stat |
Anomaly Detection in Trajectory Data with Normalizing Flows | The task of detecting anomalous data patterns is as important in practical
applications as challenging. In the context of spatial data, recognition of
unexpected trajectories brings additional difficulties, such as high
dimensionality and varying pattern lengths. We aim to tackle such a problem
from a probability density estimation point of view, since it provides an
unsupervised procedure to identify out of distribution samples. More
specifically, we pursue an approach based on normalizing flows, a recent
framework that enables complex density estimation from data with neural
networks. Our proposal computes exact model likelihood values, an important
feature of normalizing flows, for each segment of the trajectory. Then, we
aggregate the segments' likelihoods into a single coherent trajectory anomaly
score. Such a strategy enables handling possibly large sequences with different
lengths. We evaluate our methodology, named aggregated anomaly detection with
normalizing flows (GRADINGS), using real world trajectory data and compare it
with more traditional anomaly detection techniques. The promising results
obtained in the performed computational experiments indicate the feasibility of
the GRADINGS, specially the variant that considers autoregressive normalizing
flows.
| stat |
Teasing out the overall survival benefit with adjustment for treatment
switching to other therapies | In oncology clinical trials, characterizing the long-term overall survival
(OS) benefit for an experimental drug or treatment regimen (experimental group)
is often unobservable if some patients in the control group switch to drugs in
the experimental group and/or other cancer treatments after disease
progression. A key question often raised by payers and reimbursement agencies
is how to estimate the true benefit of the experimental drug group on overall
survival that would have been estimated if there were no treatment switches.
Several commonly used statistical methods are available to estimate overall
survival benefit while adjusting for treatment switching, ranging from naive
exclusion or censoring approaches to more advanced methods including inverse
probability of censoring weighting (IPCW), iterative parameter estimation (IPE)
algorithm or rank-preserving structural failure time models (RPSFTM). However,
many clinical trials now have patients switching to different treatment
regimens other than the test drugs, and the existing methods cannot handle more
complicated scenarios. To address this challenge, we propose two additional
methods: stratified RPSFTM and random-forest-based prediction. A simulation
study is conducted to assess the properties of the existing methods along with
the two newly proposed approaches.
| stat |
Efficient Learning for Clustering and Optimizing Context-Dependent
Designs | We consider a simulation optimization problem for a context-dependent
decision-making. A Gaussian mixture model is proposed to capture the
performance clustering phenomena of context-dependent designs. Under a Bayesian
framework, we develop a dynamic sampling policy to efficiently learn both the
global information of each cluster and local information of each design for
selecting the best designs in all contexts. The proposed sampling policy is
proved to be consistent and achieve the asymptotically optimal sampling ratio.
Numerical experiments show that the proposed sampling policy significantly
improves the efficiency in context-dependent simulation optimization.
| stat |
Nonstationary, Nonparametric, Nonseparable Bayesian Spatio-Temporal
Modeling Using Kernel Convolution of Order Based Dependent Dirichlet Process | In this article, using kernel convolution of order based dependent Dirichlet
process (Griffin and Steel (2006)) we construct a nonstationary, nonseparable,
nonparametric space-time process, which, as we show, satisfies desirable
properties, and includes the stationary, separable, parametric processes as
special cases. We also investigate the smoothness properties of our proposed
model. Since our model entails an infinite random series, for Bayesian model
fitting purpose we must either truncate the series or more appropriately
consider a random number of summands, which renders the model dimension a
random variable. We attack the variable dimensionality problem using
Transdimensional Transformation based Markov Chain Monte Carlo introduced by
Das and Bhattacharya (2019b), which can update all the variables and also
change dimensions in a single block using essentially a single random variable
drawn from some arbitrary density defined on a relevant support. For the sake
of completeness we also address the problem of truncating the infinite series
by providing a uniform bound on the error incurred by truncating the infinite
series.
We illustrate the effectiveness of our model and methodologies on a simulated
data set and demonstrate that our approach significantly outperforms that of
Fuentes and Reich (2013) which is based on principles somewhat similar to ours.
We also fit two real, spatial and spatio-temporal datasets with our approach
and obtain quite encouraging results in both the cases.
| stat |
Principal Ellipsoid Analysis (PEA): Efficient non-linear dimension
reduction & clustering | Even with the rise in popularity of over-parameterized models, simple
dimensionality reduction and clustering methods, such as PCA and k-means, are
still routinely used in an amazing variety of settings. A primary reason is the
combination of simplicity, interpretability and computational efficiency. The
focus of this article is on improving upon PCA and k-means, by allowing
non-linear relations in the data and more flexible cluster shapes, without
sacrificing the key advantages. The key contribution is a new framework for
Principal Elliptical Analysis (PEA), defining a simple and computationally
efficient alternative to PCA that fits the best elliptical approximation
through the data. We provide theoretical guarantees on the proposed PEA
algorithm using Vapnik-Chervonenkis (VC) theory to show strong consistency and
uniform concentration bounds. Toy experiments illustrate the performance of
PEA, and the ability to adapt to non-linear structure and complex cluster
shapes. In a rich variety of real data clustering applications, PEA is shown to
do as well as k-means for simple datasets, while dramatically improving
performance in more complex settings.
| stat |
Accelerating Nonconvex Learning via Replica Exchange Langevin Diffusion | Langevin diffusion is a powerful method for nonconvex optimization, which
enables the escape from local minima by injecting noise into the gradient. In
particular, the temperature parameter controlling the noise level gives rise to
a tradeoff between ``global exploration'' and ``local exploitation'', which
correspond to high and low temperatures. To attain the advantages of both
regimes, we propose to use replica exchange, which swaps between two Langevin
diffusions with different temperatures. We theoretically analyze the
acceleration effect of replica exchange from two perspectives: (i) the
convergence in \chi^2-divergence, and (ii) the large deviation principle. Such
an acceleration effect allows us to faster approach the global minima.
Furthermore, by discretizing the replica exchange Langevin diffusion, we obtain
a discrete-time algorithm. For such an algorithm, we quantify its
discretization error in theory and demonstrate its acceleration effect in
practice.
| stat |
Efficient Bayesian Experimental Design for Implicit Models | Bayesian experimental design involves the optimal allocation of resources in
an experiment, with the aim of optimising cost and performance. For implicit
models, where the likelihood is intractable but sampling from the model is
possible, this task is particularly difficult and therefore largely unexplored.
This is mainly due to technical difficulties associated with approximating
posterior distributions and utility functions. We devise a novel experimental
design framework for implicit models that improves upon previous work in two
ways. First, we use the mutual information between parameters and data as the
utility function, which has previously not been feasible. We achieve this by
utilising Likelihood-Free Inference by Ratio Estimation (LFIRE) to approximate
posterior distributions, instead of the traditional approximate Bayesian
computation or synthetic likelihood methods. Secondly, we use Bayesian
optimisation in order to solve the optimal design problem, as opposed to the
typically used grid search or sampling-based methods. We find that this
increases efficiency and allows us to consider higher design dimensions.
| stat |
Estimation and inference for area-wise spatial income distributions from
grouped data | Estimating income distributions plays an important role in the measurement of
inequality and poverty over space. The existing literature on income
distributions predominantly focuses on estimating an income distribution for a
country or a region separately and the simultaneous estimation of multiple
income distributions has not been discussed in spite of its practical
importance. In this work, we develop an effective method for the simultaneous
estimation and inference for area-wise spatial income distributions taking
account of geographical information from grouped data. Based on the multinomial
likelihood function for grouped data, we propose a spatial state-space model
for area-wise parameters of parametric income distributions. We provide an
efficient Bayesian approach to estimation and inference for area-wise latent
parameters, which enables us to compute area-wise summary measures of income
distributions such as mean incomes and Gini indices, not only for sampled areas
but also for areas without any samples thanks to the latent spatial state-space
structure. The proposed method is demonstrated using the Japanese
municipality-wise grouped income data. The simulation studies show the
superiority of the proposed method to a crude conventional approach which
estimates the income distributions separately.
| stat |
Additive Adversarial Learning for Unbiased Authentication | Authentication is a task aiming to confirm the truth between data instances
and personal identities. Typical authentication applications include face
recognition, person re-identification, authentication based on mobile devices
and so on. The recently-emerging data-driven authentication process may
encounter undesired biases, i.e., the models are often trained in one domain
(e.g., for people wearing spring outfits) while required to apply in other
domains (e.g., they change the clothes to summer outfits). To address this
issue, we propose a novel two-stage method that disentangles the class/identity
from domain-differences, and we consider multiple types of domain-difference.
In the first stage, we learn disentangled representations by a one-versus-rest
disentangle learning (OVRDL) mechanism. In the second stage, we improve the
disentanglement by an additive adversarial learning (AAL) mechanism. Moreover,
we discuss the necessity to avoid a learning dilemma due to disentangling
causally related types of domain-difference. Comprehensive evaluation results
demonstrate the effectiveness and superiority of the proposed method.
| stat |
What Drives Inflation and How: Evidence from Additive Mixed Models
Selected by cAIC | We analyze which forces explain inflation and how in a large panel of 124
countries from 1997 to 2015. Models motivated by economic theory are compared
to an approach based on model-based boosting and non-linearities are explicitly
considered. We provide compelling evidence that the interaction of energy price
and energy rents stand out among 40 explanatory variables. The output gap and
globalization are also relevant drivers of inflation. Credit and money growth,
a country's inflation history and demographic changes are comparably less
important while central bank related variables as well as political variables
turn out to have the least empirical relevance. In a subset of countries public
debt denomination and exchange rate arrangements also play a noteworthy role in
the inflation process. By contrast, other public-debt variables and an
inflation targeting regime have weaker explanatory power. Finally, there is
clear evidence of structural breaks in the effects since the financial crisis.
| stat |
Elastic $k$-means clustering of functional data for posterior
exploration, with an application to inference on acute respiratory infection
dynamics | We propose a new method for clustering of functional data using a $k$-means
framework. We work within the elastic functional data analysis framework, which
allows for decomposition of the overall variation in functional data into
amplitude and phase components. We use the amplitude component to partition
functions into shape clusters using an automated approach. To select an
appropriate number of clusters, we additionally propose a novel Bayesian
Information Criterion defined using a mixture model on principal components
estimated using functional Principal Component Analysis. The proposed method is
motivated by the problem of posterior exploration, wherein samples obtained
from Markov chain Monte Carlo algorithms are naturally represented as
functions. We evaluate our approach using a simulated dataset, and apply it to
a study of acute respiratory infection dynamics in San Luis Potos\'{i}, Mexico.
| stat |
Universal consistency of Wasserstein $k$-NN classifier | The Wasserstein distance provides a notion of dissimilarities between
probability measures, which has recent applications in learning of structured
data with varying size such as images and text documents. In this work, we
analyze the $k$-nearest neighbor classifier ($k$-NN) under the Wasserstein
distance and establish the universal consistency on families of distributions.
Using previous known results on the consistency of the $k$-NN classifier on
infinite dimensional metric spaces, it suffices to show that the families is a
countable union of finite dimensional components. As a result, we are able to
prove universal consistency of $k$-NN on spaces of finitely supported measures,
the space of finite wavelet series and the spaces of Gaussian measures with
commuting covariance matrices.
| stat |
A pseudo-marginal sequential Monte Carlo online smoothing algorithm | We consider online computation of expectations of additive state functionals
under general path probability measures proportional to products of
unnormalised transition densities. These transition densities are assumed to be
intractable but possible to estimate, with or without bias. Using
pseudo-marginalisation techniques we are able to extend the particle-based,
rapid incremental smoother (PaRIS) algorithm proposed in [J.Olsson and
J.Westerborn. Efficient particle-based online smoothing in general hidden
Markov models: The PaRIS algorithm. Bernoulli, 23(3):1951--1996, 2017] to this
setting. The resulting algorithm, which has a linear complexity in the number
of particles and constant memory requirements, applies to a wide range of
challenging path-space Monte Carlo problems, including smoothing in partially
observed diffusion processes and models with intractable likelihood. The
algorithm is furnished with several theoretical results, including a central
limit theorem, establishing its convergence and numerical stability. Moreover,
under strong mixing assumptions we establish a novel $O(n \varepsilon)$ bound
on the asymptotic bias of the algorithm, where $n$ is the path length and
$\varepsilon$ controls the bias of the density estimators.
| stat |
Generalized Records for Functional Time Series with Application to Unit
Root Tests | A generalization of the definition of records to functional data is proposed.
The definition is based on ranking curves using a notion of functional depth.
This approach allows us to study the curves of the number of records over time.
We focus on functional time series and apply ideas from univariate time series
to demonstrate the asymptotic distribution describing the number of records. A
unit root test is proposed as an application of functional record theory.
Through a Monte Carlo study, different scenarios of functional processes are
simulated to evaluate the performance of the unit root test. The generalized
record definition is applied on two different datasets: Annual mortality rates
in France and daily curves of wind speed at Yanbu, Saudi Arabia. The record
curves are identified and the underlying functional process is studied based on
the number of record curves observed.
| stat |
Loss-Based Variational Bayes Prediction | We propose a new method for Bayesian prediction that caters for models with a
large number of parameters and is robust to model misspecification. Given a
class of high-dimensional (but parametric) predictive models, this new approach
constructs a posterior predictive using a variational approximation to a
loss-based, or Gibbs, posterior that is directly focused on predictive
accuracy. The theoretical behavior of the new prediction approach is analyzed
and a form of optimality demonstrated. Applications to both simulated and
empirical data using high-dimensional Bayesian neural network and
autoregressive mixture models demonstrate that the approach provides more
accurate results than various alternatives, including misspecified
likelihood-based predictions.
| stat |
The vote Package: Single Transferable Vote and Other Electoral Systems
in R | We describe the vote package in R, which implements the plurality (or
first-past-the-post), two-round runoff, score, approval and single transferable
vote (STV) electoral systems, as well as methods for selecting the Condorcet
winner and loser. We emphasize the STV system, which we have found to work well
in practice for multi-winner elections with small electorates, such as
committee and council elections, and the selection of multiple job candidates.
For single-winner elections, the STV is also called instant runoff voting
(IRV), ranked choice voting (RCV), or the alternative vote (AV) system. The
package also implements the STV system with equal preferences, for the first
time in a software package, to our knowledge. It also implements a new variant
of STV, in which a minimum number of candidates from a specified group are
required to be elected. We illustrate the package with several real examples.
| stat |
Statistical significance revisited | Statistical significance measures the reliability of a result obtained from a
random experiment. We investigate the number of repetitions needed for a
statistical result to have a certain significance. In the first step, we
consider binomially distributed variables in the example of medication testing
with fixed placebo efficacy, asking how many experiments are needed in order to
achieve a significance of 95 %. In the next step, we take the probability
distribution of the placebo efficacy into account, which to the best of our
knowledge has not been done so far. Depending on the specifics, we show that in
order to obtain identical significance, it may be necessary to perform twice as
many experiments than in a setting where the placebo distribution is neglected.
We proceed by considering more general probability distributions and close with
comments on some erroneous assumptions on probability distributions which lead,
for instance, to a trivial explanation of the fat tail.
| stat |
Data Driven Robust Estimation Methods for Fixed Effects Panel Data
Models | The panel data regression models have gained increasing attention in
different areas of research including but not limited to econometrics,
environmental sciences, epidemiology, behavioral and social sciences. However,
the presence of outlying observations in panel data may often lead to biased
and inefficient estimates of the model parameters resulting in unreliable
inferences when the least squares (LS) method is applied. We propose extensions
of the M-estimation approach with a data-driven selection of tuning parameters
to achieve desirable level of robustness against outliers without loss of
estimation efficiency. The consistency and asymptotic normality of the proposed
estimators have also been proved under some mild regularity conditions. The
finite sample properties of the existing and proposed robust estimators have
been examined through an extensive simulation study and an application to
macroeconomic data. Our findings reveal that the proposed methods often
exhibits improved estimation and prediction performances in the presence of
outliers and are consistent with the traditional LS method when there is no
contamination.
| stat |
Bayesian estimates of transmission line outage rates that consider line
dependencies | Transmission line outage rates are fundamental to power system reliability
analysis. Line outages are infrequent, occurring only about once a year, so
outage data are limited. We propose a Bayesian hierarchical model that
leverages line dependencies to better estimate outage rates of individual
transmission lines from limited outage data. The Bayesian estimates have a
lower standard deviation than estimating the outage rates simply by dividing
the number of outages by the number of years of data, especially when the
number of outages is small. The Bayesian model produces more accurate
individual line outage rates, as well as estimates of the uncertainty of these
rates. Better estimates of line outage rates can improve system risk
assessment, outage prediction, and maintenance scheduling.
| stat |
The conformable fractional grey system model | The fractional order grey models (FGM) have appealed considerable interest of
research in recent years due to its higher effectiveness and flexibility than
the conventional grey models and other prediction models. However, the
definitions of the fractional order accumulation (FOA) and difference (FOD) is
computationally complex, which leads to difficulties for the theoretical
analysis and applications. In this paper, the new definition of the FOA are
proposed based on the definitions of Conformable Fractional Derivative, which
is called the Conformable Fractional Accumulation (CFA), along with its inverse
operation, the Conformable Fractional Difference (CFD). Then the new
Conformable Fractional Grey Model (CFGM) based on CFA and CFD is introduced
with detailed modelling procedures. The feasibility and simplicity and the CFGM
are shown in the numerical example. And the at last the comprehensive
real-world case studies of natural gas production forecasting in 11 countries
are presented, and results show that the CFGM is much more effective than the
existing FGM model in the 165 subcases.
| stat |
Spectral Dependence | This paper presents a general framework for modeling dependence in
multivariate time series. Its fundamental approach relies on decomposing each
signal in a system into various frequency components and then studying the
dependence properties through these oscillatory activities.The unifying theme
across the paper is to explore the strength of dependence and possible lead-lag
dynamics through filtering. The proposed framework is capable of representing
both linear and non-linear dependencies that could occur instantaneously or
after some delay(lagged dependence). Examples for studying dependence between
oscillations are illustrated through multichannel electroencephalograms. These
examples emphasized that some of the most prominent frequency domain measures
such as coherence, partial coherence,and dual-frequency coherence can be
derived as special cases under this general framework.This paper also
introduces related approaches for modeling dependence through phase-amplitude
coupling and causality of (one-sided) filtered signals.
| stat |
Penalized polytomous ordinal logistic regression using cumulative
logits. Application to network inference of zero-inflated variables | We consider the problem of variable selection when the response is ordinal,
that is an ordered categorical variable. In particular, we are interested in
selecting quantitative explanatory variables linked with the ordinal response
variable and we want to determine which predictors are relevant. In this
framework, we choose to use the polytomous ordinal logistic regression model
using cumulative logits which generalizes the logistic regression. We then
introduce the Lasso estimation of the regression coefficients using the
Frank-Wolfe algorithm. To deal with the choice of the penalty parameter, we use
the stability selection method and we develop a new method based on the
knockoffs idea. This knockoffs method is general and suitable to any regression
and besides, gives an order of importance of the covariates. Finally, we
provide some experimental results to corroborate our method. We then present an
application of this regression method for network inference of zero-inflated
variables and use it in practice on real abundance data in an agronomic
context.
| stat |
A weighting method for simultaneous adjustment for confounding and joint
exposure-outcome misclassifications | Joint misclassification of exposure and outcome variables can lead to
considerable bias in epidemiological studies of causal exposure-outcome
effects. In this paper, we present a new maximum likelihood based estimator for
the marginal causal odd-ratio that simultaneously adjusts for confounding and
several forms of joint misclassification of the exposure and outcome variables.
The proposed method relies on validation data for the construction of weights
that account for both sources of bias. The weighting estimator, which is an
extension of the exposure misclassification weighting estimator proposed by
Gravel and Platt (Statistics in Medicine, 2018), is applied to reinfarction
data. Simulation studies were carried out to study its finite sample properties
and compare it with methods that do not account for confounding or
misclassification. The new estimator showed favourable large sample properties
in the simulations. Further research is needed to study the sensitivity of the
proposed method and that of alternatives to violations of their assumptions.
The implementation of the estimator is facilitated by a new R function in an
existing R package.
| stat |
Interpretation of the individual effect under treatment spillover | Some interventions may include important spillover or dissemination effects
between study participants. For example, vaccines, cash transfers, and
education programs may exert a causal effect on participants beyond those to
whom individual treatment is assigned. In a recent paper, Buchanan et al.
provide a causal definition of the "individual effect" of an intervention in
networks of people who inject drugs. In this short note, we discuss the
interpretation of the individual effect when a spillover or dissemination
effect exists.
| stat |
Pareto Smoothed Importance Sampling | Importance weighting is a general way to adjust Monte Carlo integration to
account for draws from the wrong distribution, but the resulting estimate can
be noisy when the importance ratios have a heavy right tail. This routinely
occurs when there are aspects of the target distribution that are not well
captured by the approximating distribution, in which case more stable estimates
can be obtained by modifying extreme importance ratios. We present a new method
for stabilizing importance weights using a generalized Pareto distribution fit
to the upper tail of the distribution of the simulated importance ratios. The
method, which empirically performs better than existing methods for stabilizing
importance sampling estimates, includes stabilized effective sample size
estimates, Monte Carlo error estimates and convergence diagnostics.
| stat |
Using flexible noise models to avoid noise model misspecification in
inference of differential equation time series models | When modelling time series, it is common to decompose observed variation into
a "signal" process, the process of interest, and "noise", representing nuisance
factors that obfuscate the signal. To separate signal from noise, assumptions
must be made about both parts of the system. If the signal process is
incorrectly specified, our predictions using this model may generalise poorly;
similarly, if the noise process is incorrectly specified, we can attribute too
much or too little observed variation to the signal. With little justification,
independent Gaussian noise is typically chosen, which defines a statistical
model that is simple to implement but often misstates system uncertainty and
may underestimate error autocorrelation. There are a range of alternative noise
processes available but, in practice, none of these may be entirely
appropriate, as actual noise may be better characterised as a time-varying
mixture of these various types. Here, we consider systems where the signal is
modelled with ordinary differential equations and present classes of flexible
noise processes that adapt to a system's characteristics. Our noise models
include a multivariate normal kernel where Gaussian processes allow for
non-stationary persistence and variance, and nonparametric Bayesian models that
partition time series into distinct blocks of separate noise structures. Across
the scenarios we consider, these noise processes faithfully reproduce true
system uncertainty: that is, parameter estimate uncertainty when doing
inference using the correct noise model. The models themselves and the methods
for fitting them are scalable to large datasets and could help to ensure more
appropriate quantification of uncertainty in a host of time series models.
| stat |
Inference about complex relationships using peak height data from DNA
mixtures | In both criminal cases and civil cases there is an increasing demand for the
analysis of DNA mixtures involving relationships. The goal might be, for
example, to identify the contributors to a DNA mixture where the donors may be
related, or to infer the relationship between individuals based on a mixture.
This paper introduces an approach to modelling and computation for DNA
mixtures involving contributors with arbitrarily complex relationships. It
builds on an extension of Jacquard's condensed coefficients of identity, to
specify and compute with joint relationships, not only pairwise ones, including
the possibility of inbreeding. The methodology developed is applied to two
casework examples involving a missing person, and simulation studies of
performance, in which the ability of the methodology to recover complex
relationship information from synthetic data with known `true' family structure
is examined.
The methods used to analyse the examples are implemented in the new KinMix R
package, that extends the DNAmixtures package to allow for modelling DNA
mixtures with related contributors. KinMix inherits from DNAmixtures the
capacity to deal with mixtures with many contributors, in a time- and
space-efficient way.
| stat |
Estimation of Discrete Choice Models: A Machine Learning Approach | In this paper we propose a new method of estimation for discrete choice
demand models when individual level data are available. The method employs a
two-step procedure. Step 1 predicts the choice probabilities as functions of
the observed individual level characteristics. Step 2 estimates the structural
parameters of the model using the estimated choice probabilities at a
particular point of interest and the moment restrictions. In essence, the
method uses nonparametric approximation (followed by) moment estimation. Hence
the name---NAME. We use simulations to compare the performance of NAME with the
standard methodology. We find that our method improves precision as well as
convergence time. We supplement the analysis by providing the large sample
properties of the proposed estimator.
| stat |
Vector Autoregressive Moving Average Model with Scalar Moving Average | We show Vector Autoregressive Moving Average models with scalar Moving
Average components could be estimated by generalized least square (GLS) for
each fixed moving average polynomial. The conditional variance of the GLS model
is the concentrated covariant matrix of the moving average process. Under GLS
the likelihood function of these models has similar format to their VAR
counterparts. Maximum likelihood estimate can be done by optimizing with
gradient over the moving average parameters. These models are inexpensive
generalizations of Vector Autoregressive models. We discuss a relationship
between this result and the Borodin-Okounkov formula in operator theory.
| stat |
Two-step penalised logistic regression for multi-omic data with an
application to cardiometabolic syndrome | Building classification models that predict a binary class label on the basis
of high dimensional multi-omics datasets poses several challenges, due to the
typically widely differing characteristics of the data layers in terms of
number of predictors, type of data, and levels of noise. Previous research has
shown that applying classical logistic regression with elastic-net penalty to
these datasets can lead to poor results (Liu et al., 2018). We implement a
two-step approach to multi-omic logistic regression in which variable selection
is performed on each layer separately and a predictive model is then built
using the variables selected in the first step. Here, our approach is compared
to other methods that have been developed for the same purpose, and we adapt
existing software for multi-omic linear regression (Zhao and Zucknick, 2020) to
the logistic regression setting. Extensive simulation studies show that our
approach should be preferred if the goal is to select as many relevant
predictors as possible, as well as achieving prediction performances comparable
to those of the best competitors. Our motivating example is a cardiometabolic
syndrome dataset comprising eight 'omic data types for 2 extreme phenotype
groups (10 obese and 10 lipodystrophy individuals) and 185 blood donors. Our
proposed approach allows us to identify features that characterise
cardiometabolic syndrome at the molecular level. R code is available at
https://github.com/acabassi/logistic-regression-for-multi-omic-data.
| stat |
A comparison of group testing architectures for COVID-19 testing | An important component of every country's COVID-19 response is fast and
efficient testing - to identify and isolate cases, as well as for early
detection of local hotspots. For many countries, producing a sufficient number
of tests has been a serious limiting factor in their efforts to control
COVID-19 infections. Group testing is a well-established mathematical tool,
which can provide a substantial and inexpensive expansion of testing capacity.
In this note, we compare several popular group testing schemes in the context
of qPCR testing for COVID-19. We find that in practical settings, for
identification of individuals with COVID-19, Dorfman testing is the best choice
at prevalences up to 30%, while for estimation of COVID-19 prevalence rates in
the total population, Gibbs-Gower testing is the best choice at prevalences up
to 30% given a fixed and relatively small number of tests. For instance, at a
prevalence of up to 2%, Dorfman testing gives an efficiency gain of 3.5--8; at
1% prevalence, Gibbs-Gower testing gives an efficiency gain of 18, even when
capping the pool size at a feasible number .
This note is intended as a helpful handbook for labs implementing group
testing methods.
| stat |
A Bayesian binomial regression model with latent Gaussian processes for
modelling DNA methylation | Epigenetic observations are represented by the total number of reads from a
given pool of cells and the number of methylated reads, making it reasonable to
model this data by a binomial distribution. There are numerous factors that can
influence the probability of success in a particular region. Moreover, there is
a strong spatial (alongside the genome) dependence of these probabilities. We
incorporate dependence on the covariates and the spatial dependence of the
methylation probability for observations from a pool of cells by means of a
binomial regression model with a latent Gaussian field and a logit link
function. We apply a Bayesian approach including prior specifications on model
configurations. We run a mode jumping Markov chain Monte Carlo algorithm
(MJMCMC) across different choices of covariates in order to obtain the joint
posterior distribution of parameters and models. This also allows finding the
best set of covariates to model methylation probability within the genomic
region of interest and individual marginal inclusion probabilities of the
covariates.
| stat |
Many perspectives on Deborah Mayo's "Statistical Inference as Severe
Testing: How to Get Beyond the Statistics Wars" | The new book by philosopher Deborah Mayo is relevant to data science for
topical reasons, as she takes various controversial positions regarding
hypothesis testing and statistical practice, and also as an entry point to
thinking about the philosophy of statistics. The present article is a slightly
expanded version of a series of informal reviews and comments on Mayo's book.
We hope this discussion will introduce people to Mayo's ideas along with other
perspectives on the topics she addresses.
| stat |
Data filtering methods for SARS-CoV-2 wastewater surveillance | In the case of SARS-CoV-2 pandemic management, wastewater-based epidemiology
aims to derive information on the infection dynamics by monitoring virus
concentrations in the wastewater. However, due to the intrinsic random
fluctuations of the viral signal in the wastewater (due to e.g., dilution;
transport and fate processes in sewer system; variation in the number of
persons discharging; variations in virus excretion and water consumption per
day) the subsequent prevalence analysis may result in misleading conclusions.
It is thus helpful to apply data filtering techniques to reduce the noise in
the signal. In this paper we investigate 13 smoothing algorithms applied to the
virus signals monitored in four wastewater treatment plants in Austria. The
parameters of the algorithms have been defined by an optimization procedure
aiming for performance metrics. The results are further investigated by means
of a cluster analysis. While all algorithms are in principle applicable,
SPLINE, Generalized Additive Model and Friedman Super Smoother are recognized
as superior methods in this context (with the latter two having a tendency to
over-smoothing). A first analysis of the resulting datasets indicates the
influence of catchment size for wastewater-based epidemiology as smaller
communities both reveal a signal threshold before any relation with infection
dynamics is visible and also a higher sensitivity towards infection clusters.
| stat |
Surrogate-based Bayesian Comparison of Computationally Expensive Models:
Application to Microbially Induced Calcite Precipitation | Geochemical processes in subsurface reservoirs affected by microbial activity
change the material properties of porous media. This is a complex
biogeochemical process in subsurface reservoirs that currently contains strong
conceptual uncertainty. This means, several modeling approaches describing the
biogeochemical process are plausible and modelers face the uncertainty of
choosing the most appropriate one. Once observation data becomes available, a
rigorous Bayesian model selection accompanied by a Bayesian model
justifiability analysis could be employed to choose the most appropriate model,
i.e. the one that describes the underlying physical processes best in the light
of the available data. However, biogeochemical modeling is computationally very
demanding because it conceptualizes different phases, biomass dynamics,
geochemistry, precipitation and dissolution in porous media. Therefore, the
Bayesian framework cannot be based directly on the full computational models as
this would require too many expensive model evaluations. To circumvent this
problem, we suggest performing both Bayesian model selection and justifiability
analysis after constructing surrogates for the competing biogeochemical models.
Here, we use the arbitrary polynomial chaos expansion. We account for the
approximation error in the Bayesian analysis by introducing novel correction
factors for the resulting model weights. Thereby, we extend the Bayesian
justifiability analysis and assess model similarities for computationally
expensive models. We demonstrate the method on a representative scenario for
microbially induced calcite precipitation in a porous medium. Our extension of
the justifiability analysis provides a suitable approach for the comparison of
computationally demanding models and gives an insight on the necessary amount
of data for a reliable model performance.
| stat |
Conditional Density Estimation with Neural Networks: Best Practices and
Benchmarks | Given a set of empirical observations, conditional density estimation aims to
capture the statistical relationship between a conditional variable
$\mathbf{x}$ and a dependent variable $\mathbf{y}$ by modeling their
conditional probability $p(\mathbf{y}|\mathbf{x})$. The paper develops best
practices for conditional density estimation for finance applications with
neural networks, grounded on mathematical insights and empirical evaluations.
In particular, we introduce a noise regularization and data normalization
scheme, alleviating problems with over-fitting, initialization and
hyper-parameter sensitivity of such estimators. We compare our proposed
methodology with popular semi- and non-parametric density estimators, underpin
its effectiveness in various benchmarks on simulated and Euro Stoxx 50 data and
show its superior performance. Our methodology allows to obtain high-quality
estimators for statistical expectations of higher moments, quantiles and
non-linear return transformations, with very little assumptions about the
return dynamic.
| stat |
Online Smoothing for Diffusion Processes Observed with Noise | We introduce a methodology for online estimation of smoothing expectations
for a class of additive functionals, in the context of a rich family of
diffusion processes (that may include jumps) -- observed at discrete-time
instances. We overcome the unavailability of the transition density of the
underlying SDE by working on the augmented pathspace. The new method can be
applied, for instance, to carry out online parameter inference for the
designated class of models. Algorithms defined on the infinite-dimensional
pathspace have been developed in the last years mainly in the context of MCMC
techniques. There, the main benefit is the achievement of mesh-free mixing
times for the practical time-discretised algorithm used on a PC. Our own
methodology sets up the framework for infinite-dimensional online filtering --
an important positive practical consequence is the construct of estimates with
the variance that does not increase with decreasing mesh-size. Besides
regularity conditions, our method is, in principle, applicable under the weak
assumption -- relatively to restrictive conditions often required in the MCMC
or filtering literature of methods defined on pathspace -- that the SDE
covariance matrix is invertible.
| stat |
On Adversarial Bias and the Robustness of Fair Machine Learning | Optimizing prediction accuracy can come at the expense of fairness. Towards
minimizing discrimination against a group, fair machine learning algorithms
strive to equalize the behavior of a model across different groups, by imposing
a fairness constraint on models. However, we show that giving the same
importance to groups of different sizes and distributions, to counteract the
effect of bias in training data, can be in conflict with robustness. We analyze
data poisoning attacks against group-based fair machine learning, with the
focus on equalized odds. An adversary who can control sampling or labeling for
a fraction of training data, can reduce the test accuracy significantly beyond
what he can achieve on unconstrained models. Adversarial sampling and
adversarial labeling attacks can also worsen the model's fairness gap on test
data, even though the model satisfies the fairness constraint on training data.
We analyze the robustness of fair machine learning through an empirical
evaluation of attacks on multiple algorithms and benchmark datasets.
| stat |
NeurT-FDR: Controlling FDR by Incorporating Feature Hierarchy | Controlling false discovery rate (FDR) while leveraging the side information
of multiple hypothesis testing is an emerging research topic in modern data
science. Existing methods rely on the test-level covariates while ignoring
possible hierarchy among the covariates. This strategy may not be optimal for
complex large-scale problems, where hierarchical information often exists among
those test-level covariates. We propose NeurT-FDR which boosts statistical
power and controls FDR for multiple hypothesis testing while leveraging the
hierarchy among test-level covariates. Our method parametrizes the test-level
covariates as a neural network and adjusts the feature hierarchy through a
regression framework, which enables flexible handling of high-dimensional
features as well as efficient end-to-end optimization. We show that NeurT-FDR
has strong FDR guarantees and makes substantially more discoveries in synthetic
and real datasets compared to competitive baselines.
| stat |
Statistical analysis of periodic data in neuroscience | Many experimental paradigms in neuroscience involve driving the nervous
system with periodic sensory stimuli. Neural signals recorded with a variety of
techniques will then include phase-locked oscillations at the stimulation
frequency. The analysis of such data often involves standard univariate
statistics such as T-tests, conducted on the Fourier amplitude components
(ignoring phase), either to test for the presence of a signal, or to compare
signals across different conditions. However, the assumptions of these tests
will often be violated because amplitudes are not normally distributed, and
furthermore weak signals might be missed if the phase information is discarded.
An alternative approach is to conduct multivariate statistical tests using the
real and imaginary Fourier components. Here the performance of two multivariate
extensions of the T-test are compared: Hotelling's $T^2$ and a variant called
$T^2_{circ}$. A novel test of the assumptions of $T^2_{circ}$ is developed,
based on the condition index of the data (the square root of the ratio of
eigenvalues of a bounding ellipse), and a heuristic for excluding outliers
using the Mahalanobis distance is proposed. The $T^2_{circ}$ statistic is then
extended to multi-level designs, resulting in a new statistical test termed
$ANOVA^2_{circ}$. This has identical assumptions to $T^2_{circ}$, and is shown
to be more sensitive than MANOVA when these assumptions are met. The use of
these tests is demonstrated for two publicly available empirical data sets, and
practical guidance is suggested for choosing which test to run. Implementations
of these novel tools are provided as an R package and a Matlab toolbox, in the
hope that their wider adoption will improve the sensitivity of statistical
inferences involving periodic data.
| stat |
Recyclable Gaussian Processes | We present a new framework for recycling independent variational
approximations to Gaussian processes. The main contribution is the construction
of variational ensembles given a dictionary of fitted Gaussian processes
without revisiting any subset of observations. Our framework allows for
regression, classification and heterogeneous tasks, i.e. mix of continuous and
discrete variables over the same input domain. We exploit infinite-dimensional
integral operators based on the Kullback-Leibler divergence between stochastic
processes to re-combine arbitrary amounts of variational sparse approximations
with different complexity, likelihood model and location of the pseudo-inputs.
Extensive results illustrate the usability of our framework in large-scale
distributed experiments, also compared with the exact inference models in the
literature.
| stat |
The Probabilistic Fault Tolerance of Neural Networks in the Continuous
Limit | The loss of a few neurons in a brain rarely results in any visible loss of
function. However, the insight into what "few" means in this context is
unclear. How many random neuron failures will it take to lead to a visible loss
of function? In this paper, we address the fundamental question of the impact
of the crash of a random subset of neurons on the overall computation of a
neural network and the error in the output it produces. We study fault
tolerance of neural networks subject to small random neuron/weight crash
failures in a probabilistic setting. We give provable guarantees on the
robustness of the network to these crashes. Our main contribution is a bound on
the error in the output of a network under small random Bernoulli crashes
proved by using a Taylor expansion in the continuous limit, where close-by
neurons at a layer are similar. The failure mode we adopt in our model is
characteristic of neuromorphic hardware, a promising technology to speed up
artificial neural networks, as well as of biological networks. We show that our
theoretical bounds can be used to compare the fault tolerance of different
architectures and to design a regularizer improving the fault tolerance of a
given architecture. We design an algorithm achieving fault tolerance using a
reasonable number of neurons. In addition to the theoretical proof, we also
provide experimental validation of our results and suggest a connection to the
generalization capacity problem.
| stat |
One-Trial Correction of Legacy AI Systems and Stochastic Separation
Theorems | We consider the problem of efficient "on the fly" tuning of existing, or {\it
legacy}, Artificial Intelligence (AI) systems. The legacy AI systems are
allowed to be of arbitrary class, albeit the data they are using for computing
interim or final decision responses should posses an underlying structure of a
high-dimensional topological real vector space. The tuning method that we
propose enables dealing with errors without the need to re-train the system.
Instead of re-training a simple cascade of perceptron nodes is added to the
legacy system. The added cascade modulates the AI legacy system's decisions. If
applied repeatedly, the process results in a network of modulating rules
"dressing up" and improving performance of existing AI systems. Mathematical
rationale behind the method is based on the fundamental property of measure
concentration in high dimensional spaces. The method is illustrated with an
example of fine-tuning a deep convolutional network that has been pre-trained
to detect pedestrians in images.
| stat |
Predictive inference with Fleming--Viot-driven dependent Dirichlet
processes | We consider predictive inference using a class of temporally dependent
Dirichlet processes driven by Fleming--Viot diffusions, which have a natural
bearing in Bayesian nonparametrics and lend the resulting family of random
probability measures to analytical posterior analysis. Formulating the implied
statistical model as a hidden Markov model, we fully describe the predictive
distribution induced by these Fleming--Viot-driven dependent Dirichlet
processes, for a sequence of observations collected at a certain time given
another set of draws collected at several previous times. This is identified as
a mixture of P\'olya urns, whereby the observations can be values from the
baseline distribution or copies of previous draws collected at the same time as
in the usual P\`olya urn, or can be sampled from a random subset of the data
collected at previous times. We characterise the time-dependent weights of the
mixture which select such subsets and discuss the asymptotic regimes. We
describe the induced partition by means of a Chinese restaurant process
metaphor with a conveyor belt, whereby new customers who do not sit at an
occupied table open a new table by picking a dish either from the baseline
distribution or from a time-varying offer available on the conveyor belt. We
lay out explicit algorithms for exact and approximate posterior sampling of
both observations and partitions, and illustrate our results on predictive
problems with synthetic and real data.
| stat |
Small Area Estimation of Health Outcomes | Small area estimation (SAE) entails estimating characteristics of interest
for domains, often geographical areas, in which there may be few or no samples
available. SAE has a long history and a wide variety of methods have been
suggested, from a bewildering range of philosophical standpoints. We describe
design-based and model-based approaches and models that are specified at the
area-level and at the unit-level, focusing on health applications and fully
Bayesian spatial models. The use of auxiliary information is a key ingredient
for successful inference when response data are sparse and we discuss a number
of approaches that allow the inclusion of covariate data. SAE for HIV
prevalence, using data collected from a Demographic Health Survey in Malawi in
2015-2016, is used to illustrate a number of techniques. The potential use of
SAE techniques for outcomes related to COVID-19 is discussed.
| stat |
Probabilistic Neural Architecture Search | In neural architecture search (NAS), the space of neural network
architectures is automatically explored to maximize predictive accuracy for a
given task. Despite the success of recent approaches, most existing methods
cannot be directly applied to large scale problems because of their prohibitive
computational complexity or high memory usage. In this work, we propose a
Probabilistic approach to neural ARchitecture SEarCh (PARSEC) that drastically
reduces memory requirements while maintaining state-of-the-art computational
complexity, making it possible to directly search over more complex
architectures and larger datasets. Our approach only requires as much memory as
is needed to train a single architecture from our search space. This is due to
a memory-efficient sampling procedure wherein we learn a probability
distribution over high-performing neural network architectures. Importantly,
this framework enables us to transfer the distribution of architectures learnt
on smaller problems to larger ones, further reducing the computational cost. We
showcase the advantages of our approach in applications to CIFAR-10 and
ImageNet, where our approach outperforms methods with double its computational
cost and matches the performance of methods with costs that are three orders of
magnitude larger.
| stat |
Estimating Local Interactions Among Many Agents Who Observe Their
Neighbors | In various economic environments, people observe other people with whom they
strategically interact. We can model such information-sharing relations as an
information network, and the strategic interactions as a game on the network.
When any two agents in the network are connected either directly or indirectly
in a large network, empirical modeling using an equilibrium approach can be
cumbersome, since the testable implications from an equilibrium generally
involve all the players of the game, whereas a researcher's data set may
contain only a fraction of these players in practice. This paper develops a
tractable empirical model of linear interactions where each agent, after
observing part of his neighbors' types, not knowing the full information
network, uses best responses that are linear in his and other players' types
that he observes, based on simple beliefs about the other players' strategies.
We provide conditions on information networks and beliefs such that the best
responses take an explicit form with multiple intuitive features. Furthermore,
the best responses reveal how local payoff interdependence among agents is
translated into local stochastic dependence of their actions, allowing the
econometrician to perform asymptotic inference without having to observe all
the players in the game or having to know the precise sampling process.
| stat |
Comparison of Methods for the Assessment of Nonlinearity in Short-Term
Heart Rate Variability under different Physiopathological States | Despite the widespread diffusion of nonlinear methods for heart rate
variability (HRV) analysis, the presence and the extent to which nonlinear
dynamics contribute to short-term HRV is still controversial. This work aims at
testing the hypothesis that different types of nonlinearity can be observed in
HRV depending on the method adopted and on the physiopathological state. Two
entropy-based measures of time series complexity (normalized complexity index,
NCI) and regularity (information storage, IS), and a measure quantifying
deviations from linear correlations in a time series (Gaussian linear contrast,
GLC), are applied to short HRV recordings obtained in young (Y) and old (O)
healthy subjects and in myocardial infarction (MI) patients monitored in the
resting supine position and in the upright position reached through head-up
tilt. The method of surrogate data is employed to detect the presence of and
quantify the contribution of nonlinear dynamics to HRV. We find that the three
measures differ both in their variations across groups and conditions and in
the number and strength of nonlinear HRV dynamics detected: at rest, IS reveals
a significantly lower number of nonlinear dynamics in Y, whereas during tilt
GLC reveals significantly stronger nonlinear HRV dynamics in MI; in the
transition from rest to tilt, all measures detect a significant weakening of
nonlinear HRV dynamics in Y, while only GLC detects a significant strengthening
of such dynamics in MI. These results suggest that distinct dynamic structures,
detected with different sensitivity by nonlinear measures, lie beneath
short-term HRV in different physiological states and pathological conditions.
| stat |
Simulating an infinite mean waiting time | We consider a hybrid method to simulate the return time to the initial state
in a critical-case birth--death process. The expected value of this return time
is infinite, but its distribution asymptotically follows a power-law. Hence,
the simulation approach is to directly simulate the process, unless the
simulated time exceeds some threshold and if it does, draw the return time from
the tail of the power law.
| stat |
Bayesian Landmark-based Shape Analysis of Tumor Pathology Images | Medical imaging is a form of technology that has revolutionized the medical
field in the past century. In addition to radiology imaging of tumor tissues,
digital pathology imaging, which captures histological details in high spatial
resolution, is fast becoming a routine clinical procedure for cancer diagnosis
support and treatment planning. Recent developments in deep-learning methods
facilitate the segmentation of tumor regions at almost the cellular level from
digital pathology images. The traditional shape features that were developed
for characterizing tumor boundary roughness in radiology are not applicable.
Reliable statistical approaches to modeling tumor shape in pathology images are
in urgent need. In this paper, we consider the problem of modeling a tumor
boundary with a closed polygonal chain. A Bayesian landmark-based shape
analysis (BayesLASA) model is proposed to partition the polygonal chain into
mutually exclusive segments to quantify the boundary roughness piecewise. Our
fully Bayesian inference framework provides uncertainty estimates of both the
number and locations of landmarks. The BayesLASA outperforms a recently
developed landmark detection model for planar elastic curves in terms of
accuracy and efficiency. We demonstrate how this model-based analysis can lead
to sharper inferences than ordinary approaches through a case study on the 246
pathology images from 143 non-small cell lung cancer patients. The case study
shows that the heterogeneity of tumor boundary roughness predicts patient
prognosis (p-value < 0.001). This statistical methodology not only presents a
new model for characterizing a digitized object's shape features by using its
landmarks, but also provides a new perspective for understanding the role of
tumor surface in cancer progression.
| stat |
Continuous-Time Birth-Death MCMC for Bayesian Regression Tree Models | Decision trees are flexible models that are well suited for many statistical
regression problems. In a Bayesian framework for regression trees, Markov Chain
Monte Carlo (MCMC) search algorithms are required to generate samples of tree
models according to their posterior probabilities. The critical component of
such an MCMC algorithm is to construct good Metropolis-Hastings steps for
updating the tree topology. However, such algorithms frequently suffering from
local mode stickiness and poor mixing. As a result, the algorithms are slow to
converge. Hitherto, authors have primarily used discrete-time birth/death
mechanisms for Bayesian (sums of) regression tree models to explore the model
space. These algorithms are efficient only if the acceptance rate is high which
is not always the case. Here we overcome this issue by developing a new search
algorithm which is based on a continuous-time birth-death Markov process. This
search algorithm explores the model space by jumping between parameter spaces
corresponding to different tree structures. In the proposed algorithm, the
moves between models are always accepted which can dramatically improve the
convergence and mixing properties of the MCMC algorithm. We provide theoretical
support of the algorithm for Bayesian regression tree models and demonstrate
its performance.
| stat |
Entropy-Regularized $2$-Wasserstein Distance between Gaussian Measures | Gaussian distributions are plentiful in applications dealing in uncertainty
quantification and diffusivity. They furthermore stand as important special
cases for frameworks providing geometries for probability measures, as the
resulting geometry on Gaussians is often expressible in closed-form under the
frameworks. In this work, we study the Gaussian geometry under the
entropy-regularized 2-Wasserstein distance, by providing closed-form solutions
for the distance and interpolations between elements. Furthermore, we provide a
fixed-point characterization of a population barycenter when restricted to the
manifold of Gaussians, which allows computations through the fixed-point
iteration algorithm. As a consequence, the results yield closed-form
expressions for the 2-Sinkhorn divergence. As the geometries change by varying
the regularization magnitude, we study the limiting cases of vanishing and
infinite magnitudes, reconfirming well-known results on the limits of the
Sinkhorn divergence. Finally, we illustrate the resulting geometries with a
numerical study.
| stat |
Adaptive Bayesian SLOPE -- High-dimensional Model Selection with Missing
Values | We consider the problem of variable selection in high-dimensional settings
with missing observations among the covariates. To address this relatively
understudied problem, we propose a new synergistic procedure -- adaptive
Bayesian SLOPE -- which effectively combines the SLOPE method (sorted $l_1$
regularization) together with the Spike-and-Slab LASSO method. We position our
approach within a Bayesian framework which allows for simultaneous variable
selection and parameter estimation, despite the missing values. As with the
Spike-and-Slab LASSO, the coefficients are regarded as arising from a
hierarchical model consisting of two groups: (1) the spike for the inactive and
(2) the slab for the active. However, instead of assigning independent spike
priors for each covariate, here we deploy a joint "SLOPE" spike prior which
takes into account the ordering of coefficient magnitudes in order to control
for false discoveries. Through extensive simulations, we demonstrate
satisfactory performance in terms of power, FDR and estimation bias under a
wide range of scenarios. Finally, we analyze a real dataset consisting of
patients from Paris hospitals who underwent a severe trauma, where we show
excellent performance in predicting platelet levels. Our methodology has been
implemented in C++ and wrapped into an R package ABSLOPE for public use.
| stat |
Confidently Comparing Estimators with the c-value | Modern statistics provides an ever-expanding toolkit for estimating unknown
parameters. Consequently, applied statisticians frequently face a difficult
decision: retain a parameter estimate from a familiar method or replace it with
an estimate from a newer or complex one. While it is traditional to compare
estimators using risk, such comparisons are rarely conclusive in realistic
settings. In response, we propose the "c-value" as a measure of confidence that
a new estimate achieves smaller loss than an old estimate on a given dataset.
We show that it is unlikely that a computed c-value is large and that the new
estimate has larger loss than the old. Therefore, just as a small p-value
provides evidence to reject a null hypothesis, a large c-value provides
evidence to use a new estimate in place of the old. For a wide class of
problems and estimators, we show how to compute a c-value by first constructing
a data-dependent high-probability lower bound on the difference in loss. The
c-value is frequentist in nature, but we show that it can provide a validation
of Bayesian estimates in real data applications involving hierarchical models
and Gaussian processes.
| stat |
Default Bayesian Model Selection of Constrained Multivariate Normal
Linear Models | The multivariate normal linear model is one of the most widely employed
models for statistical inference in applied research. Special cases include
(multivariate) t testing, (M)AN(C)OVA, (multivariate) multiple regression, and
repeated measures analysis. Statistical procedures for model selection where
the models may have equality and order constraints on the model parameters of
interest are limited however. This paper presents a default Bayes factor for
this model selection problem. The default Bayes factor is based on generalized
fractional Bayes methodology where different fractions are used for different
observations and where the default prior is centered on the boundary of the
constrained space under investigation. First, the method is fully automatic and
therefore can be applied when prior information is weak or completely
unavailable. Second, using group specific fractions, the same amount of
information is used from each group resulting in a minimally informative
default prior having a matrix Cauchy distribution, resulting in a consistent
default Bayes factor. Third, numerical computation can be done using
parallelization which makes it computationally cheap. Fourth, the evidence can
be updated in a relatively simple manner when observing new data. Fifth, the
selection criterion can be applied relatively straightforwardly in the presence
of missing data that are missing at random. Applications for the social and
behavioral sciences are used for illustration.
| stat |
Multi-view Vector-valued Manifold Regularization for Multi-label Image
Classification | In computer vision, image datasets used for classification are naturally
associated with multiple labels and comprised of multiple views, because each
image may contain several objects (e.g. pedestrian, bicycle and tree) and is
properly characterized by multiple visual features (e.g. color, texture and
shape). Currently available tools ignore either the label relationship or the
view complementary. Motivated by the success of the vector-valued function that
constructs matrix-valued kernels to explore the multi-label structure in the
output space, we introduce multi-view vector-valued manifold regularization
(MV$\mathbf{^3}$MR) to integrate multiple features. MV$\mathbf{^3}$MR exploits
the complementary property of different features and discovers the intrinsic
local geometry of the compact support shared by different features under the
theme of manifold regularization. We conducted extensive experiments on two
challenging, but popular datasets, PASCAL VOC' 07 (VOC) and MIR Flickr (MIR),
and validated the effectiveness of the proposed MV$\mathbf{^3}$MR for image
classification.
| stat |
Emulating satellite drag from large simulation experiments | Obtaining accurate estimates of satellite drag coefficients in low Earth
orbit is a crucial component in positioning and collision avoidance. Simulators
can produce accurate estimates, but their computational expense is much too
large for real-time application. A pilot study showed that Gaussian process
(GP) surrogate models could accurately emulate simulations. However, cubic
runtime for training GPs means that they could only be applied to a narrow
range of input configurations to achieve the desired level of accuracy. In this
paper we show how extensions to the local approximate Gaussian Process (laGP)
method allow accurate full-scale emulation. The new methodological
contributions, which involve a multi-level global/local modeling approach, and
a set-wise approach to local subset selection, are shown to perform well in
benchmark and synthetic data settings. We conclude by demonstrating that our
method achieves the desired level of accuracy, besting simpler viable (i.e.,
computationally tractable) global and local modeling approaches, when trained
on seventy thousand core hours of drag simulations for two real-world
satellites: the Hubble space telescope (HST) and the gravity recovery and
climate experiment (GRACE).
| stat |
High dimensionality: The latest challenge to data analysis | The advent of modern technology, permitting the measurement of thousands of
characteristics simultaneously, has given rise to floods of data characterized
by many large or even huge datasets. This new paradigm presents extraordinary
challenges to data analysis and the question arises: how can conventional data
analysis methods, devised for moderate or small datasets, cope with the
complexities of modern data? The case of high dimensional data is particularly
revealing of some of the drawbacks. We look at the case where the number of
characteristics measured in an object is at least the number of observed
objects and conclude that this configuration leads to geometrical and
mathematical oddities and is an insurmountable barrier for the direct
application of traditional methodologies. If scientists are going to ignore
fundamental mathematical results arrived at in this paper and blindly use
software to analyze data, the results of their analyses may not be trustful,
and the findings of their experiments may never be validated. That is why new
methods together with the wise use of traditional approaches are essential to
progress safely through the present reality.
| stat |
Kernel Machine and Distributed Lag Models for Assessing Windows of
Susceptibility to Environmental Mixtures in Children's Health Studies | Exposures to environmental chemicals during gestation can alter health status
later in life. Most studies of maternal exposure to chemicals during pregnancy
focus on a single chemical exposure observed at high temporal resolution.
Recent research has turned to focus on exposure to mixtures of multiple
chemicals, generally observed at a single time point. We consider statistical
methods for analyzing data on chemical mixtures that are observed at a high
temporal resolution. As motivation, we analyze the association between exposure
to four ambient air pollutants observed weekly throughout gestation and birth
weight in a Boston-area prospective birth cohort. To explore patterns in the
data, we first apply methods for analyzing data on (1) a single chemical
observed at high temporal resolution, and (2) a mixture measured at a single
point in time. We highlight the shortcomings of these approaches for
temporally-resolved data on exposure to chemical mixtures. Second, we propose a
novel method, a Bayesian kernel machine regression distributed lag model
(BKMR-DLM), that simultaneously accounts for nonlinear associations and
interactions among time-varying measures of exposure to mixtures. BKMR-DLM uses
a functional weight for each exposure that parameterizes the window of
susceptibility corresponding to that exposure within a kernel machine framework
that captures non-linear and interaction effects of the multivariate exposure
on the outcome. In a simulation, we show that the proposed method can better
estimate the exposure-response function and, in high signal settings, can
identify critical windows in time during which exposure has an increased
association with the outcome. Applying the proposed methods to the birth cohort
data, we found evidence of a negative association between OC and birth weight
and that nitrate modifies the OC, EC, and sulfate exposure-response functions.
| stat |
Generalized Autoregressive Moving Average Models with GARCH Errors | One of the important and widely used classes of models for non-Gaussian time
series is the generalized autoregressive model average models (GARMA), which
specifies an ARMA structure for the conditional mean process of the underlying
time series. However, in many applications one often encounters conditional
heteroskedasticity. In this paper we propose a new class of models, referred to
as GARMA-GARCH models, that jointly specify both the conditional mean and
conditional variance processes of a general non-Gaussian time series. Under the
general modeling framework, we propose three specific models, as examples, for
proportional time series, nonnegative time series, and skewed and heavy-tailed
financial time series. Maximum likelihood estimator (MLE) and quasi Gaussian
MLE (GMLE) are used to estimate the parameters. Simulation studies and three
applications are used to demonstrate the properties of the models and the
estimation procedures.
| stat |
A note on marginal correlation based screening | Independence screening methods such as the two sample $t$-test and the
marginal correlation based ranking are among the most widely used techniques
for variable selection in ultrahigh dimensional data sets. In this short note,
simple examples are used to demonstrate potential problems with the
independence screening methods in the presence of correlated predictors. Also,
an example is considered where all important variables are independent among
themselves and all but one important variables are independent with the
unimportant variables. Furthermore, a real data example from a genome wide
association study is used to illustrate inferior performance of marginal
correlation screening compared to another screening method.
| stat |
A Computational Theory of Robust Localization Verifiability in the
Presence of Pure Outlier Measurements | The problem of localizing a set of nodes from relative pairwise measurements
is at the core of many applications such as Structure from Motion (SfM), sensor
networks, and Simultaneous Localization And Mapping (SLAM). In practical
situations, the accuracy of the relative measurements is marred by noise and
outliers; hence, we have the problem of quantifying how much we should trust
the solution returned by some given localization solver. In this work, we focus
on the question of whether an L1-norm robust optimization formulation can
recover a solution that is identical to the ground truth, under the scenario of
translation-only measurements corrupted exclusively by outliers and no noise;
we call this concept verifiability. On the theoretical side, we prove that the
verifiability of a problem depends only on the topology of the graph of
measurements, the edge support of the outliers, and their signs, while it is
independent of ground truth locations of the nodes, and of any positive scaling
of the outliers. On the computational side, we present a novel approach based
on the dual simplex algorithm that can check the verifiability of a problem,
completely characterize the space of equivalent solutions if they exist, and
identify subgraphs that are verifiable. As an application of our theory, we
provide a procedure to compute a priori probability of recovering a solution
congruent or equivalent to the ground truth given a measurement graph and the
probabilities of each edge containing an outlier.
| stat |
Estimating conditional probabilities of historical migrations in the
transatlantic slave trade using kriging and Markov decision process models | Intra-African conflicts during the collapse of the kingdom of Oyo from 1817
to 1836 resulted in the enslavement of an estimated 121,000 people who were
then transported to coastal ports via complex trade networks and loaded onto
slave ships destined for the Americas. Historians have a good record of where
these people went across the Atlantic, but little is known about where
individuals were from or enslaved \textit{within} Africa. In this work, we
develop a novel two-step statistical approach to describe the enslavement of
people given documented violent conflict, the transport of enslaved peoples
from their location of capture to their port of departure, and---given an
enslaved individual's location of departure---that person's probability of
origin. We combine spatial prediction of conflict density via Kriging with a
Markov decision process characterising intra-African transportation. The
results of this model can be visualised using an interactive web application,
plotting estimated conditional probabilities of historical migrations during
the African diaspora. These results help trace the uncertain origins of people
enslaved in this region of Africa during this time period: using the two-step
statistical methodology developed here provides a probabilistic answer to this
question.
| stat |
Causal Inference in the Time of Covid-19 | In this paper we develop statistical methods for causal inference in
epidemics. Our focus is in estimating the effect of social mobility on deaths
in the Covid-19 pandemic. We propose a marginal structural model motivated by a
modified version of a basic epidemic model. We estimate the counterfactual time
series of deaths under interventions on mobility. We conduct several types of
sensitivity analyses. We find that the data support the idea that reduced
mobility causes reduced deaths, but the conclusion comes with caveats. There is
evidence of sensitivity to model misspecification and unmeasured confounding
which implies that the size of the causal effect needs to be interpreted with
caution. While there is little doubt the the effect is real, our work
highlights the challenges in drawing causal inferences from pandemic data.
| stat |
Floodgate: inference for model-free variable importance | Many modern applications seek to understand the relationship between an
outcome variable $Y$ and a covariate $X$ in the presence of a (possibly
high-dimensional) confounding variable $Z$. Although much attention has been
paid to testing whether $Y$ depends on $X$ given $Z$, in this paper we seek to
go beyond testing by inferring the strength of that dependence. We first define
our estimand, the minimum mean squared error (mMSE) gap, which quantifies the
conditional relationship between $Y$ and $X$ in a way that is deterministic,
model-free, interpretable, and sensitive to nonlinearities and interactions. We
then propose a new inferential approach called floodgate that can leverage any
working regression function chosen by the user (allowing, e.g., it to be fitted
by a state-of-the-art machine learning algorithm or be derived from qualitative
domain knowledge) to construct asymptotic confidence bounds, and we apply it to
the mMSE gap. In addition to proving floodgate's asymptotic validity, we
rigorously quantify its accuracy (distance from confidence bound to estimand)
and robustness. We demonstrate floodgate's performance in a series of
simulations and apply it to data from the UK Biobank to infer the strengths of
dependence of platelet count on various groups of genetic mutations.
| stat |
Applications of Common Entropy for Causal Inference | We study the problem of discovering the simplest latent variable that can
make two observed discrete variables conditionally independent. The minimum
entropy required for such a latent is known as common entropy in information
theory. We extend this notion to Renyi common entropy by minimizing the Renyi
entropy of the latent variable. To efficiently compute common entropy, we
propose an iterative algorithm that can be used to discover the trade-off
between the entropy of the latent variable and the conditional mutual
information of the observed variables. We show two applications of common
entropy in causal inference: First, under the assumption that there are no
low-entropy mediators, it can be used to distinguish causation from spurious
correlation among almost all joint distributions on simple causal graphs with
two observed variables. Second, common entropy can be used to improve
constraint-based methods such as PC or FCI algorithms in the small-sample
regime, where these methods are known to struggle. We propose a modification to
these constraint-based methods to assess if a separating set found by these
algorithms is valid using common entropy. We finally evaluate our algorithms on
synthetic and real data to establish their performance.
| stat |
A Bayesian Statistics Course for Undergraduates: Bayesian Thinking,
Computing, and Research | We propose a semester-long Bayesian statistics course for undergraduate
students with calculus and probability background. We cultivate students'
Bayesian thinking with Bayesian methods applied to real data problems. We
leverage modern Bayesian computing techniques not only for implementing
Bayesian methods, but also to deepen students' understanding of the methods.
Collaborative case studies further enrich students' learning and provide
experience to solve open-ended applied problems. The course has an emphasis on
undergraduate research, where accessible academic journal articles are read,
discussed, and critiqued in class. With increased confidence and familiarity,
students take the challenge of reading, implementing, and sometimes extending
methods in journal articles for their course projects.
| stat |
Prevalence Estimation from Random Samples and Census Data with
Participation Bias | Countries officially record the number of COVID-19 cases based on medical
tests of a subset of the population with unknown participation bias. For
prevalence estimation, the official information is typically discarded and,
instead, small random survey samples are taken. We derive (maximum likelihood
and method of moment) prevalence estimators, based on a survey sample, that
additionally utilize the official information, and that are substantially more
accurate than the simple sample proportion of positive cases. Put differently,
using our estimators, the same level of precision can be obtained with
substantially smaller survey samples. We take into account the possibility of
measurement errors due to the sensitivity and specificity of the medical
testing procedure. The proposed estimators and associated confidence intervals
are implemented in the companion open source R package cape.
| stat |
Generalized tensor regression with covariates on multiple modes | We consider the problem of tensor-response regression given covariates on
multiple modes. Such data problems arise frequently in applications such as
neuroimaging, network analysis, and spatial-temporal modeling. We propose a new
family of tensor response regression models that incorporate covariates, and
establish the theoretical accuracy guarantees. Unlike earlier methods, our
estimation allows high-dimensionality in both the tensor response and the
covariate matrices on multiple modes. An efficient alternating updating
algorithm is further developed. Our proposal handles a broad range of data
types, including continuous, count, and binary observations. Through simulation
and applications to two real datasets, we demonstrate the outperformance of our
approach over the state-of-art.
| stat |
Warped Gradient-Enhanced Gaussian Process Surrogate Models for Inference
with Intractable Likelihoods | Markov chain Monte Carlo methods for intractable likelihoods, such as the
exchange algorithm, require simulations of the sufficient statistics at every
iteration of the Markov chain, which often result in expensive computations.
Surrogate models for the likelihood function have been developed to accelerate
inference algorithms in this context. However, these surrogate models tend to
be relatively inflexible, and often provide a poor approximation to the true
likelihood function. In this article, we propose the use of a warped,
gradient-enhanced, Gaussian process surrogate model for the likelihood
function, which jointly models the sample means and variances of the sufficient
statistics, and uses warping functions to capture covariance nonstationarity in
the input parameter space. We show that both the consideration of
nonstationarity and the inclusion of gradient information can be leveraged to
obtain a surrogate model that outperforms the conventional stationary Gaussian
process surrogate model when making inference, particularly in regions where
the likelihood function exhibits a phase transition. We also show that the
proposed surrogate model can be used to improve the effective sample size per
unit time when embedded in exact inferential algorithms. The utility of our
approach in speeding up inferential algorithms is demonstrated on simulated and
real-world data.
| stat |
Confidence intervals for class prevalences under prior probability shift | Point estimation of class prevalences in the presence of data set shift has
been a popular research topic for more than two decades. Less attention has
been paid to the construction of confidence and prediction intervals for
estimates of class prevalences. One little considered question is whether or
not it is necessary for practical purposes to distinguish confidence and
prediction intervals. Another question so far not yet conclusively answered is
whether or not the discriminatory power of the classifier or score at the basis
of an estimation method matters for the accuracy of the estimates of the class
prevalences. This paper presents a simulation study aimed at shedding some
light on these and other related questions.
| stat |
Onset detection: A new approach to QBH system | Query by Humming (QBH) is a system to provide a user with the song(s) which
the user hums to the system. Current QBH method requires the extraction of
onset and pitch information in order to track similarity with various versions
of different songs. However, we here focus on detecting precise onsets only and
use them to build a QBH system which is better than existing methods in terms
of speed and memory and empirically in terms of accuracy. We also provide
statistical analogy for onset detection functions and provide a measure of
error in our algorithm.
| stat |
Visualizing and comparing distributions with half-disk density strips | We propose a user-friendly graphical tool, the half-disk density strip
(HDDS), for visualizing and comparing probability density functions. The HDDS
exploits color shading for representing a distribution in an intuitive way. In
univariate settings, the half-disk density strip allows to immediately discern
the key characteristics of a density, such as symmetry, dispersion, and
multi-modality. In the multivariate settings, we define HDDS tables to
generalize the concept of contingency tables. It is an array of half-disk
density strips, which compactly displays the univariate marginal and
conditional densities of a variable of interest, together with the joint and
marginal densities of the conditioning variables. Moreover, HDDSs are by
construction well suited to easily compare pairs of densities. To highlight the
concrete benefits of the proposed methods, we show how to use HDDSs for
analyzing income distribution and life-satisfaction, conditionally on
continuous and categorical controls, from survey data. The code for
implementing HDDS methods is made available through a dedicated R package.
| stat |
Multi-Objective Hyperparameter Tuning and Feature Selection using Filter
Ensembles | Both feature selection and hyperparameter tuning are key tasks in machine
learning. Hyperparameter tuning is often useful to increase model performance,
while feature selection is undertaken to attain sparse models. Sparsity may
yield better model interpretability and lower cost of data acquisition, data
handling and model inference. While sparsity may have a beneficial or
detrimental effect on predictive performance, a small drop in performance may
be acceptable in return for a substantial gain in sparseness. We therefore
treat feature selection as a multi-objective optimization task. We perform
hyperparameter tuning and feature selection simultaneously because the choice
of features of a model may influence what hyperparameters perform well.
We present, benchmark, and compare two different approaches for
multi-objective joint hyperparameter optimization and feature selection: The
first uses multi-objective model-based optimization. The second is an
evolutionary NSGA-II-based wrapper approach to feature selection which
incorporates specialized sampling, mutation and recombination operators. Both
methods make use of parameterized filter ensembles.
While model-based optimization needs fewer objective evaluations to achieve
good performance, it incurs computational overhead compared to the NSGA-II, so
the preferred choice depends on the cost of evaluating a model on given data.
| stat |
Quantifying Artifacts over Time: Interval Estimation of a Poisson
Distribution using the Jeffreys Prior | This article presents a new method for estimating the amount of an artifact
class in use at a given moment in the past from a random assemblage of
archaeological finds. This method is based on the use of simulation, since an
analytical solution is computationally impractical. Estimating the number of
artifacts in use at any time $t$ is shown to follow a Poisson distribution,
which allows for credible intervals to be established using the Jeffreys prior.
This estimator works from minimal assumptions about the dating and duration of
finds, as well as the intensity of collection, and is applied to coinage from
four Roman-period sites excavated by the Roman Peasant Project (2009-2014). The
result provides for an estimation of the abundance of material according to an
interval of certainty.
| stat |
Residual Entropy | We describe an approach to improving model fitting and model generalization
that considers the entropy of distributions of modelling residuals. We use
simple simulations to demonstrate the observational signatures of overfitting
on ordered sequences of modelling residuals, via the autocorrelation and power
spectral density. These results motivate the conclusion that, as commonly
applied, the least squares method assumes too much when it assumes that
residuals are uncorrelated for all possible models or values of the model
parameters. We relax these too-stringent assumptions in favour of imposing an
entropy prior on the (unknown, model-dependent, but potentially marginalizable)
distribution function for residuals. We recommend a simple extension to the
Mean Squared Error loss function that approximately incorporates this prior and
can be used immediately for modelling applications where meaningfully-ordered
sequences of observations or training data can be defined.
| stat |
Sparse Laplacian Shrinkage with the Graphical Lasso Estimator for
Regression Problems | This paper considers a high-dimensional linear regression problem where there
are complex correlation structures among predictors. We propose a
graph-constrained regularization procedure, named Sparse Laplacian Shrinkage
with the Graphical Lasso Estimator (SLS-GLE). The procedure uses the estimated
precision matrix to describe the specific information on the conditional
dependence pattern among predictors, and encourages both sparsity on the
regression model and the graphical model. We introduce the Laplacian quadratic
penalty adopting the graph information, and give detailed discussions on the
advantages of using the precision matrix to construct the Laplacian matrix.
Theoretical properties and numerical comparisons are presented to show that the
proposed method improves both model interpretability and accuracy of
estimation. We also apply this method to a financial problem and prove that the
proposed procedure is successful in assets selection.
| stat |
A semi-supervised learning framework for quantitative structure-activity
regression modelling | Supervised learning models, also known as quantitative structure-activity
regression (QSAR) models, are increasingly used in assisting the process of
preclinical, small molecule drug discovery. The models are trained on data
consisting of a finite dimensional representation of molecular structures and
their corresponding target specific activities. These models can then be used
to predict the activity of previously unmeasured novel compounds. In this work
we address two problems related to this approach. The first is to estimate the
extent to which the quality of the model predictions degrades for compounds
very different from the compounds in the training data. The second is to adjust
for the screening dependent selection bias inherent in many training data sets.
In the most extreme cases, only compounds which pass an activity-dependent
screening are reported. By using a semi-supervised learning framework, we show
that it is possible to make predictions which take into account the similarity
of the testing compounds to those in the training data and adjust for the
reporting selection bias. We illustrate this approach using publicly available
structure-activity data on a large set of compounds reported by GlaxoSmithKline
(the Tres Cantos AntiMalarial Set) to inhibit in vitro P. falciparum growth.
| stat |
Neural Mixture Distributional Regression | We present neural mixture distributional regression (NMDR), a holistic
framework to estimate complex finite mixtures of distributional regressions
defined by flexible additive predictors. Our framework is able to handle a
large number of mixtures of potentially different distributions in
high-dimensional settings, allows for efficient and scalable optimization and
can be applied to recent concepts that combine structured regression models
with deep neural networks. While many existing approaches for mixture models
address challenges in optimization of such and provide results for convergence
under specific model assumptions, our approach is assumption-free and instead
makes use of optimizers well-established in deep learning. Through extensive
numerical experiments and a high-dimensional deep learning application we
provide evidence that the proposed approach is competitive to existing
approaches and works well in more complex scenarios.
| stat |
Adaptive spline fitting with particle swarm optimization | In fitting data with a spline, finding the optimal placement of knots can
significantly improve the quality of the fit. However, the challenging
high-dimensional and non-convex optimization problem associated with completely
free knot placement has been a major roadblock in using this approach. We
present a method that uses particle swarm optimization (PSO) combined with
model selection to address this challenge. The problem of overfitting due to
knot clustering that accompanies free knot placement is mitigated in this
method by explicit regularization, resulting in a significantly improved
performance on highly noisy data. The principal design choices available in the
method are delineated and a statistically rigorous study of their effect on
performance is carried out using simulated data and a wide variety of benchmark
functions. Our results demonstrate that PSO-based free knot placement leads to
a viable and flexible adaptive spline fitting approach that allows the fitting
of both smooth and non-smooth functions.
| stat |
The Shapley Value of coalition of variables provides better explanations | While Shapley Values (SV) are one of the gold standard for interpreting
machine learning models, we show that they are still poorly understood, in
particular in the presence of categorical variables or of variables of low
importance. For instance, we show that the popular practice that consists in
summing the SV of dummy variables is false as it provides wrong estimates of
all the SV in the model and implies spurious interpretations. Based on the
identification of null and active coalitions, and a coalitional version of the
SV, we provide a correct computation and inference of important variables.
Moreover, a Python library (All the experiments and simulations can be
reproduced with the publicly available library Active Coalition of Variables,
https://www.github.com/salimamoukou/acv00) that computes reliably conditional
expectations and SV for tree-based models, is implemented and compared with
state-of-the-art algorithms on toy models and real data sets.
| stat |
Estimating the error variance in a high-dimensional linear model | The lasso has been studied extensively as a tool for estimating the
coefficient vector in the high-dimensional linear model; however, considerably
less is known about estimating the error variance in this context. In this
paper, we propose the natural lasso estimator for the error variance, which
maximizes a penalized likelihood objective. A key aspect of the natural lasso
is that the likelihood is expressed in terms of the natural parameterization of
the multiparameter exponential family of a Gaussian with unknown mean and
variance. The result is a remarkably simple estimator of the error variance
with provably good performance in terms of mean squared error. These
theoretical results do not require placing any assumptions on the design matrix
or the true regression coefficients. We also propose a companion estimator,
called the organic lasso, which theoretically does not require tuning of the
regularization parameter. Both estimators do well empirically compared to
preexisting methods, especially in settings where successful recovery of the
true support of the coefficient vector is hard. Finally, we show that existing
methods can do well under fewer assumptions than previously known, thus
providing a fuller story about the problem of estimating the error variance in
high-dimensional linear models.
| stat |
A Probabilistic Assessment of the COVID-19 Lockdown on Air Quality in
the UK | In March 2020 the United Kingdom (UK) entered a nationwide lockdown period
due to the Covid-19 pandemic. As a result, levels of nitrogen dioxide (NO2) in
the atmosphere dropped. In this work, we use 550,134 NO2 data points from 237
stations in the UK to build a spatiotemporal Gaussian process capable of
predicting NO2 levels across the entire UK. We integrate several covariate
datasets to enhance the model's ability to capture the complex spatiotemporal
dynamics of NO2. Our numerical analyses show that, within two weeks of a UK
lockdown being imposed, UK NO2 levels dropped 36.8%. Further, we show that as a
direct result of lockdown NO2 levels were 29-38% lower than what they would
have been had no lockdown occurred. In accompaniment to these numerical
results, we provide a software framework that allows practitioners to easily
and efficiently fit similar models.
| stat |
Detection of foraging behavior from accelerometer data using U-Net type
convolutional networks | Narwhal is one of the most mysterious marine mammals, due to its isolated
habitat in the Arctic region. Tagging is a technology that has the potential to
explore the activities of this species, where behavioral information can be
collected from instrumented individuals. This includes accelerometer data,
diving and acoustic data as well as GPS positioning. An essential element in
understanding the ecological role of toothed whales is to characterize their
feeding behavior and estimate the amount of food consumption. Buzzes are sounds
emitted by toothed whales that are related directly to the foraging behaviors.
It is therefore of interest to measure or estimate the rate of buzzing to
estimate prey intake. The main goal of this paper is to find a way to detect
prey capture attempts directly from accelerometer data, and thus be able to
estimate food consumption without the need for the more demanding acoustic
data. We develop 3 automated buzz detection methods based on accelerometer and
depth data solely. We use a dataset from 5 narwhals instrumented in East
Greenland in 2018 to train, validate and test a logistic regression model and
the machine learning algorithms random forest and deep learning, using the
buzzes detected from acoustic data as the ground truth. The deep learning
algorithm performed best among the tested methods. We conclude that reliable
buzz detectors can be derived from high-frequency-sampling, back-mounted
accelerometer tags, thus providing an alternative tool for studies of foraging
ecology of marine mammals in their natural environments. We also compare buzz
detection with certain movement patterns, such as sudden changes in
acceleration (jerks), found in other marine mammal species for estimating prey
capture. We find that narwhals do not seem to make big jerks when foraging and
conclude that their hunting patterns in that respect differ from other marine
mammals.
| stat |
Comparing partitions through the Matching Error | With the aim to propose a non parametric hypothesis test, this paper carries
out a study on the Matching Error (ME), a comparison index of two partitions
obtained from the same data set, using for example two clustering methods. This
index is related to the misclassifica-tion error in supervised learning. Some
properties of the ME and, especially, its distribution function for the case of
two independent partitions are analyzed. Extensive simulations show the
efficiency of the ME and we propose a hypothesis test based on it.
| stat |
Additive Models for Symmetric Positive-Definite Matrices, Riemannian
Manifolds and Lie groups | In this paper an additive regression model for a symmetric positive-definite
matrix valued response and multiple scalar predictors is proposed. The model
exploits the abelian group structure inherited from either the Log-Cholesky
metric or the Log-Euclidean framework that turns the space of symmetric
positive-definite matrices into a Riemannian manifold and further a
bi-invariant Lie group. The additive model for responses in the space of
symmetric positive-definite matrices with either of these metrics is shown to
connect to an additive model on a tangent space. This connection not only
entails an efficient algorithm to estimate the component functions but also
allows to generalize the proposed additive model to general Riemannian
manifolds that might not have a Lie group structure. Optimal asymptotic
convergence rates and normality of the estimated component functions are also
established. Numerical studies show that the proposed model enjoys superior
numerical performance, especially when there are multiple predictors. The
practical merits of the proposed model are demonstrated by analyzing diffusion
tensor brain imaging data.
| stat |
Vectorized Uncertainty Propagation and Input Probability Sensitivity
Analysis | In this article we construct a theoretical and computational process for
assessing Input Probability Sensitivity Analysis (IPSA) using a Graphics
Processing Unit (GPU) enabled technique called Vectorized Uncertainty
Propagation (VUP). VUP propagates probability distributions through a
parametric computational model in a way that's computational time complexity
grows sublinearly in the number of distinct propagated input probability
distributions. VUP can therefore be used to efficiently implement IPSA, which
estimates a model's probabilistic sensitivity to measurement and parametric
uncertainty over each relevant measurement location. Theory and simulation
illustrate the effectiveness of these methods.
| stat |
Modeling a sequence of multinomial data with randomly varying
probabilities | We consider a sequence of variables having multinomial distribution with the
number of trials corresponding to these variables being large and possibly
different. The multinomial probabilities of the categories are assumed to vary
randomly depending on batches. The proposed framework is interesting from the
perspective of various applications in practice such as predicting the winner
of an election, forecasting the market share of different brands etc. In this
work, first we derive sufficient conditions of asymptotic normality of the
estimates of the multinomial cell probabilities, and corresponding suitable
transformations. Then, we consider a Bayesian setting to implement our model.
We consider hierarchical priors using multivariate normal and inverse Wishart
distributions, and establish the posterior consistency. Based on this result
and following appropriate Gibbs sampling algorithms, we can infer about
aggregate data. The methodology is illustrated in detail with two real life
applications, in the contexts of political election and sales forecasting.
Additional insights of effectiveness are also derived through a simulation
study.
| stat |
Probabilistic Performance-Pattern Decomposition (PPPD): analysis
framework and applications to stochastic mechanical systems | Since the early 1900s, numerous research efforts have been devoted to
developing quantitative solutions to stochastic mechanical systems. In general,
the problem is perceived as solved when a complete or partial probabilistic
description on the quantity of interest (QoI) is determined. However, in the
presence of complex system behavior, there is a critical need to go beyond mere
probabilistic descriptions. In fact, to gain a full understanding of the
system, it is crucial to extract physical characterizations from the
probabilistic structure of the QoI, especially when the QoI solution is
obtained in a data-driven fashion. Motivated by this perspective, the paper
proposes a framework to obtain structuralized characterizations on behaviors of
stochastic systems. The framework is named Probabilistic Performance-Pattern
Decomposition (PPPD). PPPD analysis aims to decompose complex response
behaviors, conditional to a prescribed performance state, into meaningful
patterns in the space of system responses, and to investigate how the patterns
are triggered in the space of basic random variables. To illustrate the
application of PPPD, the paper studies three numerical examples: 1) an
illustrative example with hypothetical stochastic processes input and output;
2) a stochastic Lorenz system with periodic as well as chaotic behaviors; and
3) a simplified shear-building model subjected to a stochastic ground motion
excitation.
| stat |
Qsparse-local-SGD: Distributed SGD with Quantization, Sparsification,
and Local Computations | Communication bottleneck has been identified as a significant issue in
distributed optimization of large-scale learning models. Recently, several
approaches to mitigate this problem have been proposed, including different
forms of gradient compression or computing local models and mixing them
iteratively. In this paper, we propose \emph{Qsparse-local-SGD} algorithm,
which combines aggressive sparsification with quantization and local
computation along with error compensation, by keeping track of the difference
between the true and compressed gradients. We propose both synchronous and
asynchronous implementations of \emph{Qsparse-local-SGD}. We analyze
convergence for \emph{Qsparse-local-SGD} in the \emph{distributed} setting for
smooth non-convex and convex objective functions. We demonstrate that
\emph{Qsparse-local-SGD} converges at the same rate as vanilla distributed SGD
for many important classes of sparsifiers and quantizers. We use
\emph{Qsparse-local-SGD} to train ResNet-50 on ImageNet and show that it
results in significant savings over the state-of-the-art, in the number of bits
transmitted to reach target accuracy.
| stat |
Analysis and Simulation of Extremes and Rare Events in Complex Systems | Rare weather and climate events, such as heat waves and floods, can bring
tremendous social costs. Climate data is often limited in duration and spatial
coverage, and climate forecasting has often turned to simulations of climate
models to make better predictions of rare weather events. However very long
simulations of complex models, in order to obtain accurate probability
estimates, may be prohibitively slow. It is an important scientific problem to
develop probabilistic and dynamical techniques to estimate the probabilities of
rare events accurately from limited data. In this paper we compare four modern
methods of estimating the probability of rare events: the generalized extreme
value (GEV) method from classical extreme value theory; two importance sampling
techniques, genealogical particle analysis (GPA) and the
Giardina-Kurchan-Lecomte-Tailleur (GKLT) algorithm; as well as brute force
Monte Carlo (MC). With these techniques we estimate the probabilities of rare
events in three dynamical models: the Ornstein-Uhlenbeck process, the Lorenz
'96 system and PlaSim (a climate model). We keep the computational effort
constant and see how well the rare event probability estimation of each
technique compares to a gold standard afforded by a very long run control.
Somewhat surprisingly we find that classical extreme value theory methods
outperform GPA, GKLT and MC at estimating rare events.
| stat |
A Contour Stochastic Gradient Langevin Dynamics Algorithm for
Simulations of Multi-modal Distributions | We propose an adaptively weighted stochastic gradient Langevin dynamics
algorithm (SGLD), so-called contour stochastic gradient Langevin dynamics
(CSGLD), for Bayesian learning in big data statistics. The proposed algorithm
is essentially a \emph{scalable dynamic importance sampler}, which
automatically \emph{flattens} the target distribution such that the simulation
for a multi-modal distribution can be greatly facilitated. Theoretically, we
prove a stability condition and establish the asymptotic convergence of the
self-adapting parameter to a {\it unique fixed-point}, regardless of the
non-convexity of the original energy function; we also present an error
analysis for the weighted averaging estimators. Empirically, the CSGLD
algorithm is tested on multiple benchmark datasets including CIFAR10 and
CIFAR100. The numerical results indicate its superiority over the existing
state-of-the-art algorithms in training deep neural networks.
| stat |
Uniform convergence of local Fr\'echet regression, with applications to
locating extrema and time warping for metric-space valued trajectories | Local Fr\'echet regression is a nonparametric regression method for
metric-space valued responses and Euclidean predictors, which can be utilized
to obtain estimates of smooth trajectories taking values in general metric
spaces from noisy metric-space valued random objects. We derive uniform rates
of convergence, which so far have eluded theoretical analysis of this method,
for both fixed and random target functions, where we utilize tools from
empirical processes. These results are shown to be widely applicable in
metric-space valued data analysis. In addition to simulations, we provide two
pertinent examples where these results are important: The consistent estimation
of the location of properly defined extrema in metric-space valued
trajectories, which we illustrate with the problem of locating the age of
minimum connectivity in the brain as obtained from fMRI data; and time warping
for metric-space valued trajectories, illustrated with yearly mortality
distributions across countries.
| stat |
Particle-based adaptive-lag online marginal smoothing in general
state-space models | We present a novel algorithm, an adaptive-lag smoother, approximating
efficiently, in an online fashion, sequences of expectations under the marginal
smoothing distributions in general state-space models. The algorithm evolves
recursively a bank of estimators, one for each marginal, in resemblance with
the so-called particle-based, rapid incremental smoother (PaRIS). Each
estimator is propagated until a stopping criterion, measuring the fluctuations
of the estimates, is met. The presented algorithm is furnished with theoretical
results describing its asymptotic limit and memory usage.
| stat |
TULIP: A Toolbox for Linear Discriminant Analysis with Penalties | Linear discriminant analysis (LDA) is a powerful tool in building classifiers
with easy computation and interpretation. Recent advancements in science
technology have led to the popularity of datasets with high dimensions, high
orders and complicated structure. Such datasetes motivate the generalization of
LDA in various research directions. The R package TULIP integrates several
popular high-dimensional LDA-based methods and provides a comprehensive and
user-friendly toolbox for linear, semi-parametric and tensor-variate
classification. Functions are included for model fitting, cross validation and
prediction. In addition, motivated by datasets with diverse sources of
predictors, we further include functions for covariate adjustment. Our package
is carefully tailored for low storage and high computation efficiency.
Moreover, our package is the first R package for many of these methods,
providing great convenience to researchers in this area.
| stat |
Uncertainty Quantification in Extreme Learning Machine: Analytical
Developments, Variance Estimates and Confidence Intervals | Uncertainty quantification is crucial to assess prediction quality of a
machine learning model. In the case of Extreme Learning Machines (ELM), most
methods proposed in the literature make strong assumptions on the data, ignore
the randomness of input weights or neglect the bias contribution in confidence
interval estimations. This paper presents novel estimations that overcome these
constraints and improve the understanding of ELM variability. Analytical
derivations are provided under general assumptions, supporting the
identification and the interpretation of the contribution of different
variability sources. Under both homoskedasticity and heteroskedasticity,
several variance estimates are proposed, investigated, and numerically tested,
showing their effectiveness in replicating the expected variance behaviours.
Finally, the feasibility of confidence intervals estimation is discussed by
adopting a critical approach, hence raising the awareness of ELM users
concerning some of their pitfalls. The paper is accompanied with a scikit-learn
compatible Python library enabling efficient computation of all estimates
discussed herein.
| stat |
Integrating Domain Knowledge in Data-driven Earth Observation with
Process Convolutions | The modelling of Earth observation data is a challenging problem, typically
approached by either purely mechanistic or purely data-driven methods.
Mechanistic models encode the domain knowledge and physical rules governing the
system. Such models, however, need the correct specification of all
interactions between variables in the problem and the appropriate
parameterization is a challenge in itself. On the other hand, machine learning
approaches are flexible data-driven tools, able to approximate arbitrarily
complex functions, but lack interpretability and struggle when data is scarce
or in extrapolation regimes. In this paper, we argue that hybrid learning
schemes that combine both approaches can address all these issues efficiently.
We introduce Gaussian process (GP) convolution models for hybrid modelling in
Earth observation (EO) problems. We specifically propose the use of a class of
GP convolution models called latent force models (LFMs) for EO time series
modelling, analysis and understanding. LFMs are hybrid models that incorporate
physical knowledge encoded in differential equations into a multioutput GP
model. LFMs can transfer information across time-series, cope with missing
observations, infer explicit latent functions forcing the system, and learn
parameterizations which are very helpful for system analysis and
interpretability. We consider time series of soil moisture from active (ASCAT)
and passive (SMOS, AMSR2) microwave satellites. We show how assuming a first
order differential equation as governing equation, the model automatically
estimates the e-folding time or decay rate related to soil moisture persistence
and discovers latent forces related to precipitation. The proposed hybrid
methodology reconciles the two main approaches in remote sensing parameter
estimation by blending statistical learning and mechanistic modeling.
| stat |
Profile Entropy: A Fundamental Measure for the Learnability and
Compressibility of Discrete Distributions | The profile of a sample is the multiset of its symbol frequencies. We show
that for samples of discrete distributions, profile entropy is a fundamental
measure unifying the concepts of estimation, inference, and compression.
Specifically, profile entropy a) determines the speed of estimating the
distribution relative to the best natural estimator; b) characterizes the rate
of inferring all symmetric properties compared with the best estimator over any
label-invariant distribution collection; c) serves as the limit of profile
compression, for which we derive optimal near-linear-time block and sequential
algorithms. To further our understanding of profile entropy, we investigate its
attributes, provide algorithms for approximating its value, and determine its
magnitude for numerous structural distribution families.
| stat |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.