title
stringlengths 6
244
| abstract
stringlengths 19
5.09k
| label
stringclasses 10
values |
|---|---|---|
Flexible estimation of the state dwell-time distribution in hidden
semi-Markov models | Hidden semi-Markov models generalise hidden Markov models by explicitly
modelling the time spent in a given state, the so-called dwell time, using some
distribution defined on the natural numbers. While the (shifted) Poisson and
negative binomial distribution provide natural choices for such distributions,
in practice, parametric distributions can lack the flexibility to adequately
model the dwell times. To overcome this problem, a penalised maximum likelihood
approach is proposed that allows for a flexible and data-driven estimation of
the dwell-time distributions without the need to make any distributional
assumption. This approach is suitable for direct modelling purposes or as an
exploratory tool to investigate the latent state dynamics. The feasibility and
potential of the suggested approach is illustrated by modelling muskox
movements in northeast Greenland using GPS tracking data. The proposed method
is implemented in the R-package PHSMM which is available on CRAN.
| stat |
Unsupervised Online Anomaly Detection On Irregularly Sampled Or Missing
Valued Time-Series Data Using LSTM Networks | We study anomaly detection and introduce an algorithm that processes variable
length, irregularly sampled sequences or sequences with missing values. Our
algorithm is fully unsupervised, however, can be readily extended to supervised
or semisupervised cases when the anomaly labels are present as remarked
throughout the paper. Our approach uses the Long Short Term Memory (LSTM)
networks in order to extract temporal features and find the most relevant
feature vectors for anomaly detection. We incorporate the sampling time
information to our model by modulating the standard LSTM model with time
modulation gates. After obtaining the most relevant features from the LSTM, we
label the sequences using a Support Vector Data Descriptor (SVDD) model. We
introduce a loss function and then jointly optimize the feature extraction and
sequence processing mechanisms in an end-to-end manner. Through this joint
optimization, the LSTM extracts the most relevant features for anomaly
detection later to be used in the SVDD, hence completely removes the need for
feature selection by expert knowledge. Furthermore, we provide a training
algorithm for the online setup, where we optimize our model parameters with
individual sequences as the new data arrives. Finally, on real-life datasets,
we show that our model significantly outperforms the standard approaches thanks
to its combination of LSTM with SVDD and joint optimization.
| stat |
A new method to compare the interpretability of rule-based algorithms | Interpretability is becoming increasingly important for predictive model
analysis. Unfortunately, as remarked by many authors, there is still no
consensus regarding this notion. The goal of this article is to propose a
definition of the notion of interpretability that allows comparisons of
rule-based algorithms. This definition consists of three terms, each one being
quantitatively measured with a simple formula: predictivity, stability and
simplicity. While predictivity has been extensively studied to measure the
accuracy of predictive algorithms, stability is based on the Dice-Sorensen
index for comparing two sets of rules generated by an algorithm using two
independent samples. The simplicity is based on the sum of the length of the
rules derived from the predictive model. The new measure for the
interpretability of a rule-based algorithm is a weighted sum of the three terms
mentioned above. We use the new measure to compare the interpretability of
several rule-based algorithms, specifically CART, RuleFit, Node Harvest,
Covering algorithm and SIRUS for the regression case, and CART, PART and RIPPER
for the classification case
| stat |
BLOB : A Probabilistic Model for Recommendation that Combines Organic
and Bandit Signals | A common task for recommender systems is to build a pro le of the interests
of a user from items in their browsing history and later to recommend items to
the user from the same catalog. The users' behavior consists of two parts: the
sequence of items that they viewed without intervention (the organic part) and
the sequences of items recommended to them and their outcome (the bandit part).
In this paper, we propose Bayesian Latent Organic Bandit model (BLOB), a
probabilistic approach to combine the 'or-ganic' and 'bandit' signals in order
to improve the estimation of recommendation quality. The bandit signal is
valuable as it gives direct feedback of recommendation performance, but the
signal quality is very uneven, as it is highly concentrated on the
recommendations deemed optimal by the past version of the recom-mender system.
In contrast, the organic signal is typically strong and covers most items, but
is not always relevant to the recommendation task. In order to leverage the
organic signal to e ciently learn the bandit signal in a Bayesian model we
identify three fundamental types of distances, namely action-history,
action-action and history-history distances. We implement a scalable
approximation of the full model using variational auto-encoders and the local
re-paramerization trick. We show using extensive simulation studies that our
method out-performs or matches the value of both state-of-the-art organic-based
recommendation algorithms, and of bandit-based methods (both value and
policy-based) both in organic and bandit-rich environments.
| stat |
ET-Lasso: A New Efficient Tuning of Lasso-type Regularization for
High-Dimensional Data | The L1 regularization (Lasso) has proven to be a versatile tool to select
relevant features and estimate the model coefficients simultaneously and has
been widely used in many research areas such as genomes studies, finance, and
biomedical imaging. Despite its popularity, it is very challenging to guarantee
the feature selection consistency of Lasso especially when the dimension of the
data is huge. One way to improve the feature selection consistency is to select
an ideal tuning parameter. Traditional tuning criteria mainly focus on
minimizing the estimated prediction error or maximizing the posterior model
probability, such as cross-validation and BIC, which may either be
time-consuming or fail to control the false discovery rate (FDR) when the
number of features is extremely large. The other way is to introduce
pseudo-features to learn the importance of the original ones. Recently, the
Knockoff filter is proposed to control the FDR when performing feature
selection. However, its performance is sensitive to the choice of the expected
FDR threshold. Motivated by these ideas, we propose a new method using
pseudo-features to obtain an ideal tuning parameter. In particular, we present
the Efficient Tuning of Lasso (ET-Lasso) to separate active and inactive
features by adding permuted features as pseudo-features in linear models. The
pseudo-features are constructed to be inactive by nature, which can be used to
obtain a cutoff to select the tuning parameter that separates active and
inactive features. Experimental studies on both simulations and real-world data
applications are provided to show that ET-Lasso can effectively and efficiently
select active features under a wide range of scenarios
| stat |
Ensemble Adversarial Training: Attacks and Defenses | Adversarial examples are perturbed inputs designed to fool machine learning
models. Adversarial training injects such examples into training data to
increase robustness. To scale this technique to large datasets, perturbations
are crafted using fast single-step methods that maximize a linear approximation
of the model's loss. We show that this form of adversarial training converges
to a degenerate global minimum, wherein small curvature artifacts near the data
points obfuscate a linear approximation of the loss. The model thus learns to
generate weak perturbations, rather than defend against strong ones. As a
result, we find that adversarial training remains vulnerable to black-box
attacks, where we transfer perturbations computed on undefended models, as well
as to a powerful novel single-step attack that escapes the non-smooth vicinity
of the input data via a small random step. We further introduce Ensemble
Adversarial Training, a technique that augments training data with
perturbations transferred from other models. On ImageNet, Ensemble Adversarial
Training yields models with strong robustness to black-box attacks. In
particular, our most robust model won the first round of the NIPS 2017
competition on Defenses against Adversarial Attacks. However, subsequent work
found that more elaborate black-box attacks could significantly enhance
transferability and reduce the accuracy of our models.
| stat |
Valid Inference Corrected for Outlier Removal | Ordinary least square (OLS) estimation of a linear regression model is
well-known to be highly sensitive to outliers. It is common practice to (1)
identify and remove outliers by looking at the data and (2) to fit OLS and form
confidence intervals and p-values on the remaining data as if this were the
original data collected. This standard "detect-and-forget" approach has been
shown to be problematic, and in this paper we highlight the fact that it can
lead to invalid inference and show how recently developed tools in selective
inference can be used to properly account for outlier detection and removal.
Our inferential procedures apply to a general class of outlier removal
procedures that includes several of the most commonly used approaches. We
conduct simulations to corroborate the theoretical results, and we apply our
method to three real data sets to illustrate how our inferential results can
differ from the traditional detect-and-forget strategy. A companion R package,
outference, implements these new procedures with an interface that matches the
functions commonly used for inference with lm in R.
| stat |
PAC-Bayes and Domain Adaptation | We provide two main contributions in PAC-Bayesian theory for domain
adaptation where the objective is to learn, from a source distribution, a
well-performing majority vote on a different, but related, target distribution.
Firstly, we propose an improvement of the previous approach we proposed in
Germain et al. (2013), which relies on a novel distribution pseudodistance
based on a disagreement averaging, allowing us to derive a new tighter domain
adaptation bound for the target risk. While this bound stands in the spirit of
common domain adaptation works, we derive a second bound (introduced in Germain
et al., 2016) that brings a new perspective on domain adaptation by deriving an
upper bound on the target risk where the distributions' divergence-expressed as
a ratio-controls the trade-off between a source error measure and the target
voters' disagreement. We discuss and compare both results, from which we obtain
PAC-Bayesian generalization bounds. Furthermore, from the PAC-Bayesian
specialization to linear classifiers, we infer two learning algorithms, and we
evaluate them on real data.
| stat |
A Systematic Approach to Surveillance and Detection of Hierarchical
Healthcare Cost Drivers and Utilization Offsets | There is strong interest among healthcare payers to identify emerging
healthcare cost drivers to support early intervention. However, many challenges
arise in analyzing large, high dimensional, and noisy healthcare data. In this
paper, we propose a systematic approach that utilizes hierarchical search
strategies and enhanced statistical process control (SPC) algorithms to surface
high impact cost drivers. Our approach aims to provide interpretable, detailed,
and actionable insights of detected change patterns attributing to multiple
clinical factors. We also proposed an algorithm to identify comparable
treatment offsets at the population level and quantify the cost impact on their
utilization changes. To illustrate our approach, we apply it to the IBM Watson
Health MarketScan Commercial Database and organized the detected emerging
drivers into 5 categories for reporting. We also discuss some findings in this
analysis and potential actions in mitigating the impact of the drivers.
| stat |
Scaling up Kernel Ridge Regression via Locality Sensitive Hashing | Random binning features, introduced in the seminal paper of Rahimi and Recht
(2007), are an efficient method for approximating a kernel matrix using
locality sensitive hashing. Random binning features provide a very simple and
efficient way of approximating the Laplace kernel but unfortunately do not
apply to many important classes of kernels, notably ones that generate smooth
Gaussian processes, such as the Gaussian kernel and Matern kernel. In this
paper, we introduce a simple weighted version of random binning features and
show that the corresponding kernel function generates Gaussian processes of any
desired smoothness. We show that our weighted random binning features provide a
spectral approximation to the corresponding kernel matrix, leading to efficient
algorithms for kernel ridge regression. Experiments on large scale regression
datasets show that our method outperforms the accuracy of random Fourier
features method.
| stat |
Direct semi-parametric estimation of the state price density implied in
option prices | We present a model for direct semi-parametric estimation of the State Price
Density (SPD) implied in quoted option prices. We treat the observed prices as
expected values of possible pay-offs at maturity, weighted by the unknown
probability density function. We model the logarithm of the latter as a smooth
function while matching the expected values of the potential pay-offs with the
observed prices. This leads to a special case of the penalized composite link
model. Our estimates do not rely on any parametric assumption on the underlying
asset price dynamics and are consistent with no-arbitrage conditions. The model
shows excellent performance in simulations and in application to real data.
| stat |
Slice Sampling for General Completely Random Measures | Completely random measures provide a principled approach to creating flexible
unsupervised models, where the number of latent features is infinite and the
number of features that influence the data grows with the size of the data set.
Due to the infinity the latent features, posterior inference requires either
marginalization---resulting in dependence structures that prevent efficient
computation via parallelization and conjugacy---or finite truncation, which
arbitrarily limits the flexibility of the model. In this paper we present a
novel Markov chain Monte Carlo algorithm for posterior inference that
adaptively sets the truncation level using auxiliary slice variables, enabling
efficient, parallelized computation without sacrificing flexibility. In
contrast to past work that achieved this on a model-by-model basis, we provide
a general recipe that is applicable to the broad class of completely random
measure-based priors. The efficacy of the proposed algorithm is evaluated on
several popular nonparametric models, demonstrating a higher effective sample
size per second compared to algorithms using marginalization as well as a
higher predictive performance compared to models employing fixed truncations.
| stat |
A Normal Approximation Method for Statistics in Knockouts | The authors give an approximation method for Bayesian inference in arena
model, which is focused on paired comparisons with eliminations and
bifurcations. The approximation method simplifies the inference by reducing
parameters and introducing normal distribution functions into the computation
of posterior distribution, which is largely based on an important property of
normal random variables. Maximum a posteriori probability (MAP) and Bayesian
prediction are then used to mine the information from the past pairwise
comparison data, such as an individual's strength or volatility and his
possible future results. We conduct a simulation to show the accuracy and
stability of the approximation method and demonstrate the algorithm on
nonlinear parameter inference as well as prediction problem arising in the FIFA
World Cup.
| stat |
Backward importance sampling for online estimation of state space models | This paper proposes a new Sequential Monte Carlo algorithm to perform online
estimation in the context of state space models when either the transition
density of the latent state or the conditional likelihood of an observation
given a state is intractable. In this setting, obtaining low variance
estimators of expectations under the posterior distributions of the unobserved
states given the observations is a challenging task. Following recent
theoretical results for pseudo-marginal sequential Monte Carlo smoothers, a
pseudo-marginal backward importance sampling step is introduced to estimate
such expectations. This new step allows to reduce very significantly the
computational time of the existing numerical solutions based on an
acceptance-rejection procedure for similar performance, and to broaden the
class of eligible models for such methods. For instance, in the context of
multivariate stochastic differential equations, the proposed algorithm makes
use of unbiased estimates of the unknown transition densities under much weaker
assumptions than standard alternatives. The performance of this estimator is
assessed for high-dimensional discrete-time latent data models, for recursive
maximum likelihood estimation in the context of partially observed diffusion
process, and in the case of a bidimensional partially observed stochastic
Lotka-Volterra model.
| stat |
A Framework for Causal Discovery in non-intervenable systems | Many frameworks exist to infer cause and effect relations in complex
nonlinear systems but a complete theory is lacking. A new framework is
presented that is fully nonlinear, provides a complete information theoretic
disentanglement of causal processes, allows for nonlinear interactions between
causes, identifies the causal strength of missing or unknown processes, and can
analyze systems that cannot be represented on standard graphs. The basic
building blocks are information theoretic measures such as (conditional) mutual
information and a new concept called certainty that monotonically increases
with the information available about the target process. The framework is
presented in detail and compared with other existing frameworks, and the
treatment of confounders is discussed. It is tested on several highly
simplified stochastic processes to demonstrate how blocking and gateways are
handled, and on the chaotic Lorentz 1963 system. It is shown that the framework
provides information on the local dynamics, but also reveals information on the
larger scale structure of the underlying attractor. While there are systems
with structures that the framework cannot disentangle, it is argued that any
causal framework that is based on integrated quantities will miss out
potentially important information of the underlying probability density
functions.
| stat |
A Machine Learning Approach to Improving Occupational Income Scores | Historical studies of labor markets frequently lack data on individual
income. The occupational income score (OCCSCORE) is often used as an
alternative measure of labor market outcomes. We consider the consequences of
using OCCSCORE when researchers are interested in earnings regressions. We
estimate race and gender earnings gaps in modern decennial Censuses as well as
the 1915 Iowa State Census. Using OCCSCORE biases results towards zero and can
result in estimated gaps of the wrong sign. We use a machine learning approach
to construct a new adjusted score based on industry, occupation, and
demographics. The new income score provides estimates closer to earnings
regressions. Lastly, we consider the consequences for estimates of
intergenerational mobility elasticities.
| stat |
Decision-Making with Auto-Encoding Variational Bayes | To make decisions based on a model fit with auto-encoding variational Bayes
(AEVB), practitioners often let the variational distribution serve as a
surrogate for the posterior distribution. This approach yields biased estimates
of the expected risk, and therefore leads to poor decisions for two reasons.
First, the model fit with AEVB may not equal the underlying data distribution.
Second, the variational distribution may not equal the posterior distribution
under the fitted model. We explore how fitting the variational distribution
based on several objective functions other than the ELBO, while continuing to
fit the generative model based on the ELBO, affects the quality of downstream
decisions. For the probabilistic principal component analysis model, we
investigate how importance sampling error, as well as the bias of the model
parameter estimates, varies across several approximate posteriors when used as
proposal distributions. Our theoretical results suggest that a posterior
approximation distinct from the variational distribution should be used for
making decisions. Motivated by these theoretical results, we propose learning
several approximate proposals for the best model and combining them using
multiple importance sampling for decision-making. In addition to toy examples,
we present a full-fledged case study of single-cell RNA sequencing. In this
challenging instance of multiple hypothesis testing, our proposed approach
surpasses the current state of the art.
| stat |
Assessing the contagiousness of mass shootings with nonparametric Hawkes
processes | Gun violence and mass shootings are high-profile epidemiological issues
facing the United States with questions regarding their contagiousness gaining
prevalence in news media. Through the use of nonparametric Hawkes processes, we
examine the evidence for the existence of contagiousness within a catalog of
mass shootings and highlight the broader benefits of using such nonparametric
point process models in modeling the occurrence of such events.
| stat |
Gradient-based Bayesian Experimental Design for Implicit Models using
Mutual Information Lower Bounds | We introduce a framework for Bayesian experimental design (BED) with implicit
models, where the data-generating distribution is intractable but sampling from
it is still possible. In order to find optimal experimental designs for such
models, our approach maximises mutual information lower bounds that are
parametrised by neural networks. By training a neural network on sampled data,
we simultaneously update network parameters and designs using stochastic
gradient-ascent. The framework enables experimental design with a variety of
prominent lower bounds and can be applied to a wide range of scientific tasks,
such as parameter estimation, model discrimination and improving future
predictions. Using a set of intractable toy models, we provide a comprehensive
empirical comparison of prominent lower bounds applied to the aforementioned
tasks. We further validate our framework on a challenging system of stochastic
differential equations from epidemiology.
| stat |
Bayesian inference, model selection and likelihood estimation using fast
rejection sampling: the Conway-Maxwell-Poisson distribution | Bayesian inference for models with intractable likelihood functions
represents a challenging suite of problems in modern statistics. In this work
we analyse the Conway-Maxwell-Poisson (COM-Poisson) distribution, a two
parameter generalisation of the Poisson distribution. COM-Poisson regression
modelling allows the flexibility to model dispersed count data as part of a
generalised linear model (GLM) with a COM-Poisson response, where exogenous
covariates control the mean and dispersion level of the response. The major
difficulty with COM-Poisson regression is that the likelihood function contains
multiple intractable normalising constants and is not amenable to standard
inference and MCMC techniques. Recent work by Chanialidis et al. (2017) has
seen the development of a sampler to draw random variates from the COM-Poisson
likelihood using a rejection sampling algorithm. We provide a new rejection
sampler for the COM-Poisson distribution which significantly reduces the CPU
time required to perform inference for COM-Poisson regression models. A novel
extension of this work shows that for any intractable likelihood function with
an associated rejection sampler it is possible to construct unbiased estimators
of the intractable likelihood which proves useful for model selection or for
use within pseudo-marginal MCMC algorithms (Andrieu and Roberts, 2009). We
demonstrate all of these methods on a real-world dataset of takeover bids.
| stat |
$S^{2}$-LBI: Stochastic Split Linearized Bregman Iterations for
Parsimonious Deep Learning | This paper proposes a novel Stochastic Split Linearized Bregman Iteration
($S^{2}$-LBI) algorithm to efficiently train the deep network. The $S^{2}$-LBI
introduces an iterative regularization path with structural sparsity. Our
$S^{2}$-LBI combines the computational efficiency of the LBI, and model
selection consistency in learning the structural sparsity. The computed
solution path intrinsically enables us to enlarge or simplify a network, which
theoretically, is benefited from the dynamics property of our $S^{2}$-LBI
algorithm. The experimental results validate our $S^{2}$-LBI on MNIST and
CIFAR-10 dataset. For example, in MNIST, we can either boost a network with
only 1.5K parameters (1 convolutional layer of 5 filters, and 1 FC layer),
achieves 98.40\% recognition accuracy; or we simplify $82.5\%$ of parameters in
LeNet-5 network, and still achieves the 98.47\% recognition accuracy. In
addition, we also have the learning results on ImageNet, which will be added in
the next version of our report.
| stat |
Estimating the real burden of disease under a pandemic situation: The
SARS-CoV2 case | The present paper introduces a new model used to study and analyse the severe
acute respiratory syndrome coronavirus 2 (SARS-CoV2) epidemic-reported-data
from Spain. This is a Hidden Markov Model whose hidden layer is a regeneration
process with Poisson immigration, Po-INAR(1), together with a mechanism that
allows the estimation of the under-reporting in non-stationary count time
series. A novelty of the model is that the expectation of the innovations in
the unobserved process is a time-dependent function defined in such a way that
information about the spread of an epidemic, as modelled through a
Susceptible-Infectious-Removed dynamical system, is incorporated into the
model. In addition, the parameter controlling the intensity of the
under-reporting is also made to vary with time to adjust to possible
seasonality or trend in the data. Maximum likelihood methods are used to
estimate the parameters of the model.
| stat |
Time Series Forecasting With Deep Learning: A Survey | Numerous deep learning architectures have been developed to accommodate the
diversity of time series datasets across different domains. In this article, we
survey common encoder and decoder designs used in both one-step-ahead and
multi-horizon time series forecasting -- describing how temporal information is
incorporated into predictions by each model. Next, we highlight recent
developments in hybrid deep learning models, which combine well-studied
statistical models with neural network components to improve pure methods in
either category. Lastly, we outline some ways in which deep learning can also
facilitate decision support with time series data.
| stat |
Is the NUTS algorithm correct? | This paper is devoted to investigate whether the popular No U-turn (NUTS)
sampling algorithm is correct, i.e.\ whether the target probability
distribution is \emph{exactly} conserved by the algorithm. It turns out that
one of the Gibbs substeps used in the algorithm cannot always be guaranteed to
be correct.
| stat |
Quantifying spatio-temporal boundary condition uncertainty for the North
American deglaciation | Ice sheet models are used to study the deglaciation of North America at the
end of the last ice age (past 21,000 years), so that we might understand
whether and how existing ice sheets may reduce or disappear under climate
change. Though ice sheet models have a few parameters controlling physical
behaviour of the ice mass, they also require boundary conditions for climate
(spatio-temporal fields of temperature and precipitation, typically on regular
grids and at monthly intervals). The behaviour of the ice sheet is highly
sensitive to these fields, and there is relatively little data from geological
records to constrain them as the land was covered with ice. We develop a
methodology for generating a range of plausible boundary conditions, using a
low-dimensional basis representation of the spatio-temporal input. We derive
this basis by combining key patterns, extracted from a small ensemble of
climate model simulations of the deglaciation, with sparse spatio-temporal
observations. By jointly varying the ice sheet parameters and basis vector
coefficients, we run ensembles of the Glimmer ice sheet model that
simultaneously explore both climate and ice sheet model uncertainties. We use
these to calibrate the ice sheet physics and boundary conditions for Glimmer,
by ruling out regions of the joint coefficient and parameter space via history
matching. We use binary ice/no ice observations from reconstructions of past
ice sheet margin position to constrain this space by introducing a novel metric
for history matching to binary data.
| stat |
Pathwise Derivatives for Multivariate Distributions | We exploit the link between the transport equation and derivatives of
expectations to construct efficient pathwise gradient estimators for
multivariate distributions. We focus on two main threads. First, we use null
solutions of the transport equation to construct adaptive control variates that
can be used to construct gradient estimators with reduced variance. Second, we
consider the case of multivariate mixture distributions. In particular we show
how to compute pathwise derivatives for mixtures of multivariate Normal
distributions with arbitrary means and diagonal covariances. We demonstrate in
a variety of experiments in the context of variational inference that our
gradient estimators can outperform other methods, especially in high
dimensions.
| stat |
A Mechanistic Model of Annual Sulfate Concentrations in the United
States | We develop a mechanistic model to analyze the impact of sulfur dioxide
emissions from coal-fired power plants on average sulfate concentrations in the
central United States. A multivariate Ornstein-Uhlenbeck (OU) process is used
to approximate the dynamics of the underlying space-time chemical transport
process, and its distributional properties are leveraged to specify novel
probability models for spatial data (i.e., spatially-referenced data with no
temporal replication) that are viewed as either a snapshot or a time-averaged
observation of the OU process. Air pollution transport dynamics determine the
mean and covariance structure of our atmospheric sulfate model, allowing us to
infer which process dynamics are driving observed air pollution concentrations.
We use these inferred dynamics to assess the regulatory impact of flue-gas
desulfurization (FGD) technologies on human exposure to sulfate aerosols.
| stat |
Orthogonal Structure Search for Efficient Causal Discovery from
Observational Data | The problem of inferring the direct causal parents of a response variable
among a large set of explanatory variables is of high practical importance in
many disciplines. Recent work exploits stability of regression coefficients or
invariance properties of models across different experimental conditions for
reconstructing the full causal graph. These approaches generally do not scale
well with the number of the explanatory variables and are difficult to extend
to nonlinear relationships. Contrary to existing work, we propose an approach
which even works for observational data alone, while still offering theoretical
guarantees including the case of partially nonlinear relationships. Our
algorithm requires only one estimation for each variable and in our experiments
we apply our causal discovery algorithm even to large graphs, demonstrating
significant improvements compared to well established approaches.
| stat |
A Smoother Way to Train Structured Prediction Models | We present a framework to train a structured prediction model by performing
smoothing on the inference algorithm it builds upon. Smoothing overcomes the
non-smoothness inherent to the maximum margin structured prediction objective,
and paves the way for the use of fast primal gradient-based optimization
algorithms. We illustrate the proposed framework by developing a novel primal
incremental optimization algorithm for the structural support vector machine.
The proposed algorithm blends an extrapolation scheme for acceleration and an
adaptive smoothing scheme and builds upon the stochastic variance-reduced
gradient algorithm. We establish its worst-case global complexity bound and
study several practical variants, including extensions to deep structured
prediction. We present experimental results on two real-world problems, namely
named entity recognition and visual object localization. The experimental
results show that the proposed framework allows us to build upon efficient
inference algorithms to develop large-scale optimization algorithms for
structured prediction which can achieve competitive performance on the two
real-world problems.
| stat |
Analyzing count data using a time series model with an exponentially
decaying covariance structure | Count data appears in various disciplines. In this work, a new method to
analyze time series count data has been proposed. The method assumes
exponentially decaying covariance structure, a special class of the Mat\'ern
covariance function, for the latent variable in a Poisson regression model. It
is implemented in a Bayesian framework, with the help of Gibbs sampling and
ARMS sampling techniques. The proposed approach provides reliable estimates for
the covariate effects and estimates the extent of variability explained by the
temporally dependent process and the white noise process. The method is
flexible, allows irregular spaced data, and can be extended naturally to bigger
datasets. The Bayesian implementation helps us to compute the posterior
predictive distribution and hence is more appropriate and attractive for count
data forecasting problems. Two real life applications of different flavors are
included in the paper. These two examples and a short simulation study
establish that the proposed approach has good inferential and predictive
abilities and performs better than the other competing models.
| stat |
A stable and adaptive polygenic signal detection method based on
repeated sample splitting | Focusing on polygenic signal detection in high dimensional genetic
association studies of complex traits, we develop an adaptive test for
generalized linear models to accommodate different alternatives. To facilitate
valid post-selection inference for high dimensional data, our study here
adheres to the original sampling-splitting principle but does so, repeatedly,
to increase stability of the inference. We show the asymptotic null
distributions of the proposed test for both fixed and diverging number of
variants. We also show the asymptotic properties of the proposed test under
local alternatives, providing insights on why power gain attributed to variable
selection and weighting can compensate for efficiency loss due to sample
splitting. We support our analytical findings through extensive simulation
studies and two applications. The proposed procedure is computationally
efficient and has been implemented as the R package DoubleCauchy.
| stat |
Regularized spectral methods for clustering signed networks | We study the problem of $k$-way clustering in signed graphs. Considerable
attention in recent years has been devoted to analyzing and modeling signed
graphs, where the affinity measure between nodes takes either positive or
negative values. Recently, Cucuringu et al. [CDGT 2019] proposed a spectral
method, namely SPONGE (Signed Positive over Negative Generalized Eigenproblem),
which casts the clustering task as a generalized eigenvalue problem optimizing
a suitably defined objective function. This approach is motivated by social
balance theory, where the clustering task aims to decompose a given network
into disjoint groups, such that individuals within the same group are connected
by as many positive edges as possible, while individuals from different groups
are mainly connected by negative edges. Through extensive numerical
simulations, SPONGE was shown to achieve state-of-the-art empirical
performance. On the theoretical front, [CDGT 2019] analyzed SPONGE and the
popular Signed Laplacian method under the setting of a Signed Stochastic Block
Model (SSBM), for $k=2$ equal-sized clusters, in the regime where the graph is
moderately dense.
In this work, we build on the results in [CDGT 2019] on two fronts for the
normalized versions of SPONGE and the Signed Laplacian. Firstly, for both
algorithms, we extend the theoretical analysis in [CDGT 2019] to the general
setting of $k \geq 2$ unequal-sized clusters in the moderately dense regime.
Secondly, we introduce regularized versions of both methods to handle sparse
graphs -- a regime where standard spectral methods underperform -- and provide
theoretical guarantees under the same SSBM model. To the best of our knowledge,
regularized spectral methods have so far not been considered in the setting of
clustering signed graphs. We complement our theoretical results with an
extensive set of numerical experiments on synthetic data.
| stat |
Gradient-Based Markov Chain Monte Carlo for Bayesian Inference With
Non-Differentiable Priors | The use of non-differentiable priors in Bayesian statistics has become
increasingly popular, in particular in Bayesian imaging analysis. Current state
of the art methods are approximate in the sense that they replace the posterior
with a smooth approximation via Moreau-Yosida envelopes, and apply
gradient-based discretized diffusions to sample from the resulting
distribution. We characterize the error of the Moreau-Yosida approximation and
propose a novel implementation using underdamped Langevin dynamics. In
misson-critical cases, however, replacing the posterior with an approximation
may not be a viable option. Instead, we show that Piecewise-Deterministic
Markov Processes (PDMP) can be utilized for exact posterior inference from
distributions satisfying almost everywhere differentiability. Furthermore, in
contrast with diffusion-based methods, the suggested PDMP-based samplers place
no assumptions on the prior shape, nor require access to a computationally
cheap proximal operator, and consequently have a much broader scope of
application. Through detailed numerical examples, including a
non-differentiable circular distribution and a non-convex genomics model, we
elucidate the relative strengths of these sampling methods on problems of
moderate to high dimensions, underlining the benefits of PDMP-based methods
when accurate sampling is decisive.
| stat |
Prevalence Threshold and the Geometry of Screening Curves | The relationship between a screening tests' positive predictive value,
$\rho$, and its target prevalence, $\phi$, is proportional - though not linear
in all but a special case. In consequence, there is a point of local extrema of
curvature defined only as a function of the sensitivity $a$ and specificity $b$
beyond which the rate of change of a test's $\rho$ drops precipitously relative
to $\phi$. Herein, we show the mathematical model exploring this phenomenon and
define the $prevalence$ $threshold$ ($\phi_e$) point where this change occurs
as:
$\phi_e=\frac{\sqrt{a\left(-b+1\right)}+b-1}{(\varepsilon-1)}$ where
$\varepsilon$ = $a$+$b$.
From the prevalence threshold we deduce a more generalized relationship
between prevalence and positive predictive value as a function of
$\varepsilon$, which represents a fundamental theorem of screening, herein
defined as:
$\displaystyle\lim_{\varepsilon \to 2}{\displaystyle
\int_{0}^{1}}{\rho(\phi)d\phi} = 1$
Understanding the concepts described in this work can help contextualize the
validity of screening tests in real time, and help guide the interpretation of
different clinical scenarios in which screening is undertaken.
| stat |
A Maximum Entropy Procedure to Solve Likelihood Equations | In this article we provide initial findings regarding the problem of solving
likelihood equations by means of a maximum entropy approach. Unlike standard
procedures that require equating at zero the score function of the
maximum-likelihood problem, we propose an alternative strategy where the score
is instead used as external informative constraint to the maximization of the
convex Shannon's entropy function. The problem involves the re-parameterization
of the score parameters as expected values of discrete probability
distributions where probabilities need to be estimated. This leads to a simpler
situation where parameters are searched in smaller (hyper) simplex space. We
assessed our proposal by means of empirical case studies and a simulation
study, this latter involving the most critical case of logistic regression
under data separation. The results suggested that the maximum entropy
re-formulation of the score problem solves the likelihood equation problem.
Similarly, when maximum-likelihood estimation is difficult, as for the case of
logistic regression under separation, the maximum entropy proposal achieved
results (numerically) comparable to those obtained by the Firth's
Bias-corrected approach. Overall, these first findings reveal that a maximum
entropy solution can be considered as an alternative technique to solve the
likelihood equation.
| stat |
MCMC assisted by Belief Propagation | Markov Chain Monte Carlo (MCMC) and Belief Propagation (BP) are the most
popular algorithms for computational inference in Graphical Models (GM). In
principle, MCMC is an exact probabilistic method which, however, often suffers
from exponentially slow mixing. In contrast, BP is a deterministic method,
which is typically fast, empirically very successful, however in general
lacking control of accuracy over loopy graphs. In this paper, we introduce MCMC
algorithms correcting the approximation error of BP, i.e., we provide a way to
compensate for BP errors via a consecutive BP-aware MCMC. Our framework is
based on the Loop Calculus (LC) approach which allows expressing the BP error
as a sum of weighted generalized loops. Although the full series is
computationally intractable, it is known that a truncated series, summing up
all 2-regular loops, is computable in polynomial-time for planar pair-wise
binary GMs and it also provides a highly accurate approximation empirically.
Motivated by this, we first propose a polynomial-time approximation MCMC scheme
for the truncated series of general (non-planar) pair-wise binary models. Our
main idea here is to use the Worm algorithm, known to provide fast mixing in
other (related) problems, and then design an appropriate rejection scheme to
sample 2-regular loops. Furthermore, we also design an efficient rejection-free
MCMC scheme for approximating the full series. The main novelty underlying our
design is in utilizing the concept of cycle basis, which provides an efficient
decomposition of the generalized loops. In essence, the proposed MCMC schemes
run on transformed GM built upon the non-trivial BP solution, and our
experiments show that this synthesis of BP and MCMC outperforms both direct
MCMC and bare BP schemes.
| stat |
Propensity Score Matching underestimates Real Treatment Effect, in a
simulated theoretical multivariate model | Propensity Score Matching (PSM) is an useful method to reduce the impact
ofTreatment - Selection Bias in the estimation of causal effects in
observational studies. After matching, the PSM significantly reduces the sample
under investigation, which may lead to other possible biases. In this sense, we
want to analyse the behaviour of this PSM compared with other widely used
method to deal with non-comparable groups, as is the Multivariate Regression
Model (MRM). Monte Carlo Simulations are made to construct groups with
different effects in order to compare the behaviour of PSM and MRM estimating
this effects. Also the Treatment - Selection Bias reduction for the PSM is
calculated. With the PSM a reduction in the Treatment - Selection Bias is
achieved, with a reduction in the Relative Real Treatment Effect Estimation
Error, but despite of this bias reduction and estimation error reduction, the
MRM significantly reduces more this estimation error compared with the PSM.
Also the PSM leads to a not insignificant reduction of the sample. This loss of
information derived from the matching process may lead to another not known
bias, and thus, to the inaccurate of the effect estimation compared with the
MRM.
| stat |
User Sentiment as a Success Metric: Persistent Biases Under Full
Randomization | We study user sentiment (reported via optional surveys) as a metric for fully
randomized A/B tests. Both user-level covariates and treatment assignment can
impact response propensity. We propose a set of consistent estimators for the
average and local treatment effects on treated and respondent users. We show
that our problem can be mapped onto the intersection of the missing data
problem and observational causal inference, and we identify conditions under
which consistent estimators exist. We evaluate the performance of estimators
via simulation studies and find that more complicated models do not necessarily
provide superior performance.
| stat |
Ramifications of Approximate Posterior Inference for Bayesian Deep
Learning in Adversarial and Out-of-Distribution Settings | Deep neural networks have been successful in diverse discriminative
classification tasks, although, they are poorly calibrated often assigning high
probability to misclassified predictions. Potential consequences could lead to
trustworthiness and accountability of the models when deployed in real
applications, where predictions are evaluated based on their confidence scores.
Existing solutions suggest the benefits attained by combining deep neural
networks and Bayesian inference to quantify uncertainty over the models'
predictions for ambiguous datapoints. In this work we propose to validate and
test the efficacy of likelihood based models in the task of out of distribution
detection (OoD). Across different datasets and metrics we show that Bayesian
deep learning models on certain occasions marginally outperform conventional
neural networks and in the event of minimal overlap between in/out distribution
classes, even the best models exhibit a reduction in AUC scores in detecting
OoD data. Preliminary investigations indicate the potential inherent role of
bias due to choices of initialisation, architecture or activation functions. We
hypothesise that the sensitivity of neural networks to unseen inputs could be a
multi-factor phenomenon arising from the different architectural design choices
often amplified by the curse of dimensionality. Furthermore, we perform a study
to find the effect of the adversarial noise resistance methods on in and
out-of-distribution performance, as well as, also investigate adversarial noise
robustness of Bayesian deep learners.
| stat |
Importance subsampling: improving power system planning under
climate-based uncertainty | Recent studies indicate that the effects of inter-annual climate-based
variability in power system planning are significant and that long samples of
demand & weather data (spanning multiple decades) should be considered. At the
same time, modelling renewable generation such as solar and wind requires high
temporal resolution to capture fluctuations in output levels. In many realistic
power system models, using long samples at high temporal resolution is
computationally unfeasible. This paper introduces a novel subsampling approach,
referred to as "importance subsampling", allowing the use of multiple decades
of demand & weather data in power system planning models at reduced
computational cost. The methodology can be applied in a wide class of
optimisation-based power system simulations. A test case is performed on a
model of the United Kingdom created using the open-source modelling framework
Calliope and 36 years of hourly demand and wind data. Standard data reduction
approaches such as using individual years or clustering into representative
days lead to significant errors in estimates of optimal system design.
Furthermore, the resultant power systems lead to supply capacity shortages,
raising questions of generation capacity adequacy. In contrast, "importance
subsampling" leads to accurate estimates of optimal system design at greatly
reduced computational cost, with resultant power systems able to meet demand
across all 36 years of demand & weather scenarios.
| stat |
Functional Isolation Forest | For the purpose of monitoring the behavior of complex infrastructures (e.g.
aircrafts, transport or energy networks), high-rate sensors are deployed to
capture multivariate data, generally unlabeled, in quasi continuous-time to
detect quickly the occurrence of anomalies that may jeopardize the smooth
operation of the system of interest. The statistical analysis of such massive
data of functional nature raises many challenging methodological questions. The
primary goal of this paper is to extend the popular Isolation Forest (IF)
approach to Anomaly Detection, originally dedicated to finite dimensional
observations, to functional data. The major difficulty lies in the wide variety
of topological structures that may equip a space of functions and the great
variety of patterns that may characterize abnormal curves. We address the issue
of (randomly) splitting the functional space in a flexible manner in order to
isolate progressively any trajectory from the others, a key ingredient to the
efficiency of the algorithm. Beyond a detailed description of the algorithm,
computational complexity and stability issues are investigated at length. From
the scoring function measuring the degree of abnormality of an observation
provided by the proposed variant of the IF algorithm, a Functional Statistical
Depth function is defined and discussed as well as a multivariate functional
extension. Numerical experiments provide strong empirical evidence of the
accuracy of the extension proposed.
| stat |
A case for new neural network smoothness constraints | How sensitive should machine learning models be to input changes? We tackle
the question of model smoothness and show that it is a useful inductive bias
which aids generalization, adversarial robustness, generative modeling and
reinforcement learning. We explore current methods of imposing smoothness
constraints and observe they lack the flexibility to adapt to new tasks, they
don't account for data modalities, they interact with losses, architectures and
optimization in ways not yet fully understood. We conclude that new advances in
the field are hinging on finding ways to incorporate data, tasks and learning
into our definitions of smoothness.
| stat |
R-miss-tastic: a unified platform for missing values methods and
workflows | Missing values are unavoidable when working with data. Their occurrence is
exacerbated as more data from different sources become available. However, most
statistical models and visualization methods require complete data, and
improper handling of missing data results in information loss, or biased
analyses. Since the seminal work of Rubin (1976), there has been a burgeoning
literature on missing values with heterogeneous aims and motivations. This has
resulted in the development of various methods, formalizations, and tools
(including a large number of R packages and Python modules). However, for
practitioners, it remains challenging to decide which method is most suited for
their problem, partially because handling missing data is still not a topic
systematically covered in statistics or data science curricula.
To help address this challenge, we have launched a unified platform:
"R-miss-tastic", which aims to provide an overview of standard missing values
problems, methods, how to handle them in analyses, and relevant implementations
of methodologies. In the same perspective, we have also developed several
pipelines in R and Python to allow for a hands-on illustration of how to handle
missing values in various statistical tasks such as estimation and prediction,
while ensuring reproducibility of the analyses. This will hopefully also
provide some guidance on deciding which method to choose for a specific problem
and data. The objective of this work is not only to comprehensively organize
materials, but also to create standardized analysis workflows, and to provide a
common ground for discussions among the community. This platform is thus suited
for beginners, students, more advanced analysts and researchers.
| stat |
Consistency of Extreme Learning Machines and Regression under
Non-Stationary Dependence for ML-Enhanced Moving Objects | Supervised learning by extreme learning machines resp. neural networks with
random weights is studied under a non-stationary spatial-temporal sampling
design which especially addresses settings where an autonomous object moving in
a non-stationary spatial environment collects and analyzes data. The stochastic
model especially allows for spatial heterogeneity and weak dependence. As
efficient and computationally cheap learning methods (unconstrained) least
squares, ridge regression and $ \ell_r $-penalized least squares (including the
LASSO) is studied. Consistency and asymptotic normality of the least squares
and ridge regression estimates as well as corresponding consistency results for
the $ \ell_r $-penalty are shown under weak conditions. The resuts also cover
bounds for the sample squared predicition error.
| stat |
Post-sampling crowdsourced data to allow reliable statistical inference:
the case of food price indices in Nigeria | Sound policy and decision making in developing countries is often limited by
the lack of timely and reliable data. Crowdsourced data may provide a valuable
alternative for data collection and analysis, e. g. in remote and insecure
areas or of poor accessibility where traditional methods are difficult or
costly. However, crowdsourced data are not directly usable to draw sound
statistical inference. Indeed, its use involves statistical problems because
data do not obey any formal sampling design and may also suffer from various
non-sampling errors. To overcome this, we propose the use of a special form of
post-stratification with which crowdsourced data are reweighted prior their use
in an inferential context. An example in Nigeria illustrates the applicability
of the method.
| stat |
A Bayesian Non-linear State Space Copula Model to Predict Air Pollution
in Beijing | Air pollution is a serious issue that currently affects many industrial
cities in the world and can cause severe illness to the population. In
particular, it has been proven that extreme high levels of airborne
contaminants have dangerous short-term effects on human health, in terms of
increased hospital admissions for cardiovascular and respiratory diseases and
increased mortality risk. For these reasons, accurate estimation and prediction
of airborne pollutant concentration is crucial. In this paper, we propose a
flexible novel approach to model hourly measurements of fine particulate matter
and meteorological data collected in Beijing in 2014. We show that the standard
state space model, based on Gaussian assumptions, does not correctly capture
the time dynamics of the observations. Therefore, we propose a non-linear
non-Gaussian state space model where both the observation and the state
equations are defined by copula specifications, and we perform Bayesian
inference using the Hamiltonian Monte Carlo method. The proposed copula state
space approach is very flexible, since it allows us to separately model the
marginals and to accommodate a wide variety of dependence structures in the
data dynamics. We show that the proposed approach allows us not only to predict
particulate matter measurements, but also to investigate the effects of user
specified climate scenarios.
| stat |
Spatial variation in the basic reproduction number of COVID-19: A
systematic review | OBJECTIVES: Estimates of the basic reproduction number (R0) of COVID-19 vary
across countries. This paper aims to characterise the spatial variability in R0
across the first six months of the global COVID-19 outbreak, and to explore
social factors that impact R0 estimates at national and regional level.
METHODS: We searched PubMed, LitCOVID and the WHO COVID-19 database from
January to June 2020. Peer-reviewed English-language papers were included that
provided R0 estimates. For each study, the value of the estimate, country under
study and publication month were extracted. The median R0 value was calculated
per country, and the median and variance were calculated per region. For each
country with an R0 estimate, the Human Development Index (HDI), Sustainable
Mobility Index (SMI), median age, population density and development status
were obtained from external sources.
RESULTS: A total of 81 studies were included in the analysis. These studies
provided at least one estimate of R0, along with sufficient methodology to
explain how the value was calculated. Values of R0 ranged between 0.48 and
14.8, and between 0.48 and 6.7 when excluding outliers.
CONCLUSIONS: This systematic review provides a comprehensive overview of the
estimates of the basic reproduction number of COVID-19 globally and highlights
the spatial heterogeneity in R0. Higher values were recorded in more developed
countries, and countries with an older population or more sustainable mobility.
Countries with higher population density had lower R0 estimates. For most
regions, variability in R0 spiked initially before reducing and stabilising as
more estimates became available.
| stat |
Power divergence approach for one-shot device testing under competing
risks | Most work on one-shot devices assume that there is only one possible cause of
device failure. However, in practice, it is often the case that the products
under study can experience any one of various possible causes of failure.
Robust estimators and Wald-type tests are developed here for the case of
one-shot devices under competing risks. An extensive simulation study
illustrates the robustness of these divergence-based estimators and test
procedures based on them. A data-driven procedure is proposed for choosing the
optimal estimator for any given data set which is then applied to an example in
the context of survival analysis.
| stat |
Variational Nonlinear System Identification | This paper considers parameter estimation for nonlinear state-space models,
which is an important but challenging problem. We address this challenge by
employing a variational inference (VI) approach, which is a principled method
that has deep connections to maximum likelihood estimation. This VI approach
ultimately provides estimates of the model as solutions to an optimisation
problem, which is deterministic, tractable and can be solved using standard
optimisation tools. A specialisation of this approach for systems with additive
Gaussian noise is also detailed. The proposed method is examined numerically on
a range of simulation and real examples with a focus on robustness to parameter
initialisations; we additionally perform favourable comparisons against
state-of-the-art alternatives.
| stat |
Functional Lagged Regression with Sparse Noisy Observations | A functional (lagged) time series regression model involves the regression of
scalar response time series on a time series of regressors that consists of a
sequence of random functions. In practice, the underlying regressor curve time
series are not always directly accessible, but are latent processes observed
(sampled) only at discrete measurement locations. In this paper, we consider
the so-called sparse observation scenario where only a relatively small number
of measurement locations have been observed, possibly different for each curve.
The measurements can be further contaminated by additive measurement error. A
spectral approach to the estimation of the model dynamics is considered. The
spectral density of the regressor time series and the cross-spectral density
between the regressors and response time series are estimated by kernel
smoothing methods from the sparse observations. The impulse response regression
coefficients of the lagged regression model are then estimated by means of
ridge regression (Tikhonov regularisation) or PCA regression (spectral
truncation). The latent functional time series are then recovered by means of
prediction, conditioning on all the observed observed data. The performance and
implementation of our methods are illustrated by means of a simulation study
and the analysis of meteorological data.
| stat |
Regulating Greed Over Time in Multi-Armed Bandits | In retail, there are predictable yet dramatic time-dependent patterns in
customer behavior, such as periodic changes in the number of visitors, or
increases in customers just before major holidays. The current paradigm of
multi-armed bandit analysis does not take these known patterns into account.
This means that for applications in retail, where prices are fixed for periods
of time, current bandit algorithms will not suffice. This work provides a
remedy that takes the time-dependent patterns into account, and we show how
this remedy is implemented for the UCB, $\varepsilon$-greedy, and UCB-L
algorithms, and also through a new policy called the variable arm pool
algorithm. In the corrected methods, exploitation (greed) is regulated over
time, so that more exploitation occurs during higher reward periods, and more
exploration occurs in periods of low reward. In order to understand why regret
is reduced with the corrected methods, we present a set of bounds that provide
insight into why we would want to exploit during periods of high reward, and
discuss the impact on regret. Our proposed methods perform well in experiments,
and were inspired by a high-scoring entry in the Exploration and Exploitation 3
contest using data from Yahoo$!$ Front Page. That entry heavily used
time-series methods to regulate greed over time, which was substantially more
effective than other contextual bandit methods.
| stat |
A Test for Differential Ascertainment in Case-Control Studies with
Application to Child Maltreatment | We propose a method to test for the presence of differential ascertainment in
case-control studies, when data are collected by multiple sources. We show
that, when differential ascertainment is present, the use of only the observed
cases leads to severe bias in the computation of the odds ratio. We can
alleviate the effect of such bias using the estimates that our method of
testing for differential ascertainment naturally provides. We apply it to a
dataset obtained from the National Violent Death Reporting System, with the
goal of checking for the presence of differential ascertainment by race in the
count of deaths caused by child maltreatment.
| stat |
Explaining predictive models using Shapley values and non-parametric
vine copulas | The original development of Shapley values for prediction explanation relied
on the assumption that the features being described were independent. If the
features in reality are dependent this may lead to incorrect explanations.
Hence, there have recently been attempts of appropriately modelling/estimating
the dependence between the features. Although the proposed methods clearly
outperform the traditional approach assuming independence, they have their
weaknesses. In this paper we propose two new approaches for modelling the
dependence between the features.
Both approaches are based on vine copulas, which are flexible tools for
modelling multivariate non-Gaussian distributions able to characterise a wide
range of complex dependencies.
The performance of the proposed methods is evaluated on simulated data sets
and a real data set. The experiments demonstrate that the vine copula
approaches give more accurate approximations to the true Shapley values than
its competitors.
| stat |
mecor: An R package for measurement error correction in linear
regression models with a continuous outcome | Measurement error in a covariate or the outcome of regression models is
common, but is often ignored, even though measurement error can lead to
substantial bias in the estimated covariate-outcome association. While several
texts on measurement error correction methods are available, these methods
remain seldomly applied. To improve the use of measurement error correction
methodology, we developed mecor, an R package that implements measurement error
correction methods for regression models with continuous outcomes. Measurement
error correction requires information about the measurement error model and its
parameters. This information can be obtained from four types of studies, used
to estimate the parameters of the measurement error model: an internal
validation study, a replicates study, a calibration study and an external
validation study. In the package mecor, regression calibration methods and a
maximum likelihood method are implemented to correct for measurement error in a
continuous covariate in regression analyses. Additionally, methods of moments
methods are implemented to correct for measurement error in the continuous
outcome in regression analyses. Variance estimation of the corrected estimators
is provided in closed form and using the bootstrap.
| stat |
Deep Ordinal Regression using Optimal Transport Loss and Unimodal Output
Probabilities | We propose a framework for deep ordinal regression, based on unimodal output
distribution and optimal transport loss. Despite being seemingly appropriate,
in many recent works the unimodality requirement is either absent, or
implemented using soft targets, which do not guarantee unimodal outputs at
inference. In addition, we argue that the standard maximum likelihood objective
is not suitable for ordinal regression problems, and that optimal transport is
better suited for this task, as it naturally captures the order of the classes.
Inspired by the well-known Proportional Odds model, we propose to modify its
design by using an architectural mechanism which guarantees that the model
output distribution will be unimodal. We empirically analyze the different
components of our propose approach and demonstrate their contribution to the
performance of the model. Experimental results on three real-world datasets
demonstrate that our proposed approach performs on par with several recently
proposed deep learning approaches for deep ordinal regression with unimodal
output probabilities, while having guarantee on the output unimodality. In
addition, we demonstrate that the level of prediction uncertainty of the model
correlates with its accuracy.
| stat |
Estimating length of hospital stay, with regard to associated factors; A
model to enhance healthcare service efficiency and reduce healthcare resource
consumption | To assess the efficiency and the resource consumption level in healthcare
scope, many economic and social factors have to be considered. An index which
has recently been studied by the researchers, is length of hospital stay (LOS)
defined as how long each patient is hospitalized. Detecting and controlling the
factors affecting this index can help to reduce healthcare costs and also to
improve healthcare service efficiency. The effects of three major factors, say,
the season during which the patient is hospitalized, the patients age and
his/her gender on LOS pertaining to 82718 patients receiving healthcare service
from the hospitals under contract to insurance organization in Tehran, have
been analyzed, using unbalanced factorial design. The test results imply that
the whole model is significant (P-value=0<0.01), the separate effects of all
three factors on LOS are significant (P-value=0<0.01) and the only significant
interaction at alpha=0.01 is the one between gender and age group
(P-value=0<0.01). moreover, no two age groups have the significant equal means
and age groups 1-10 and 26-40 years possess the minimum and maximum means of
LOS, respectively. LOS means in winter and autumn are equal, being the maximum.
Due to the significance of these effects, allocating specified budget and
resources with regard to these factors will help hospitals and healthcare
centers enhance their efficiency and remain within their budget.
| stat |
Fast selection of nonlinear mixed effect models using penalized
likelihood | Nonlinear Mixed effects models are hidden variables models that are widely
used in many field such as pharmacometrics. In such models, the distribution
characteristics of hidden variables can be specified by including several
parameters such as covariates or correlations which must be selected. Recent
development of pharmacogenomics has brought averaged/high dimensional problems
to the field of nonlinear mixed effects modeling for which standard covariates
selection techniques like stepwise methods are not well suited. This work
proposes to select covariates and correlation parameters using a penalized
likelihood approach. The penalized likelihood problem is solved using a
stochastic proximal gradient algorithm to avoid inner-outer iterations. Speed
of convergence of the proximal gradient algorithm is improved by the use of
component-wise adaptive gradient step sizes. The practical implementation and
tuning of the proximal gradient algorithm is explored using simulations.
Calibration of regularization parameters is performed by minimizing the
Bayesian Information Criterion using particle swarm optimization, a zero order
optimization procedure. The use of warm restart and parallelization allows to
reduce significantly computing time. The performance of the proposed method
compared to the traditional grid search strategy is explored using simulated
data. Finally, an application to real data from two pharmacokinetics studies is
provided, one studying an antifibrinolitic and the other studying an
antibiotic.
| stat |
SIRUS: Stable and Interpretable RUle Set for Classification | State-of-the-art learning algorithms, such as random forests or neural
networks, are often qualified as "black-boxes" because of the high number and
complexity of operations involved in their prediction mechanism. This lack of
interpretability is a strong limitation for applications involving critical
decisions, typically the analysis of production processes in the manufacturing
industry. In such critical contexts, models have to be interpretable, i.e.,
simple, stable, and predictive. To address this issue, we design SIRUS (Stable
and Interpretable RUle Set), a new classification algorithm based on random
forests, which takes the form of a short list of rules. While simple models are
usually unstable with respect to data perturbation, SIRUS achieves a remarkable
stability improvement over cutting-edge methods. Furthermore, SIRUS inherits a
predictive accuracy close to random forests, combined with the simplicity of
decision trees. These properties are assessed both from a theoretical and
empirical point of view, through extensive numerical experiments based on our
R/C++ software implementation sirus available from CRAN.
| stat |
A Review of Innovation-Based Methods to Jointly Estimate Model and
Observation Error Covariance Matrices in Ensemble Data Assimilation | Data assimilation combines forecasts from a numerical model with
observations. Most of the current data assimilation algorithms consider the
model and observation error terms as additive Gaussian noise, specified by
their covariance matrices Q and R, respectively. These error covariances, and
specifically their respective amplitudes, determine the weights given to the
background (i.e., the model forecasts) and to the observations in the solution
of data assimilation algorithms (i.e., the analysis). Consequently, Q and R
matrices significantly impact the accuracy of the analysis. This review aims to
present and to discuss, with a unified framework, different methods to jointly
estimate the Q and R matrices using ensemble-based data assimilation
techniques. Most of the methodologies developed to date use the innovations,
defined as differences between the observations and the projection of the
forecasts onto the observation space. These methodologies are based on two main
statistical criteria: (i) the method of moments, in which the theoretical and
empirical moments of the innovations are assumed to be equal, and (ii) methods
that use the likelihood of the observations, themselves contained in the
innovations. The reviewed methods assume that innovations are Gaussian random
variables, although extension to other distributions is possible for
likelihood-based methods. The methods also show some differences in terms of
levels of complexity and applicability to high-dimensional systems. The
conclusion of the review discusses the key challenges to further develop
estimation methods for Q and R. These challenges include taking into account
time-varying error covariances, using limited observational coverage,
estimating additional deterministic error terms, or accounting for correlated
noises.
| stat |
A hierarchical Bayesian model to find brain-behaviour associations in
incomplete data sets | Canonical Correlation Analysis (CCA) and its regularised versions have been
widely used in the neuroimaging community to uncover multivariate associations
between two data modalities (e.g., brain imaging and behaviour). However, these
methods have inherent limitations: (1) statistical inferences about the
associations are often not robust; (2) the associations within each data
modality are not modelled; (3) missing values need to be imputed or removed.
Group Factor Analysis (GFA) is a hierarchical model that addresses the first
two limitations by providing Bayesian inference and modelling modality-specific
associations. Here, we propose an extension of GFA that handles missing data,
and highlight that GFA can be used as a predictive model. We applied GFA to
synthetic and real data consisting of brain connectivity and non-imaging
measures from the Human Connectome Project (HCP). In synthetic data, GFA
uncovered the underlying shared and specific factors and predicted correctly
the non-observed data modalities in complete and incomplete data sets. In the
HCP data, we identified four relevant shared factors, capturing associations
between mood, alcohol and drug use, cognition, demographics and
psychopathological measures and the default mode, frontoparietal control,
dorsal and ventral networks and insula, as well as two factors describing
associations within brain connectivity. In addition, GFA predicted a set of
non-imaging measures from brain connectivity. These findings were consistent in
complete and incomplete data sets, and replicated previous findings in the
literature. GFA is a promising tool that can be used to uncover associations
between and within multiple data modalities in benchmark datasets (such as,
HCP), and easily extended to more complex models to solve more challenging
tasks.
| stat |
Contemporary statistical inference for infectious disease models using
Stan | This paper is concerned with the application of recent statistical advances
to inference of infectious disease dynamics. We describe the fitting of a class
of epidemic models using Hamiltonian Monte Carlo and Variational Inference as
implemented in the freely available Stan software. We apply the two methods to
real data from outbreaks as well as routinely collected observations. Our
results suggest that both inference methods are computationally feasible in
this context, and show a trade-off between statistical efficiency versus
computational speed. The latter appears particularly relevant for real-time
applications.
| stat |
Hybrid safe-strong rules for efficient optimization in lasso-type
problems | The lasso model has been widely used for model selection in data mining,
machine learning, and high-dimensional statistical analysis. However, with the
ultrahigh-dimensional, large-scale data sets now collected in many real-world
applications, it is important to develop algorithms to solve the lasso that
efficiently scale up to problems of this size. Discarding features from certain
steps of the algorithm is a powerful technique for increasing efficiency and
addressing the Big Data challenge. In this paper, we propose a family of hybrid
safe-strong rules (HSSR) which incorporate safe screening rules into the
sequential strong rule (SSR) to remove unnecessary computational burden. In
particular, we present two instances of HSSR, namely SSR-Dome and SSR-BEDPP,
for the standard lasso problem. We further extend SSR-BEDPP to the elastic net
and group lasso problems to demonstrate the generalizability of the hybrid
screening idea. Extensive numerical experiments with synthetic and real data
sets are conducted for both the standard lasso and the group lasso problems.
Results show that our proposed hybrid rules can substantially outperform
existing state-of-the-art rules.
| stat |
A Wavelet-Based Independence Test for Functional Data with an
Application to MEG Functional Connectivity | Measuring and testing the dependency between multiple random functions is
often an important task in functional data analysis. In the literature, a
model-based method relies on a model which is subject to the risk of model
misspecification, while a model-free method only provides a correlation measure
which is inadequate to test independence. In this paper, we adopt the
Hilbert-Schmidt Independence Criterion (HSIC) to measure the dependency between
two random functions. We develop a two-step procedure by first pre-smoothing
each function based on its discrete and noisy measurements and then applying
the HSIC to recovered functions. To ensure the compatibility between the two
steps such that the effect of the pre-smoothing error on the subsequent HSIC is
asymptotically negligible, we propose to use wavelet soft-thresholding for
pre-smoothing and Besov-norm-induced kernels for HSIC. We also provide the
corresponding asymptotic analysis. The superior numerical performance of the
proposed method over existing ones is demonstrated in a simulation study.
Moreover, in an magnetoencephalography (MEG) data application, the functional
connectivity patterns identified by the proposed method are more anatomically
interpretable than those by existing methods.
| stat |
Error control in the numerical posterior distribution in the Bayesian UQ
analysis of a semilinear evolution PDE | We elaborate on results obtained in \cite{christen2018} for controlling the
numerical posterior error for Bayesian UQ problems, now considering forward
maps arising from the solution of a semilinear evolution partial differential
equation. Results in \cite{christen2018} demand an estimate for the absolute
global error (AGE) of the numeric forward map. Our contribution is a numerical
method for computing the AGE for semilinear evolution PDEs and shows the
potential applicability of \cite{christen2018} in this important wide range
family of PDEs. Numerical examples are given to illustrate the efficiency of
the proposed method, obtaining numerical posterior distributions for unknown
parameters that are nearly identical to the corresponding theoretical
posterior, by keeping their Bayes factor close to 1.
| stat |
Reversible Jump PDMP Samplers for Variable Selection | A new class of Markov chain Monte Carlo (MCMC) algorithms, based on
simulating piecewise deterministic Markov processes (PDMPs), have recently
shown great promise: they are non-reversible, can mix better than standard MCMC
algorithms, and can use subsampling ideas to speed up computation in big data
scenarios. However, current PDMP samplers can only sample from posterior
densities that are differentiable almost everywhere, which precludes their use
for model choice. Motivated by variable selection problems, we show how to
develop reversible jump PDMP samplers that can jointly explore the discrete
space of models and the continuous space of parameters. Our framework is
general: it takes any existing PDMP sampler, and adds two types of
trans-dimensional moves that allow for the addition or removal of a variable
from the model. We show how the rates of these trans-dimensional moves can be
calculated so that the sampler has the correct invariant distribution.
Simulations show that the new samplers can mix better than standard MCMC
algorithms. Our empirical results show they are also more efficient than
gradient-based samplers that avoid model choice through use of continuous
spike-and-slab priors which replace a point mass at zero for each parameter
with a density concentrated around zero.
| stat |
ASIED: A Bayesian Adaptive Subgroup-Identification Enrichment Design | Developing targeted therapies based on patients' baseline characteristics and
genomic profiles such as biomarkers has gained growing interests in recent
years. Depending on patients' clinical characteristics, the expression of
specific biomarkers or their combinations, different patient subgroups could
respond differently to the same treatment. An ideal design, especially at the
proof of concept stage, should search for such subgroups and make dynamic
adaptation as the trial goes on. When no prior knowledge is available on
whether the treatment works on the all-comer population or only works on the
subgroup defined by one biomarker or several biomarkers, it is necessary to
incorporate the adaptive estimation of the heterogeneous treatment effect to
the decision-making at interim analyses. To address this problem, we propose an
Adaptive Subgroup-Identification Enrichment Design, ASIED, to simultaneously
search for predictive biomarkers, identify the subgroups with differential
treatment effects, and modify study entry criteria at interim analyses when
justified. More importantly, we construct robust quantitative decision-making
rules for population enrichment when the interim outcomes are heterogeneous in
the context of a multilevel target product profile, which defines the minimal
and targeted levels of treatment effect. Through extensive simulations, the
ASIED is demonstrated to achieve desirable operating characteristics and
compare favorably against alternatives.
| stat |
Generalising rate heterogeneity across sites in statistical
phylogenetics | Phylogenetics uses alignments of molecular sequence data to learn about
evolutionary trees relating species. Along branches, sequence evolution is
modelled using a continuous-time Markov process characterised by an
instantaneous rate matrix. Early models assumed the same rate matrix governed
substitutions at all sites of the alignment, ignoring variation in evolutionary
pressures. Substantial improvements in phylogenetic inference and model fit
were achieved by augmenting these models with multiplicative random effects
that describe the result of variation in selective constraints and allow sites
to evolve at different rates which linearly scale a baseline rate matrix.
Motivated by this pioneering work, we consider an extension using a quadratic,
rather than linear, transformation. The resulting models allow for variation in
the selective coefficients of different types of point mutation at a site in
addition to variation in selective constraints.
We derive properties of the extended models. For certain non-stationary
processes, the extension gives a model that allows variation in sequence
composition both across sites and taxa. We adopt a Bayesian approach, describe
an MCMC algorithm for posterior inference and provide software. Our quadratic
models are applied to alignments spanning the tree of life and compared with
site-homogeneous and linear models.
| stat |
Sampling-Free Variational Inference of Bayesian Neural Networks by
Variance Backpropagation | We propose a new Bayesian Neural Net formulation that affords variational
inference for which the evidence lower bound is analytically tractable subject
to a tight approximation. We achieve this tractability by (i) decomposing ReLU
nonlinearities into the product of an identity and a Heaviside step function,
(ii) introducing a separate path that decomposes the neural net expectation
from its variance. We demonstrate formally that introducing separate latent
binary variables to the activations allows representing the neural network
likelihood as a chain of linear operations. Performing variational inference on
this construction enables a sampling-free computation of the evidence lower
bound which is a more effective approximation than the widely applied Monte
Carlo sampling and CLT related techniques. We evaluate the model on a range of
regression and classification tasks against BNN inference alternatives, showing
competitive or improved performance over the current state-of-the-art.
| stat |
Convergence and finite sample approximations of entropic regularized
Wasserstein distances in Gaussian and RKHS settings | This work studies the convergence and finite sample approximations of
entropic regularized Wasserstein distances in the Hilbert space setting. Our
first main result is that for Gaussian measures on an infinite-dimensional
Hilbert space, convergence in the 2-Sinkhorn divergence is {\it strictly
weaker} than convergence in the exact 2-Wasserstein distance. Specifically, a
sequence of centered Gaussian measures converges in the 2-Sinkhorn divergence
if the corresponding covariance operators converge in the Hilbert-Schmidt norm.
This is in contrast to the previous known result that a sequence of centered
Gaussian measures converges in the exact 2-Wasserstein distance if and only if
the covariance operators converge in the trace class norm. In the reproducing
kernel Hilbert space (RKHS) setting, the {\it kernel Gaussian-Sinkhorn
divergence}, which is the Sinkhorn divergence between Gaussian measures defined
on an RKHS, defines a semi-metric on the set of Borel probability measures on a
Polish space, given a characteristic kernel on that space. With the
Hilbert-Schmidt norm convergence, we obtain {\it dimension-independent}
convergence rates for finite sample approximations of the kernel
Gaussian-Sinkhorn divergence, with the same order as the Maximum Mean
Discrepancy. These convergence rates apply in particular to Sinkhorn divergence
between Gaussian measures on Euclidean and infinite-dimensional Hilbert spaces.
The sample complexity for the 2-Wasserstein distance between Gaussian measures
on Euclidean space, while dimension-dependent and larger than that of the
Sinkhorn divergence, is exponentially faster than the worst case scenario in
the literature.
| stat |
Algorithmic Linearly Constrained Gaussian Processes | We algorithmically construct multi-output Gaussian process priors which
satisfy linear differential equations. Our approach attempts to parametrize all
solutions of the equations using Gr\"obner bases. If successful, a push forward
Gaussian process along the paramerization is the desired prior. We consider
several examples from physics, geomathematics and control, among them the full
inhomogeneous system of Maxwell's equations. By bringing together stochastic
learning and computer algebra in a novel way, we combine noisy observations
with precise algebraic computations.
| stat |
Rejoinder: On nearly assumption-free tests of nominal confidence
interval coverage for causal parameters estimated by machine learning | This is the rejoinder to the discussion by Kennedy, Balakrishnan and
Wasserman on the paper "On nearly assumption-free tests of nominal confidence
interval coverage for causal parameters estimated by machine learning"
published in Statistical Science.
| stat |
How to gamble with non-stationary $\mathcal{X}$-armed bandits and have
no regrets | In $\mathcal{X}$-armed bandit problem an agent sequentially interacts with
environment which yields a reward based on the vector input the agent provides.
The agent's goal is to maximise the sum of these rewards across some number of
time steps. The problem and its variations have been a subject of numerous
studies, suggesting sub-linear and some times optimal strategies. The given
paper introduces a novel variation of the problem. We consider an environment,
which can abruptly change its behaviour an unknown number of times. To that end
we propose a novel strategy and prove it attains sub-linear cumulative regret.
Moreover, in case of highly smooth relation between an action and the
corresponding reward, the method is nearly optimal. The theoretical result are
supported by experimental study.
| stat |
Optimal Design for A/B Testing in the Presence of Covariates and Network
Connection | A/B testing, also known as controlled experiments, refers to the statistical
procedure of conducting an experiment to compare two treatments applied to
different testing subjects. For example, many companies offering online
services frequently to conduct A/B testing on their users who are connected in
social networks. Since two connected users usually share some similar traits,
we assume that their measurements are related to their network adjacency. In
this paper, we assume that the users, or the test subjects of the experiments,
are connected on an undirected network. The subjects' responses are affected by
the treatment assignment, the observed covariate features, as well as the
network connection. We include the variation from these three sources in a
conditional autoregressive model. Based on this model, we propose a design
criterion on treatment allocation that minimizes the variance of the estimated
treatment effect. Since the design criterion depends on an unknown network
correlation parameter, we propose a Bayesian optimal design method and a hybrid
solution approach to obtain the optimal design. Examples via synthetic and real
social networks are shown to demonstrate the performances of the proposed
approach.
| stat |
Intensity Estimation for Poisson Process with Compositional Noise | Intensity estimation for Poisson processes is a classical problem and has
been extensively studied over the past few decades. Practical observations,
however, often contain compositional noise, i.e. a nonlinear shift along the
time axis, which makes standard methods not directly applicable. The key
challenge is that these observations are not "aligned", and registration
procedures are required for successful estimation. In this paper, we propose an
alignment-based framework for positive intensity estimation. We first show that
the intensity function is area-preserved with respect to compositional noise.
Such a property implies that the time warping is only encoded in the normalized
intensity, or density, function. Then, we decompose the estimation of the
intensity by the product of the estimated total intensity and estimated
density. The estimation of the density relies on a metric which measures the
phase difference between two density functions. An asymptotic study shows that
the proposed estimation algorithm provides a consistent estimator for the
normalized intensity. We then extend the framework to estimating non-negative
intensity functions. The success of the proposed estimation algorithms is
illustrated using two simulations. Finally, we apply the new framework in a
real data set of neural spike trains, and find that the newly estimated
intensities provide better classification accuracy than previous methods.
| stat |
Practical Transfer Learning for Bayesian Optimization | Bayesian optimization has become a standard technique for hyperparameter
optimization of machine learning algorithms. We consider the setting where
previous optimization runs are available, and we wish to transfer their
outcomes to a new optimization run and thereby accelerate the search. We
develop a new hyperparameter-free ensemble model for Bayesian optimization,
based on a linear combination of Gaussian Processes and Agnostic Bayesian
Learning of Ensembles. We show that this is a generalization of two existing
transfer learning extensions to Bayesian optimization and establish a
worst-case bound compared to vanilla Bayesian optimization. Using a large
collection of hyperparameter optimization benchmark problems, we demonstrate
that our contributions substantially reduce optimization time compared to
standard Gaussian process-based Bayesian optimization and improve over the
current state-of-the-art for warm-starting Bayesian optimization.
| stat |
Contour Models for Boundaries Enclosing Star-Shaped and Approximately
Star-Shaped Polygons | Boundaries on spatial fields divide regions with particular features from
surrounding background areas. These boundaries are often described with contour
lines. To measure and record these boundaries, contours are often represented
as ordered sequences of spatial points that connect to form a line. Methods to
identify boundary lines from interpolated spatial fields are well-established.
Less attention has been paid to how to model sequences of connected spatial
points. For data of the latter form, we introduce the Gaussian Star-shaped
Contour Model (GSCM). GSMCs generate sequences of spatial points via generating
sets of distances in various directions from a fixed starting point. The GSCM
is designed for modeling contours that enclose regions that are star-shaped
polygons or approximately star-shaped polygons. Metrics are introduced to
assess the extent to which a polygon deviates from star-shaped. Simulation
studies illustrate the performance of the GSCM in various scenarios and an
analysis of Arctic sea ice edge contour data highlights how GSCMs can be
applied to observational data.
| stat |
Sparse Functional Principal Component Analysis in High Dimensions | Functional principal component analysis (FPCA) is a fundamental tool and has
attracted increasing attention in recent decades, while existing methods are
restricted to data with a single or finite number of random functions (much
smaller than the sample size $n$). In this work, we focus on high-dimensional
functional processes where the number of random functions $p$ is comparable to,
or even much larger than $n$. Such data are ubiquitous in various fields such
as neuroimaging analysis, and cannot be properly modeled by existing methods.
We propose a new algorithm, called sparse FPCA, which is able to model
principal eigenfunctions effectively under sensible sparsity regimes. While
sparsity assumptions are standard in multivariate statistics, they have not
been investigated in the complex context where not only is $p$ large, but also
each variable itself is an intrinsically infinite-dimensional process. The
sparsity structure motivates a thresholding rule that is easy to compute
without nonparametric smoothing by exploiting the relationship between
univariate orthonormal basis expansions and multivariate Kahunen-Lo\`eve (K-L)
representations. We investigate the theoretical properties of the resulting
estimators, and illustrate the performance with simulated and real data
examples.
| stat |
Rdimtools: An R package for Dimension Reduction and Intrinsic Dimension
Estimation | Discovering patterns of the complex high-dimensional data is a long-standing
problem. Dimension Reduction (DR) and Intrinsic Dimension Estimation (IDE) are
two fundamental thematic programs that facilitate geometric understanding of
the data. We present Rdimtools - an R package that supports 133 DR and 17 IDE
algorithms whose extent makes multifaceted scrutiny of the data in one place
easier. Rdimtools is distributed under the MIT license and is accessible from
CRAN, GitHub, and its package website, all of which deliver instruction for
installation, self-contained examples, and API documentation.
| stat |
The population-attributable fraction for time-dependent exposures and
competing risks - A discussion on estimands | The population-attributable fraction (PAF) quantifies the public health
impact of a harmful exposure. Despite being a measure of significant importance
an estimand accommodating complicated time-to-event data is not clearly
defined. We discuss current estimands of the PAF used to quantify the public
health impact of an internal time-dependent exposure for data subject to
competing outcomes. To overcome some limitations, we proposed a novel estimand
which is based on dynamic prediction by landmarking. In a profound simulation
study, we discuss interpretation and performance of the various estimands and
their estimators. The methods are applied to a large French database to
estimate the health impact of ventilator-associated pneumonia for patients in
intensive care.
| stat |
Improving the Accuracy of Principal Component Analysis by the Maximum
Entropy Method | Classical Principal Component Analysis (PCA) approximates data in terms of
projections on a small number of orthogonal vectors. There are simple
procedures to efficiently compute various functions of the data from the PCA
approximation. The most important function is arguably the Euclidean distance
between data items, This can be used, for example, to solve the approximate
nearest neighbor problem. We use random variables to model the inherent
uncertainty in such approximations, and apply the Maximum Entropy Method to
infer the underlying probability distribution. We propose using the expected
values of distances between these random variables as improved estimates of the
distance. We show by analysis and experimentally that in most cases results
obtained by our method are more accurate than what is obtained by the classical
approach. This improves the accuracy of a classical technique that have been
used with little change for over 100 years.
| stat |
Generalized inverse-Gaussian frailty models with application to TARGET
neuroblastoma data | A new class of survival frailty models based on the Generalized
Inverse-Gaussian (GIG) distributions is proposed. We show that the GIG frailty
models are flexible and mathematically convenient like the popular gamma
frailty model. Furthermore, our proposed class is robust and does not present
some computational issues experienced by the gamma model. By assuming a
piecewise-exponential baseline hazard function, which gives a semiparametric
flavour for our frailty class, we propose an EM-algorithm for estimating the
model parameters and provide an explicit expression for the information matrix.
Simulated results are addressed to check the finite sample behavior of the
EM-estimators and also to study the performance of the GIG models under
misspecification. We apply our methodology to a TARGET (Therapeutically
Applicable Research to Generate Effective Treatments) data about survival time
of patients with neuroblastoma cancer and show some advantages of the GIG
frailties over existing models in the literature.
| stat |
Competition-based control of the false discovery proportion | Target-decoy competition (TDC) is commonly used in the computational mass
spectrometry community for false discovery rate (FDR) control (6). Recently,
TDC's underlying approach of competition-based FDR control has gained
significant popularity in other fields after Barber and Candes laid its
theoretical foundation in a more general setting (1). However, any approach
that aims to control the FDR, which is defined as the expected value of the
false discovery proportion (FDP), suffers from a problem. Specifically, even
when successfully controlling the FDR al level {\alpha}, the FDP in our list of
discoveries can significantly exceed {\alpha}. We offer two new procedures to
address this problem. TDC-SD rigorously controls the FDP in the competition
setup by guaranteeing that the FDP is bounded by {\alpha} at any desired
confidence level. The complementary TDC-UB is designed to bound the FDP for any
list of top scoring target discoveries. For example, TDC-UB can give a
95%-confidence upper-bound on the FDP when using the canonical TDC procedure
for controlling the FDR. We demonstrate the utility of these procedures using
synthetic and real data in multiple domains.
| stat |
MRT-SS Calculator: An R Shiny Application for Sample Size Calculation in
Micro-Randomized Trials | The micro-randomized trial (MRT) is a new experimental design which allows
for the investigation of the proximal effects of a "just-in-time" treatment,
often provided via a mobile device as part of a mobile health intervention. As
with a traditional randomized controlled trial, computing the minimum required
sample size to achieve a desired power is a crucial step in designing an MRT.
We present MRT-SS Calculator, an online sample-size calculator for
micro-randomized trials, built with R Shiny. MRT-SS Calculator requires
specification of time-varying patterns for the proximal treatment effect and
expected treatment availability. We illustrate the implementation of MRT-SS
Calculator using a mobile health trial, HeartSteps. The application can be
accessed from https://pengliao.shinyapps.io/mrt-calculator.
| stat |
A Statistical Perspective on the Challenges in Molecular Microbial
Biology | High throughput sequencing (HTS)-based technology enables identifying and
quantifying non-culturable microbial organisms in all environments. Microbial
sequences have enhanced our understanding of the human microbiome, the soil and
plant environment, and the marine environment. All molecular microbial data
pose statistical challenges due to contamination sequences from reagents, batch
effects, unequal sampling, and undetected taxa. Technical biases and
heteroscedasticity have the strongest effects, but different strains across
subjects and environments also make direct differential abundance testing
unwieldy. We provide an introduction to a few statistical tools that can
overcome some of these difficulties and demonstrate those tools on an example.
We show how standard statistical methods, such as simple hierarchical mixture
and topic models, can facilitate inferences on latent microbial communities. We
also review some nonparametric Bayesian approaches that combine visualization
and uncertainty quantification. The intersection of molecular microbial biology
and statistics is an exciting new venue. Finally, we list some of the important
open problems that would benefit from more careful statistical method
development.
| stat |
Stochastic Bandits with Delay-Dependent Payoffs | Motivated by recommendation problems in music streaming platforms, we propose
a nonstationary stochastic bandit model in which the expected reward of an arm
depends on the number of rounds that have passed since the arm was last pulled.
After proving that finding an optimal policy is NP-hard even when all model
parameters are known, we introduce a class of ranking policies provably
approximating, to within a constant factor, the expected reward of the optimal
policy. We show an algorithm whose regret with respect to the best ranking
policy is bounded by $\widetilde{\mathcal{O}}\big(\!\sqrt{kT}\big)$, where $k$
is the number of arms and $T$ is time. Our algorithm uses only
$\mathcal{O}\big(k\ln\ln T\big)$ switches, which helps when switching between
policies is costly. As constructing the class of learning policies requires
ordering the arms according to their expectations, we also bound the number of
pulls required to do so. Finally, we run experiments to compare our algorithm
against UCB on different problem instances.
| stat |
Inverse Ising inference from high-temperature re-weighting of
observations | Maximum Likelihood Estimation (MLE) is the bread and butter of system
inference for stochastic systems. In some generality, MLE will converge to the
correct model in the infinite data limit. In the context of physical approaches
to system inference, such as Boltzmann machines, MLE requires the arduous
computation of partition functions summing over all configurations, both
observed and unobserved. We present here a conceptually and computationally
transparent data-driven approach to system inference that is based on the
simple question: How should the Boltzmann weights of observed configurations be
modified to make the probability distribution of observed configurations close
to a flat distribution? This algorithm gives accurate inference by using only
observed configurations for systems with a large number of degrees of freedom
where other approaches are intractable.
| stat |
Multivariate mixed membership modeling: Inferring domain-specific risk
profiles | Characterizing the shared memberships of individuals in a classification
scheme poses severe interpretability issues, even when using a moderate number
of classes (say 4). Mixed membership models quantify this phenomenon, but they
typically focus on goodness-of-fit more than on interpretable inference. To
achieve a good numerical fit, these models may in fact require many extreme
profiles, making the results difficult to interpret. We introduce a new class
of multivariate mixed membership models that, when variables can be partitioned
into subject-matter based domains, can provide a good fit to the data using
fewer profiles than standard formulations. The proposed model explicitly
accounts for the blocks of variables corresponding to the distinct domains
along with a cross-domain correlation structure, which provides new information
about shared membership of individuals in a complex classification scheme. We
specify a multivariate logistic normal distribution for the membership vectors,
which allows easy introduction of auxiliary information leveraging a latent
multivariate logistic regression. A Bayesian approach to inference, relying on
P\'olya gamma data augmentation, facilitates efficient posterior computation
via Markov Chain Monte Carlo. We apply this methodology to a spatially explicit
study of malaria risk over time on the Brazilian Amazon frontier.
| stat |
Learning Stochastic Closures Using Ensemble Kalman Inversion | Although the governing equations of many systems, when derived from first
principles, may be viewed as known, it is often too expensive to numerically
simulate all the interactions they describe. Therefore researchers often seek
simpler descriptions that describe complex phenomena without numerically
resolving all the interacting components. Stochastic differential equations
(SDEs) arise naturally as models in this context. The growth in data
acquisition, both through experiment and through simulations, provides an
opportunity for the systematic derivation of SDE models in many disciplines.
However, inconsistencies between SDEs and real data at short time scales often
cause problems, when standard statistical methodology is applied to parameter
estimation. The incompatibility between SDEs and real data can be addressed by
deriving sufficient statistics from the time-series data and learning
parameters of SDEs based on these. Following this approach, we formulate the
fitting of SDEs to sufficient statistics from real data as an inverse problem
and demonstrate that this inverse problem can be solved by using ensemble
Kalman inversion (EKI). Furthermore, we create a framework for non-parametric
learning of drift and diffusion terms by introducing hierarchical, refinable
parameterizations of unknown functions, using Gaussian process regression. We
demonstrate the proposed methodology for the fitting of SDE models, first in a
simulation study with a noisy Lorenz '63 model, and then in other applications,
including dimension reduction in deterministic chaotic systems arising in the
atmospheric sciences, large-scale pattern modeling in climate dynamics, and
simplified models for key observables arising in molecular dynamics. The
results confirm that the proposed methodology provides a robust and systematic
approach to fitting SDE models to real data.
| stat |
Constructing Likelihood Functions for Interval-valued Random Variables | There is a growing need for the ability to analyse interval-valued data.
However, existing descriptive frameworks to achieve this ignore the process by
which interval-valued data are typically constructed; namely by the aggregation
of real-valued data generated from some underlying process. In this article we
develop the foundations of likelihood based statistical inference for random
intervals that directly incorporates the underlying generative procedure into
the analysis. That is, it permits the direct fitting of models for the
underlying real-valued data given only the random interval-valued summaries.
This generative approach overcomes several problems associated with existing
methods, including the rarely satisfied assumption of within-interval
uniformity. The new methods are illustrated by simulated and real data
analyses.
| stat |
Identifying and Predicting Parkinson's Disease Subtypes through
Trajectory Clustering via Bipartite Networks | Parkinson's disease (PD) is a common neurodegenerative disease with a high
degree of heterogeneity in its clinical features, rate of progression, and
change of variables over time. In this work, we present a novel data-driven,
network-based Trajectory Profile Clustering (TPC) algorithm for 1)
identification of PD subtypes and 2) early prediction of disease progression in
individual patients. Our subtype identification is based not only on PD
variables, but also on their complex patterns of progression, providing a
useful tool for the analysis of large heterogenous, longitudinal data.
Specifically, we cluster patients based on the similarity of their trajectories
through a time series of bipartite networks connecting patients to demographic,
clinical, and genetic variables. We apply this approach to demographic and
clinical data from the Parkinson's Progression Markers Initiative (PPMI)
dataset and identify 3 patient clusters, consistent with 3 distinct PD
subtypes, each with a characteristic variable progression profile.
Additionally, TPC predicts an individual patient's subtype and future disease
trajectory, based on baseline assessments. Application of our approach resulted
in 74% accurate subtype prediction in year 5 in a test/validation cohort.
Furthermore, we show that genetic variability can be integrated seamlessly in
our TPC approach. In summary, using PD as a model for chronic progressive
diseases, we show that TPC leverages high-dimensional longitudinal datasets for
subtype identification and early prediction of individual disease subtype. We
anticipate this approach will be broadly applicable to multidimensional
longitudinal datasets in diverse chronic diseases.
| stat |
Policy Analysis using Synthetic Controls in Continuous-Time | Counterfactual estimation using synthetic controls is one of the most
successful recent methodological developments in causal inference. Despite its
popularity, the current description only considers time series aligned across
units and synthetic controls expressed as linear combinations of observed
control units. We propose a continuous-time alternative that models the latent
counterfactual path explicitly using the formalism of controlled differential
equations. This model is directly applicable to the general setting of
irregularly-aligned multivariate time series and may be optimized in rich
function spaces -- thereby improving on some limitations of existing
approaches.
| stat |
Directional Assessment of Traffic Flow Extremes | We analyze extremes of traffic flow profiles composed of traffic counts over
a day. The data is essentially curves and determining which trajectory should
be classified as extreme is not straight forward. To assess the extremes of the
traffic flow curves in a coherent way, we use a directional definition of
extremeness and apply the dimension reduction technique called principal
component analysis (PCA) in an asymmetric norm. In the classical PCA one
reduces the dimensions of the data by projecting it in the direction of the
largest variation of the projection around its mean. In the PCA in an
asymmetric norm one chooses the projection directions, such that the
asymmetrically weighted variation around a tail index -- an expectile -- of the
data is the largest possible. Expectiles are tail measures that generalize the
mean in a similar manner as quantiles generalize the median. Focusing on the
asymmetrically weighted variation around an expectile of the data, we find the
appropriate projection directions and the low dimensional representation of the
traffic flow profiles that uncover different patterns in their extremes. Using
the traffic flow data from the roundabout on Ernst-Reuter-Platz in the city
center of Berlin, Germany, we estimate, visualize and interpret the resulting
principal expectile components. The corresponding directional extremes of the
traffic flow profiles are simple to identify and to connect to their location-
and time-related specifics. Their shapes are driven by their scores on each
principal expectile component which is useful for extracting and analyzing
traffic patterns. Our approach to dimensionality reduction towards the
directional extremes of traffic flow extends the related methodological basis
and gives promising results for subsequent analysis, prediction and control of
traffic flow patterns.
| stat |
Domain Generalization via Multidomain Discriminant Analysis | Domain generalization (DG) aims to incorporate knowledge from multiple source
domains into a single model that could generalize well on unseen target
domains. This problem is ubiquitous in practice since the distributions of the
target data may rarely be identical to those of the source data. In this paper,
we propose Multidomain Discriminant Analysis (MDA) to address DG of
classification tasks in general situations. MDA learns a domain-invariant
feature transformation that aims to achieve appealing properties, including a
minimal divergence among domains within each class, a maximal separability
among classes, and overall maximal compactness of all classes. Furthermore, we
provide the bounds on excess risk and generalization error by learning theory
analysis. Comprehensive experiments on synthetic and real benchmark datasets
demonstrate the effectiveness of MDA.
| stat |
A Practical Two-Sample Test for Weighted Random Graphs | Network (graph) data analysis is a popular research topic in statistics and
machine learning. In application, one is frequently confronted with graph
two-sample hypothesis testing where the goal is to test the difference between
two graph populations. Several statistical tests have been devised for this
purpose in the context of binary graphs. However, many of the practical
networks are weighted and existing procedures can't be directly applied to
weighted graphs. In this paper, we study the weighted graph two-sample
hypothesis testing problem and propose a practical test statistic. We prove
that the proposed test statistic converges in distribution to the standard
normal distribution under the null hypothesis and analyze its power
theoretically. The simulation study shows that the proposed test has
satisfactory performance and it substantially outperforms the existing
counterpart in the binary graph case. A real data application is provided to
illustrate the method.
| stat |
Stein Point Markov Chain Monte Carlo | An important task in machine learning and statistics is the approximation of
a probability measure by an empirical measure supported on a discrete point
set. Stein Points are a class of algorithms for this task, which proceed by
sequentially minimising a Stein discrepancy between the empirical measure and
the target and, hence, require the solution of a non-convex optimisation
problem to obtain each new point. This paper removes the need to solve this
optimisation problem by, instead, selecting each new point based on a Markov
chain sample path. This significantly reduces the computational cost of Stein
Points and leads to a suite of algorithms that are straightforward to
implement. The new algorithms are illustrated on a set of challenging Bayesian
inference problems, and rigorous theoretical guarantees of consistency are
established.
| stat |
Convergence of Gradient Descent on Separable Data | We provide a detailed study on the implicit bias of gradient descent when
optimizing loss functions with strictly monotone tails, such as the logistic
loss, over separable datasets. We look at two basic questions: (a) what are the
conditions on the tail of the loss function under which gradient descent
converges in the direction of the $L_2$ maximum-margin separator? (b) how does
the rate of margin convergence depend on the tail of the loss function and the
choice of the step size? We show that for a large family of super-polynomial
tailed losses, gradient descent iterates on linear networks of any depth
converge in the direction of $L_2$ maximum-margin solution, while this does not
hold for losses with heavier tails. Within this family, for simple linear
models we show that the optimal rates with fixed step size is indeed obtained
for the commonly used exponentially tailed losses such as logistic loss.
However, with a fixed step size the optimal convergence rate is extremely slow
as $1/\log(t)$, as also proved in Soudry et al. (2018). For linear models with
exponential loss, we further prove that the convergence rate could be improved
to $\log (t) /\sqrt{t}$ by using aggressive step sizes that compensates for the
rapidly vanishing gradients. Numerical results suggest this method might be
useful for deep networks.
| stat |
A Condition Number for Hamiltonian Monte Carlo | Hamiltonian Monte Carlo is a popular sampling technique for smooth target
densities. The scale lengths of the target have long been known to influence
integration error and sampling efficiency. However, quantitative measures
intrinsic to the target have been lacking. In this paper, we restrict attention
to the multivariate Gaussian and the leapfrog integrator, and obtain a
condition number corresponding to sampling efficiency. This number, based on
the spectral and Schatten norms, quantifies the number of leapfrog steps needed
to efficiently sample. We demonstrate its utility by using this condition
number to analyze HMC preconditioning techniques. We also find the condition
number of large inverse Wishart matrices, from which we derive burn-in
heuristics.
| stat |
Anomaly and Novelty detection for robust semi-supervised learning | Three important issues are often encountered in Supervised and
Semi-Supervised Classification: class-memberships are unreliable for some
training units (label noise), a proportion of observations might depart from
the main structure of the data (outliers) and new groups in the test set may
have not been encountered earlier in the learning phase (unobserved classes).
The present work introduces a robust and adaptive Discriminant Analysis rule,
capable of handling situations in which one or more of the afore-mentioned
problems occur. Two EM-based classifiers are proposed: the first one that
jointly exploits the training and test sets (transductive approach), and the
second one that expands the parameter estimate using the test set, to complete
the group structure learned from the training set (inductive approach).
Experiments on synthetic and real data, artificially adulterated, are provided
to underline the benefits of the proposed method.
| stat |
Bayesian model selection in the $\mathcal{M}$-open setting --
Approximate posterior inference and probability-proportional-to-size
subsampling for efficient large-scale leave-one-out cross-validation | Comparison of competing statistical models is an essential part of
psychological research. From a Bayesian perspective, various approaches to
model comparison and selection have been proposed in the literature. However,
the applicability of these approaches strongly depends on the assumptions about
the model space $\mathcal{M}$, the so-called model view. Furthermore,
traditional methods like leave-one-out cross-validation (LOO-CV) estimate the
expected log predictive density (ELPD) of a model to investigate how the model
generalises out-of-sample, which quickly becomes computationally inefficient
when sample size becomes large. Here, we provide a tutorial on approximate
Pareto-smoothed importance sampling leave-one-out cross-validation (PSIS-LOO),
a computationally efficient method for Bayesian model comparison. First, we
discuss several model views and the available Bayesian model comparison methods
in each. We then use Bayesian logistic regression as a running example how to
apply the method in practice, and show that it outperforms other methods like
LOO-CV or information criteria in terms of computational effort while providing
similarly accurate ELPD estimates. In a second step, we show how even
large-scale models can be compared efficiently by using posterior
approximations in combination with probability-proportional-to-size
subsampling. We show how to compare competing models based on the ELPD
estimates provided, and how to conduct posterior predictive checks to safeguard
against overconfidence in one of the models under consideration. We conclude
that the method is attractive for mathematical psychologists who aim at
comparing several competing statistical models, which are possibly
high-dimensional and in the big-data regime.
| stat |
Orthogonal Estimation of Wasserstein Distances | Wasserstein distances are increasingly used in a wide variety of applications
in machine learning. Sliced Wasserstein distances form an important subclass
which may be estimated efficiently through one-dimensional sorting operations.
In this paper, we propose a new variant of sliced Wasserstein distance, study
the use of orthogonal coupling in Monte Carlo estimation of Wasserstein
distances and draw connections with stratified sampling, and evaluate our
approaches experimentally in a range of large-scale experiments in generative
modelling and reinforcement learning.
| stat |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.