title
stringlengths 6
244
| abstract
stringlengths 19
5.09k
| label
stringclasses 10
values |
|---|---|---|
A simulation study of semiparametric estimation in copula models based
on minimum Alpha-Divergence | The purpose of this paper is to introduce two semiparametric methods for the
estimation of copula parameter. These methods are based on minimum
Alpha-Divergence between a non-parametric estimation of copula density using
local likelihood probit transformation method and a true copula density
function. A Monte Carlo study is performed to measure the performance of these
methods based on Hellinger distance and Neyman divergence as special cases of
Alpha-Divergence. Simulation results are compared to the Maximum
Pseudo-Likelihood (MPL) estimation as a conventional estimation method in
well-known bivariate copula models. These results show that the proposed method
based on Minimum Pseudo Hellinger Distance estimation has a good performance in
small sample size and weak dependency situations. The parameter estimation
methods are applied to a real data set in Hydrology.
| stat |
Optimally adaptive Bayesian spectral density estimation for stationary
and nonstationary processes | This article improves on existing methods to estimate the spectral density of
stationary and nonstationary time series assuming a Gaussian process prior. By
optimising an appropriate eigendecomposition using a smoothing spline
covariance structure, our method more appropriately models both smooth and
rough data. We further justify the utility of this optimal eigendecomposition
by investigating the performance of alternative covariance functions other than
smoothing splines. We show that the optimal eigendecomposition provides a
material improvement, while the other covariance functions under examination do
not, all performing comparatively well as the smoothing spline. During our
computational investigation, we introduce new validation metrics for the
spectral density estimate, inspired from the physical sciences. We validate our
models in an extensive simulation study and demonstrate superior performance
with real data.
| stat |
Trajectory Optimization for Unknown Constrained Systems using
Reinforcement Learning | In this paper, we propose a reinforcement learning-based algorithm for
trajectory optimization for constrained dynamical systems. This problem is
motivated by the fact that for most robotic systems, the dynamics may not
always be known. Generating smooth, dynamically feasible trajectories could be
difficult for such systems. Using sampling-based algorithms for motion planning
may result in trajectories that are prone to undesirable control jumps.
However, they can usually provide a good reference trajectory which a
model-free reinforcement learning algorithm can then exploit by limiting the
search domain and quickly finding a dynamically smooth trajectory. We use this
idea to train a reinforcement learning agent to learn a dynamically smooth
trajectory in a curriculum learning setting. Furthermore, for generalization,
we parameterize the policies with goal locations, so that the agent can be
trained for multiple goals simultaneously. We show result in both simulated
environments as well as real experiments, for a $6$-DoF manipulator arm
operated in position-controlled mode to validate the proposed idea. We compare
the proposed ideas against a PID controller which is used to track a designed
trajectory in configuration space. Our experiments show that our RL agent
trained with a reference path outperformed a model-free PID controller of the
type commonly used on many robotic platforms for trajectory tracking.
| stat |
Inverse Probability Weighted Estimators of Vaccine Effects Accommodating
Partial Interference and Censoring | Estimating population-level effects of a vaccine is challenging because there
may be interference, i.e., the outcome of one individual may depend on the
vaccination status of another individual. Partial interference occurs when
individuals can be partitioned into groups such that interference occurs only
within groups. In the absence of interference, inverse probability weighted
(IPW) estimators are commonly used to draw inference about causal effects of an
exposure or treatment. Tchetgen Tchetgen and VanderWeele (2012) proposed a
modified IPW estimator for causal effects in the presence of partial
interference. Motivated by a cholera vaccine study in Bangladesh, this paper
considers an extension of the Tchetgen Tchetgen and VanderWeele IPW estimator
to the setting where the outcome is subject to right censoring using inverse
probability of censoring weights (IPCW). Censoring weights are estimated using
proportional hazards frailty models. The large sample properties of the IPCW
estimators are derived, and simulation studies are presented demonstrating the
estimators' performance in finite samples. The methods are then used to analyze
data from the cholera vaccine study.
| stat |
Practical Bayesian Learning of Neural Networks via Adaptive Optimisation
Methods | We introduce a novel framework for the estimation of the posterior
distribution over the weights of a neural network, based on a new probabilistic
interpretation of adaptive optimisation algorithms such as AdaGrad and Adam. We
demonstrate the effectiveness of our Bayesian Adam method, Badam, by
experimentally showing that the learnt uncertainties correctly relate to the
weights' predictive capabilities by weight pruning. We also demonstrate the
quality of the derived uncertainty measures by comparing the performance of
Badam to standard methods in a Thompson sampling setting for multi-armed
bandits, where good uncertainty measures are required for an agent to balance
exploration and exploitation.
| stat |
Exact Dimensionality Selection for Bayesian PCA | We present a Bayesian model selection approach to estimate the intrinsic
dimensionality of a high-dimensional dataset. To this end, we introduce a novel
formulation of the probabilisitic principal component analysis model based on a
normal-gamma prior distribution. In this context, we exhibit a closed-form
expression of the marginal likelihood which allows to infer an optimal number
of components. We also propose a heuristic based on the expected shape of the
marginal likelihood curve in order to choose the hyperparameters. In
non-asymptotic frameworks, we show on simulated data that this exact
dimensionality selection approach is competitive with both Bayesian and
frequentist state-of-the-art methods.
| stat |
Variable Selection with Second-Generation P-Values | Many statistical methods have been proposed for variable selection in the
past century, but few perform this task well. The current standard bearers for
variable selection include smoothly clipped absolute deviation (SCAD), adaptive
lasso (AL), and minimax concave penalty with penalized linear unbiased
selection (MC+). In practice, however, these algorithms often struggle to
balance support recovery and parameter estimation, despite well-established
oracle behavior for variable selection in certain settings. Here we report on a
novel application of second-generation p-values (SGPVs) for variable selection,
which we call Penalized regression with SGPVs (ProSGPV). This approach has
tangible advantages in balancing support recovery and parameter estimation. The
ProSGPV approach captures the true model at the best rate achieved by the
current standards, is easier to implement in practice, and yields parameter
estimates with the smallest mean absolute error. Even with strong collinearity
in the feature space, the ProSGPV approach can maintain its good performance by
using a simple pre-screening step. Here we report on extensive simulations and
two real-world applications comparing these approaches. Our ProSGPV algorithm
is a fast and intuitive approach for variable selection that leverages the
advantages of second-generation p-values.
| stat |
Deep Active Learning with Adaptive Acquisition | Model selection is treated as a standard performance boosting step in many
machine learning applications. Once all other properties of a learning problem
are fixed, the model is selected by grid search on a held-out validation set.
This is strictly inapplicable to active learning. Within the standardized
workflow, the acquisition function is chosen among available heuristics a
priori, and its success is observed only after the labeling budget is already
exhausted. More importantly, none of the earlier studies report a unique
consistently successful acquisition heuristic to the extent to stand out as the
unique best choice. We present a method to break this vicious circle by
defining the acquisition function as a learning predictor and training it by
reinforcement feedback collected from each labeling round. As active learning
is a scarce data regime, we bootstrap from a well-known heuristic that filters
the bulk of data points on which all heuristics would agree, and learn a policy
to warp the top portion of this ranking in the most beneficial way for the
character of a specific data distribution. Our system consists of a Bayesian
neural net, the predictor, a bootstrap acquisition function, a probabilistic
state definition, and another Bayesian policy network that can effectively
incorporate this input distribution. We observe on three benchmark data sets
that our method always manages to either invent a new superior acquisition
function or to adapt itself to the a priori unknown best performing heuristic
for each specific data set.
| stat |
Sparse online variational Bayesian regression | This work considers variational Bayesian inference as an inexpensive and
scalable alternative to a fully Bayesian approach in the context of
sparsity-promoting priors. In particular, the priors considered arise from
scale mixtures of Normal distributions with a generalized inverse Gaussian
mixing distribution. This includes the variational Bayesian LASSO as an
inexpensive and scalable alternative to the Bayesian LASSO introduced in [56].
It also includes priors which more strongly promote sparsity. For linear models
the method requires only the iterative solution of deterministic least squares
problems. Furthermore, for $n\rightarrow \infty$ data points and p unknown
covariates the method can be implemented exactly online with a cost of O(p$^3$)
in computation and O(p$^2$) in memory. For large p an approximation is able to
achieve promising results for a cost of O(p) in both computation and memory.
Strategies for hyper-parameter tuning are also considered. The method is
implemented for real and simulated data. It is shown that the performance in
terms of variable selection and uncertainty quantification of the variational
Bayesian LASSO can be comparable to the Bayesian LASSO for problems which are
tractable with that method, and for a fraction of the cost. The present method
comfortably handles n = p = 131,073 on a laptop in minutes, and n = 10$^5$, p =
10$^6$ overnight.
| stat |
Estimation and Inference for Moments of Ratios with Robustness against
Large Trimming Bias | Empirical researchers often trim observations with small denominator A when
they estimate moments of the form E[B/A]. Large trimming is a common practice
to mitigate variance, but it incurs large trimming bias. This paper provides a
novel method of correcting large trimming bias. If a researcher is willing to
assume that the joint distribution between A and B is smooth, then a large
trimming bias may be estimated well. With the bias correction, we also develop
a valid and robust inference result for E[B/A].
| stat |
Bayesreef: A Bayesian inference framework for modelling reef growth in
response to environmental change and biological dynamics | Estimating the impact of environmental processes on vertical reef development
in geological time is a very challenging task. pyReef-Core is a deterministic
carbonate stratigraphic forward model designed to simulate the key biological
and environmental processes that determine vertical reef accretion and
assemblage changes in fossil reef drill cores. We present a Bayesian framework
called Bayesreef for the estimation and uncertainty quantification of
parameters in pyReef-Core that represent environmental conditions affecting the
growth of coral assemblages on geological timescales. We demonstrate the
existence of multimodal posterior distributions and investigate the challenges
of sampling using Markov chain Monte-Carlo (MCMC) methods, which includes
parallel tempering MCMC. We use synthetic reef-core to investigate fundamental
issues and then apply the methodology to a selected reef-core from the Great
Barrier Reef in Australia. The results show that Bayesreef accurately estimates
and provides uncertainty quantification of the selected parameters that
represent the environment and ecological conditions in pyReef-Core. Bayesreef
provides insights into the complex posterior distributions of parameters in
pyReef-Core, which provides the groundwork for future research in this area.
| stat |
Concordance Rate of a Four-Quadrant Plot for Repeated Measurements | Before new clinical measurement methods are implemented in clinical practice,
it must be confirmed whether their results are equivalent to those of existing
methods. The agreement of the trend between these methods is evaluated using
the four-quadrant plot, which describes the trend of change in each difference
of the two measurement methods' values in sequential time points, and the
plot's concordance rate, which is calculated using the sum of data points in
the four-quadrant plot that agree with this trend divided by the number of all
accepted data points. However, the conventional concordance rate does not
consider the covariance between the data on individual subjects, which may
affect its proper evaluation. Therefore, we proposed a new concordance rate
calculated by each individual according to the number of agreement. Moreover,
this proposed method can set a parameter that the minimum concordant number
between two measurement techniques. The parameter can provide a more detailed
interpretation of the degree of agreement. A numerical simulation conducted
with several factors indicated that the proposed method resulted in a more
accurate evaluation. We also showed a real data and compared the proposed
method with the conventional approach. Then, we concluded the discussion with
the implementation in clinical studies.
| stat |
The DURATIONS randomised trial design: estimation targets, analysis
methods and operating characteristics | Background. Designing trials to reduce treatment duration is important in
several therapeutic areas, including TB and antibiotics. We recently proposed a
new randomised trial design to overcome some of the limitations of standard
two-arm non-inferiority trials. This DURATIONS design involves randomising
patients to a number of duration arms, and modelling the so-called
duration-response curve. This article investigates the operating
characteristics (type-1 and type-2 errors) of different statistical methods of
drawing inference from the estimated curve. Methods. Our first estimation
target is the shortest duration non-inferior to the control (maximum) duration
within a specific risk difference margin. We compare different methods of
estimating this quantity, including using model confidence bands, the delta
method and bootstrap. We then explore the generalisability of results to
estimation targets which focus on absolute event rates, risk ratio and gradient
of the curve. Results. We show through simulations that, in most scenarios and
for most of the estimation targets, using the bootstrap to estimate variability
around the target duration leads to good results for DURATIONS
design-appropriate quantities analogous to power and type-1 error. Using model
confidence bands is not recommended, while the delta method leads to inflated
type-1 error in some scenarios, particularly when the optimal duration is very
close to one of the randomised durations. Conclusions. Using the bootstrap to
estimate the optimal duration in a DURATIONS design has good operating
characteristics in a wide range of scenarios, and can be used with confidence
by researchers wishing to design a DURATIONS trial to reduce treatment
duration. Uncertainty around several different targets can be estimated with
this bootstrap approach.
| stat |
Protein Structure Parameterization via Mobius Distributions on the Torus | Proteins constitute a large group of macromolecules with a multitude of
functions for all living organisms. Proteins achieve this by adopting distinct
three-dimensional structures encoded by the sequence of their constituent amino
acids in one or more polypeptides. In this paper, the statistical modelling of
the protein backbone torsion angles is considered. Two new distributions are
proposed for toroidal data by applying the M\"obius transformation to the
bivariate von Mises distribution. Marginal and conditional distributions in
addition to sine-skewed versions of the proposed models are also developed.
Three big data sets consisting of bivariate information about protein domains
are analysed to illustrate the strength of the flexible proposed models.
Finally, a simulation study is done to evaluate the obtained maximum likelihood
estimates and also to find the best method of generating samples from the
proposed models to use as the proposal distributions in the Markov Chain Monte
Carlo sampling method for predicting the 3D structure of proteins.
| stat |
Learning a high-dimensional classification rule using auxiliary outcomes | Correlated outcomes are common in many practical problems. Based on a
decomposition of estimation bias into two types, within-subspace and
against-subspace, we develop a robust approach to estimating the classification
rule for the outcome of interest with the presence of auxiliary outcomes in
high-dimensional settings. The proposed method includes a pooled estimation
step using all outcomes to gain efficiency, and a subsequent calibration step
using only the outcome of interest to correct both types of biases. We show
that when the pooled estimator has a low estimation error and a sparse
against-subspace bias, the calibrated estimator can achieve a lower estimation
error than that when using only the single outcome of interest. An inference
procedure for the calibrated estimator is also provided. Simulations and a real
data analysis are conducted to justify the superiority of the proposed method.
| stat |
Learning formation energy of inorganic compounds using matrix variate
deep Gaussian process | Future advancement of engineering applications is dependent on design of
novel materials with desired properties. Enormous size of known chemical space
necessitates use of automated high throughput screening to search the desired
material. The high throughput screening uses quantum chemistry calculations to
predict material properties, however, computational complexity of these
calculations often imposes prohibitively high cost on the search for desired
material. This critical bottleneck is resolved by using deep machine learning
to emulate the quantum computations. However, the deep learning algorithms
require a large training dataset to ensure an acceptable generalization, which
is often unavailable a-priory. In this paper, we propose a deep Gaussian
process based approach to develop an emulator for quantum calculations. We
further propose a novel molecular descriptor that enables implementation of the
proposed approach. As demonstrated in this paper, the proposed approach can be
implemented using a small dataset. We demonstrate efficacy of our approach for
prediction of formation energy of inorganic molecules.
| stat |
Learning Signal Subgraphs from Longitudinal Brain Networks with
Symmetric Bilinear Logistic Regression | Modern neuroimaging technologies, combined with state-of-the-art data
processing pipelines, have made it possible to collect longitudinal
observations of an individual's brain connectome at different ages. It is of
substantial scientific interest to study how brain connectivity varies over
time in relation to human cognitive traits. In brain connectomics, the
structural brain network for an individual corresponds to a set of
interconnections among brain regions. We propose a symmetric bilinear logistic
regression to learn a set of small subgraphs relevant to a binary outcome from
longitudinal brain networks as well as estimating the time effects of the
subgraphs. We enforce the extracted signal subgraphs to have clique structure
which has appealing interpretations as they can be related to neurological
circuits. The time effect of each signal subgraph reflects how its predictive
effect on the outcome varies over time, which may improve our understanding of
interactions between the aging of brain structure and neurological disorders.
Application of this method on longitudinal brain connectomics and cognitive
capacity data shows interesting discovery of relevant interconnections among a
small set of brain regions in frontal and temporal lobes with better predictive
performance than competitors.
| stat |
Bayesian Modeling of the Structural Connectome for Studying Alzheimer
Disease | We study possible relations between the structure of the connectome, white
matter connecting different regions of brain, and Alzheimer disease. Regression
models in covariates including age, gender and disease status for the extent of
white matter connecting each pair of regions of brain are proposed. Subject We
study possible relations between the Alzheimer's disease progression and the
structure of the connectome, white matter connecting different regions of
brain. Regression models in covariates including age, gender and disease status
for the extent of white matter connecting each pair of regions of brain are
proposed. Subject inhomogeneity is also incorporated in the model through
random effects with an unknown distribution. As there are large number of pairs
of regions, we also adopt a dimension reduction technique through graphon
(Lovasz and Szegedy (2006)) functions, which reduces functions of pairs of
regions to functions of regions. The connecting graphon functions are
considered unknown but assumed smoothness allows putting priors of low
complexity on them. We pursue a nonparametric Bayesian approach by assigning a
Dirichlet process scale mixture of zero mean normal prior on the distributions
of the random effects and finite random series of tensor products of B-splines
priors on the underlying graphon functions. Markov chain Monte Carlo
techniques, for drawing samples for the posterior distributions are developed.
The proposed Bayesian method overwhelmingly outperforms similar ANCOVA models
in the simulation setup. The proposed Bayesian approach is applied on a dataset
of 100 subjects and 83 brain regions and key regions implicated in the changing
connectome are identified.
| stat |
Bayesian Topological Learning for Classifying the Structure of
Biological Networks | Actin cytoskeleton networks generate local topological signatures due to the
natural variations in the number, size, and shape of holes of the networks.
Persistent homology is a method that explores these topological properties of
data and summarizes them as persistence diagrams. In this work, we analyze and
classify these filament networks by transforming them into persistence diagrams
whose variability is quantified via a Bayesian framework on the space of
persistence diagrams. The proposed generalized Bayesian framework adopts an
independent and identically distributed cluster point process characterization
of persistence diagrams and relies on a substitution likelihood argument. This
framework provides the flexibility to estimate the posterior cardinality
distribution of points in a persistence diagram and the posterior spatial
distribution simultaneously. We present a closed form of the posteriors under
the assumption of Gaussian mixtures and binomials for prior intensity and
cardinality respectively. Using this posterior calculation, we implement a
Bayes factor algorithm to classify the actin filament networks and benchmark it
against several state-of-the-art classification methods.
| stat |
Accelerating Bayesian inference in hydrological modeling with a
mechanistic emulator | As in many fields of dynamic modeling, the long runtime of hydrological
models hinders Bayesian inference of model parameters from data. By replacing a
model with an approximation of its output as a function of input and/or
parameters, emulation allows us to complete this task by trading-off accuracy
for speed. We combine (i) the use of a mechanistic emulator, (ii)
low-discrepancy sampling of the parameter space, and (iii) iterative refinement
of the design data set, to perform Bayesian inference with a very small design
data set constructed with 128 model runs in a parameter space of up to eight
dimensions. In our didactic example we use a model implemented with the
hydrological simulator SWMM that allows us to compare our inference results
against those derived with the full model. This comparison demonstrates that
iterative improvements lead to reasonable results with a very small design data
set.
| stat |
Accuracy and stability of solar variable selection comparison under
complicated dependence structures | In this paper we focus on the empirical variable-selection peformance of
subsample-ordered least angle regression (Solar) -- a novel ultrahigh
dimensional redesign of lasso -- on the empirical data with complicated
dependence structures and, hence, severe multicollinearity and grouping effect
issues. Previous researches show that Solar largely alleviates several known
high-dimensional issues with least-angle regression and $\mathcal{L}_1$
shrinkage. Also, With the same computation load, solar yields substantiali
mprovements over two lasso solvers (least-angle regression for lasso and
coordinate-descent) in terms of the sparsity (37-64\% reduction in the average
number of selected variables), stability and accuracy of variable selection.
Simulations also demonstrate that solar enhances the robustness of variable
selection to different settings of the irrepresentable condition and to
variations in the dependence structures assumed in regression analysis. To
confirm that the improvements are also available for empirical researches, we
choose the prostate cancer data and the Sydney house price data and apply two
lasso solvers, elastic net and Solar on them for comparison. The results shows
that (i) lasso is affected by the grouping effect and randomly drop variables
with high correlations, resulting unreliable and uninterpretable results; (ii)
elastic net is more robust to grouping effect; however, it completely lose
variable-selection sparsity when the dependence structure of the data is
complicated; (iii) solar demonstrates its superior robustness to complicated
dependence structures and grouping effect, returning variable-selection results
with better stability and sparsity. The code can be found at
https://github.com/isaac2math/solar_application
| stat |
Bigger data, better questions, and a return to fourth down behavior: an
introduction to a special issue on tracking data in the National football
League | Most historical National Football League (NFL) analysis, both mainstream and
academic, has relied on public, play-level data to generate team and player
comparisons. Given the number of oft omitted variables that impact on-field
results, such as play call, game situation, and opponent strength, findings
tend to be more anecdotal than actionable. With the release of player tracking
data, however, analysts can better ask and answer questions to isolate skill
and strategy. In this article, we highlight the limitations of traditional
analyses, and use a decades-old punching bag for analysts, fourth-down
strategy, as a microcosm for why tracking data is needed. Specifically, we
assert that, in absence of using the precise yardage needed for a first down,
past findings supporting an aggressive fourth down strategy may have been
overstated. Next, we synthesize recent work that comprises this special Journal
of Quantitative Analysis in Sports issue into player tracking data in football.
Finally, we conclude with some best practices and limitations regarding usage
of this data. The release of player tracking data marks a transition for the
league and its' analysts, and we hope this issue helps guide innovation in
football analytics for years to come.
| stat |
Autoregressive Networks | We propose a first-order autoregressive model for dynamic network processes
in which edges change over time while nodes remain unchanged. The model depicts
the dynamic changes explicitly. It also facilitates simple and efficient
statistical inference such as the maximum likelihood estimators which are
proved to be (uniformly) consistent and asymptotically normal. The model
diagnostic checking can be carried out easily using a permutation test. The
proposed model can apply to any network processes with various underlying
structures but with independent edges. As an illustration, an autoregressive
stochastic block model has been investigated in depth, which characterizes the
latent communities by the transition probabilities over time. This leads to a
more effective spectral clustering algorithm for identifying the latent
communities. Inference for a change point is incorporated into the
autoregressive stochastic block model to cater for possible structure changes.
The developed asymptotic theory as well as the simulation study affirms the
performance of the proposed methods. Application with three real data sets
illustrates both relevance and usefulness of the proposed models.
| stat |
A non-inferiority test for R-squared with random regressors | Determining the lack of association between an outcome variable and a number
of different explanatory variables is frequently necessary in order to
disregard a proposed model. This paper proposes a non-inferiority test for the
coefficient of determination (or squared multiple correlation coefficient),
R-squared, in a linear regression analysis with random predictors. The test is
derived from inverting a one-sided confidence interval based on a scaled
central F distribution.
| stat |
Probabilistic Forecasting of the Arctic Sea Ice Edge with Contour
Modeling | Sea ice, or frozen ocean water, freezes and melts every year in the Arctic.
Forecasts of where sea ice will be located weeks to months in advance have
become more important as the amount of sea ice declines due to climate change,
for maritime planning and other uses. Typical sea ice forecasts are made with
ensemble models, physics-based models of sea ice and the surrounding ocean and
atmosphere. This paper introduces Mixture Contour Forecasting, a method to
forecast sea ice probabilistically using a mixture of two distributions, one
based on post-processed output from ensembles and the other on observed sea ice
patterns in recent years. At short lead times, these forecasts are better
calibrated than unadjusted dynamic ensemble forecasts and other statistical
reference forecasts. To produce these forecasts, a statistical technique is
introduced that directly models the sea ice edge contour, the boundary around
the region that is ice-covered. Mixture Contour Forecasting and reference
methods are evaluated for monthly sea ice forecasts for 2008-2016 at lead times
ranging from 0.5-6.5 months using one of the European Centre for Medium-Range
Weather Forecasts ensembles.
| stat |
Using Bayesian latent Gaussian graphical models to infer symptom
associations in verbal autopsies | Learning dependence relationships among variables of mixed types provides
insights in a variety of scientific settings and is a well-studied problem in
statistics. Existing methods, however, typically rely on copious, high quality
data to accurately learn associations. In this paper, we develop a method for
scientific settings where learning dependence structure is essential, but data
are sparse and have a high fraction of missing values. Specifically, our work
is motivated by survey-based cause of death assessments known as verbal
autopsies (VAs). We propose a Bayesian approach to characterize dependence
relationships using a latent Gaussian graphical model that incorporates
informative priors on the marginal distributions of the variables. We
demonstrate such information can improve estimation of the dependence
structure, especially in settings with little training data. We show that our
method can be integrated into existing probabilistic cause-of-death assignment
algorithms and improves model performance while recovering dependence patterns
between symptoms that can inform efficient questionnaire design in future data
collection.
| stat |
Recurrent Deep Divergence-based Clustering for simultaneous feature
learning and clustering of variable length time series | The task of clustering unlabeled time series and sequences entails a
particular set of challenges, namely to adequately model temporal relations and
variable sequence lengths. If these challenges are not properly handled, the
resulting clusters might be of suboptimal quality. As a key solution, we
present a joint clustering and feature learning framework for time series based
on deep learning. For a given set of time series, we train a recurrent network
to represent, or embed, each time series in a vector space such that a
divergence-based clustering loss function can discover the underlying cluster
structure in an end-to-end manner. Unlike previous approaches, our model
inherently handles multivariate time series of variable lengths and does not
require specification of a distance-measure in the input space. On a diverse
set of benchmark datasets we illustrate that our proposed Recurrent Deep
Divergence-based Clustering approach outperforms, or performs comparable to,
previous approaches.
| stat |
TVOR: Finding Discrete Total Variation Outliers among Histograms | Pearson's chi-squared test can detect outliers in the data distribution of a
given set of histograms. However, in fields such as demographics (for e.g.
birth years), outliers may be more easily found in terms of the histogram
smoothness where techniques such as Whipple's or Myers' indices handle
successfully only specific anomalies. This paper proposes smoothness outliers
detection among histograms by using the relation between their discrete total
variations (DTV) and their respective sample sizes. This relation is
mathematically derived to be applicable in all cases and simplified by an
accurate linear model. The deviation of the histogram's DTV from the value
predicted by the model is used as the outlier score and the proposed method is
named Total Variation Outlier Recognizer (TVOR). TVOR requires no prior
assumptions about the histograms' samples' distribution, it has no
hyperparameters that require tuning, it is not limited to only specific
patterns, and it is applicable to histograms with the same bins. Each bin can
have an arbitrary interval that can also be unbounded. TVOR finds DTV outliers
easier than Pearson's chi-squared test. In case of distribution outliers, the
opposite holds. TVOR is tested on real census data and it successfully finds
suspicious histograms. The source code is given at
https://github.com/DiscreteTotalVariation/TVOR.
| stat |
Introducing the treatment hierarchy question in network meta-analysis | Background: Comparative effectiveness research using network meta-analysis
can present a hierarchy of competing treatments, from the least to most
preferable option. However, the research question associated with the hierarchy
of multiple interventions is never clearly defined in published reviews.
Methods and Results: We introduce the notion of a treatment hierarchy question
that describes the criterion for choosing a specific treatment over one or more
competing alternatives. For example, stakeholders might ask which treatment is
most likely to improve mean survival by at least 2 years or which treatment is
associated with the longest mean survival. The answers to these two questions
are not necessarily the same. We discuss the most commonly used ranking metrics
(quantities that describe or compare the estimated treatment-specific effects),
how the metrics produce a treatment hierarchy and the type of treatment
hierarchy question that each metric can answer. We show that the ranking
metrics encompass the uncertainty in the estimation of the treatment effects in
different ways, which results in different treatment hierarchies. Conclusions:
Network meta-analyses that aim to rank treatments should state in the protocol
the treatment hierarchy question they aim to address and employ the appropriate
ranking metric to answer it.
| stat |
Prediction of the Subjective Impression of Passenger Car Roll Dynamics
on the Driver Based on Frequency-Domain Characteristic Values | Characteristic values are essential for the design and assessment of driving
dynamics during the early stages of the development process of passenger cars.
Compared to other aspects of vehicle dynamics however, the relationship between
measurable parameters and the subjective perception of vehicle roll dynamics
has not been researched extensively. In this paper, a study is presented in
which several variants of a vehicle with an electronically controlled
suspension were rated by test subjects regarding its roll dynamics and measured
in a standardized driving manoeuvre. The resulting subjective ratings and
objective characteristic values are then used to derive models to predict the
subjective liking of several roll dynamics aspects based on objective
frequency-domain parameters. Finally, the resulting prediction models are
validated using measurements of additional vehicles.
| stat |
Evaluating Ensemble Post-Processing for Wind Power Forecasts | Capturing the uncertainty in probabilistic wind power forecasts is
challenging, especially when uncertain input variables, such as the weather
play a role. Since ensemble weather predictions aim to capture this uncertainty
in the weather system accurately, they can be used to propagate this
uncertainty through to subsequent wind power forecasting models. However, as
weather ensemble systems are known to be biased and underdispersed,
meteorologists post-process the ensembles. This post-processing can
successfully correct the biases in the weather variables but has not been
evaluated thoroughly in the context of subsequent forecasts, such as wind power
generation.
The present paper evaluates multiple strategies for applying ensemble
post-processing to probabilistic wind power forecasts. We use Ensemble Model
Output Statistics (EMOS) as the post-processing method and evaluate four
possible strategies: only using the raw ensembles without post-processing, a
one-step strategy where only the weather ensembles are post-processed, a
one-step strategy where we only post-process the power ensembles, and a
two-step strategy where we post-process both the weather and power ensembles.
Results show that post-processing improves the probabilistic forecasting
accuracy and that the post-processing of the final power ensemble forecast is
the crucial step.
| stat |
D2P-Fed: Differentially Private Federated Learning With Efficient
Communication | In this paper, we propose the discrete Gaussian based differentially private
federated learning (D2P-Fed), a unified scheme to achieve both differential
privacy (DP) and communication efficiency in federated learning (FL). In
particular, compared with the only prior work taking care of both aspects,
D2P-Fed provides stronger privacy guarantee, better composability and smaller
communication cost. The key idea is to apply the discrete Gaussian noise to the
private data transmission. We provide complete analysis of the privacy
guarantee, communication cost and convergence rate of D2P-Fed. We evaluated
D2P-Fed on INFIMNIST and CIFAR10. The results show that D2P-Fed outperforms
the-state-of-the-art by 4.7% to 13.0% in terms of model accuracy while saving
one third of the communication cost.
| stat |
Dynamic principal component regression for forecasting functional time
series in a group structure | When generating social policies and pricing annuity at national and
subnational levels, it is essential both to forecast mortality accurately and
ensure that forecasts at the subnational level add up to the forecasts at the
national level. This has motivated recent developments in forecasting
functional time series in a group structure, where static principal component
analysis is used. In the presence of moderate to strong temporal dependence,
static principal component analysis designed for independent and identically
distributed functional data may be inadequate. Thus, through using the dynamic
functional principal component analysis, we consider a functional time series
forecasting method with static and dynamic principal component regression to
forecast each series in a group structure. Through using the regional
age-specific mortality rates in Japan obtained from the Japanese Mortality
Database (2019), we investigate the point and interval forecast accuracies of
our proposed extension, and subsequently make recommendations.
| stat |
Learning spatially-correlated temporal dictionaries for calcium imaging | Calcium imaging has become a fundamental neural imaging technique, aiming to
recover the individual activity of hundreds of neurons in a cortical region.
Current methods (mostly matrix factorization) are aimed at detecting neurons in
the field-of-view and then inferring the corresponding time-traces. In this
paper, we reverse the modeling and instead aim to minimize the spatial
inference, while focusing on finding the set of temporal traces present in the
data. We reframe the problem in a dictionary learning setting, where the
dictionary contains the time-traces and the sparse coefficient are spatial
maps. We adapt dictionary learning to calcium imaging by introducing
constraints on the norms and correlations of the time-traces, and incorporating
a hierarchical spatial filtering model that correlates the time-trace usage
over the field-of-view. We demonstrate on synthetic and real data that our
solution has advantages regarding initialization, implicitly inferring number
of neurons and simultaneously detecting different neuronal types.
| stat |
A Priori Estimates of the Population Risk for Two-layer Neural Networks | New estimates for the population risk are established for two-layer neural
networks. These estimates are nearly optimal in the sense that the error rates
scale in the same way as the Monte Carlo error rates. They are equally
effective in the over-parametrized regime when the network size is much larger
than the size of the dataset. These new estimates are a priori in nature in the
sense that the bounds depend only on some norms of the underlying functions to
be fitted, not the parameters in the model, in contrast with most existing
results which are a posteriori in nature. Using these a priori estimates, we
provide a perspective for understanding why two-layer neural networks perform
better than the related kernel methods.
| stat |
Elastic Coupled Co-clustering for Single-Cell Genomic Data | The recent advances in single-cell technologies have enabled us to profile
genomic features at unprecedented resolution and datasets from multiple domains
are available, including datasets that profile different types of genomic
features and datasets that profile the same type of genomic features across
different species. These datasets typically have different powers in
identifying the unknown cell types through clustering, and data integration can
potentially lead to a better performance of clustering algorithms. In this
work, we formulate the problem in an unsupervised transfer learning framework,
which utilizes knowledge learned from auxiliary dataset to improve the
clustering performance of target dataset. The degree of shared information
among the target and auxiliary datasets can vary, and their distributions can
also be different. To address these challenges, we propose an elastic coupled
co-clustering based transfer learning algorithm, by elastically propagating
clustering knowledge obtained from the auxiliary dataset to the target dataset.
Implementation on single-cell genomic datasets shows that our algorithm greatly
improves clustering performance over the traditional learning algorithms. The
source code and data sets are available at
https://github.com/cuhklinlab/elasticC3.
| stat |
Drug-Drug Adverse Effect Prediction with Graph Co-Attention | Complex or co-existing diseases are commonly treated using drug combinations,
which can lead to higher risk of adverse side effects. The detection of
polypharmacy side effects is usually done in Phase IV clinical trials, but
there are still plenty which remain undiscovered when the drugs are put on the
market. Such accidents have been affecting an increasing proportion of the
population (15% in the US now) and it is thus of high interest to be able to
predict the potential side effects as early as possible. Systematic
combinatorial screening of possible drug-drug interactions (DDI) is challenging
and expensive. However, the recent significant increases in data availability
from pharmaceutical research and development efforts offer a novel paradigm for
recovering relevant insights for DDI prediction. Accordingly, several recent
approaches focus on curating massive DDI datasets (with millions of examples)
and training machine learning models on them. Here we propose a neural network
architecture able to set state-of-the-art results on this task---using the type
of the side-effect and the molecular structure of the drugs alone---by
leveraging a co-attentional mechanism. In particular, we show the importance of
integrating joint information from the drug pairs early on when learning each
drug's representation.
| stat |
Causal Discovery of a River Network from its Extremes | Causal inference for extremes aims to discover cause and effect relations
between large observed values of random variables. Over the last years, a
number of methods have been proposed for solving the Hidden River Problem, with
the Danube data set as benchmark. In this paper, we provide \QTree, a new and
simple algorithm to solve the Hidden River Problem that outperforms existing
methods. \QTree\ returns a directed graph and achieves almost perfect recovery
on the Danube as well as on new data from the Lower Colorado River. It can
handle missing data, has an automated parameter tuning procedure, and runs in
time $O(n |V|^2)$, where $n$ is the number of observations and $|V|$ the number
of nodes in the graph. \QTree\ relies on qualitative aspects of the max-linear
Bayesian network model.
| stat |
Exact simulation of the Wright-Fisher diffusion | The Wright-Fisher family of diffusion processes is a widely used class of
evolutionary models. However, simulation is difficult because there is no known
closed-form formula for its transition function. In this article we demonstrate
that it is in fact possible to simulate exactly from a broad class of
Wright-Fisher diffusion processes and their bridges. For those diffusions
corresponding to reversible, neutral evolution, our key idea is to exploit an
eigenfunction expansion of the transition function; this approach even applies
to its infinite-dimensional analogue, the Fleming-Viot process. We then develop
an exact rejection algorithm for processes with more general drift functions,
including those modelling natural selection, using ideas from retrospective
simulation. Our approach also yields methods for exact simulation of the moment
dual of the Wright-Fisher diffusion, the ancestral process of an infinite-leaf
Kingman coalescent tree. We believe our new perspective on diffusion simulation
holds promise for other models admitting a transition eigenfunction expansion.
| stat |
Tuning parameter calibration for prediction in personalized medicine | Personalized medicine has become an important part of medicine, for instance
predicting individual drug responses based on genomic information. However,
many current statistical methods are not tailored to this task, because they
overlook the individual heterogeneity of patients. In this paper, we look at
personalized medicine from a linear regression standpoint. We introduce an
alternative version of the ridge estimator and target individuals by
establishing a tuning parameter calibration scheme that minimizes prediction
errors of individual patients. In stark contrast, classical schemes such as
cross-validation minimize prediction errors only on average. We show that our
pipeline is optimal in terms of oracle inequalities, fast, and highly effective
both in simulations and on real data.
| stat |
Robust Sparse Bayesian Infinite Factor Models | Most of previous works and applications of Bayesian factor model have assumed
the normal likelihood regardless of its validity. We propose a Bayesian factor
model for heavy-tailed high-dimensional data based on multivariate Student-$t$
likelihood to obtain better covariance estimation. We use multiplicative gamma
process shrinkage prior and factor number adaptation scheme proposed in
Bhattacharya & Dunson [Biometrika (2011) 291-306]. Since a naive Gibbs sampler
for the proposed model suffers from slow mixing, we propose a Markov Chain
Monte Carlo algorithm where fast mixing of Hamiltonian Monte Carlo is exploited
for some parameters in proposed model. Simulation results illustrate the gain
in performance of covariance estimation for heavy-tailed high-dimensional data.
We also provide a theoretical result that the posterior of the proposed model
is weakly consistent under reasonable conditions. We conclude the paper with
the application of proposed factor model on breast cancer metastasis prediction
given DNA signature data of cancer cell.
| stat |
Lasso tuning through the flexible-weighted bootstrap | Regularized regression approaches such as the Lasso have been widely adopted
for constructing sparse linear models in high-dimensional datasets. A
complexity in fitting these models is the tuning of the parameters which
control the level of introduced sparsity through penalization. The most common
approach to select the penalty parameter is through $k$-fold cross-validation.
While cross-validation is used to minimise the empirical prediction error,
approaches such as the $m$-out-of-$n$ paired bootstrap which use smaller
training datasets provide consistency in selecting the non-zero coefficients in
the oracle model, performing well in an asymptotic setting but having
limitations when $n$ is small. In fact, for models such as the Lasso there is a
monotonic relationship between the size of training sets and the penalty
parameter. We propose a generalization of these methods for selecting the
regularization parameter based on a flexible-weighted bootstrap procedure that
mimics the $m$-out-of-$n$ bootstrap and overcomes its challenges for all sample
sizes. Through simulation studies we demonstrate that when selecting a penalty
parameter, the choice of weights in the bootstrap procedure can be used to
dictate the size of the penalty parameter and hence the sparsity of the fitted
model. We empirically illustrate our weighted bootstrap procedure by applying
the Lasso to integrate clinical and microRNA data in the modeling of
Alzheimer's disease. In both the real and simulated data we find a narrow part
of the parameter space to perform well, emulating an $m$-out-of-$n$ bootstrap,
and that our procedure can be used to improve interpretation of other
optimization heuristics.
| stat |
Bayesian Modeling of Spatial Molecular Profiling Data via Gaussian
Process | The location, timing, and abundance of gene expression (both mRNA and
proteins) within a tissue define the molecular mechanisms of cell functions.
Recent technology breakthroughs in spatial molecular profiling, including
imaging-based technologies and sequencing-based technologies, have enabled the
comprehensive molecular characterization of single cells while preserving their
spatial and morphological contexts. This new bioinformatics scenario calls for
effective and robust computational methods to identify genes with spatial
patterns. We represent a novel Bayesian hierarchical model to analyze spatial
transcriptomics data, with several unique characteristics. It models the
zero-inflated and over-dispersed counts by deploying a zero-inflated negative
binomial model that greatly increases model stability and robustness. Besides,
the Bayesian inference framework allows us to borrow strength in parameter
estimation in a de novo fashion. As a result, the proposed model shows
competitive performances in accuracy and robustness over existing methods in
both simulation studies and two real data applications. The related R/C++
source code is available at https://github.com/Minzhe/BOOST-GP.
| stat |
Bias-Variance Trade-off and Overlearning in Dynamic Decision Problems | Modern Monte Carlo-type approaches to dynamic decision problems are
reformulated as empirical loss minimization, allowing direct applications of
classical results from statistical machine learning. These computational
methods are then analyzed in this framework to demonstrate their effectiveness
as well as their susceptibility to generalization error. Standard uses of
classical results prove potential overlearning, thus bias-variance trade-off,
by connecting over-trained networks to anticipating controls. On the other
hand, non-asymptotic estimates based on Rademacher complexity show the
convergence of these algorithms for sufficiently large training sets. A
numerically studied stylized example illustrates these possibilities, including
the importance of problem dimension in the degree of overlearning, and the
effectiveness of this approach.
| stat |
Population pharmacokinetics of levobupivacaine during a transversus
abdominis plane block in children | BACKGROUND:Levobupivacaine is commonly used during transversus abdominis
plane block in pediatric patients. However, the dosing regimen is still
empirical, and the pharmacokinetic properties of levobupivacaine are not
considered. Here, the pharmacokinetics of levobupivacaine during an
ultrasound-guided transversus abdominis plane block were evaluated to optimize
dosing regimen, with regard to the between-subject variability and the volume
of levobupivacaine injected.METHOD:The clinical trial (prospective, randomized,
double-blind study protocol) was conducted in 40 children aged 1 to 5 years,
who were scheduled for inguinal surgery. Each patient received 0.4 mg/kg of
levobupivacaine with a volume of local anesthesia solution adjusted to 0.2
mL/kg of 0.2% or 0.4 mL/kg of 0.1% levobupivacaine. Blood samples were
collected at 5, 15, 20, 25, 30, 45, 60, and 75 min following the block
injection. The population pharmacokinetic analysis was performed using the
NONMEM software.RESULTS:From the pharmacokinetic parameters obtained, median
Cmax, tmax, and area under the concentration versus time curve were 0.315 mg/L,
17 min, and 41 mg/L. min, respectively. Between-subject variability (BSV) of
clearance was explained by weight. At the dose regimen of 0.4 mg/kg, none of
the infants showed signs of toxicity, but in 13 patients, transversus abdominis
plane block failed. After analysis, BSV for absorption rate constant,
distribution volume, and clearance were 81%, 47%, and 41%, respectively.
Residual unexplained variability was estimated to be 14%.CONCLUSION:For
improved efficiency in the pediatric population, the dose of levobupivacaine
should be greater than 0.4 mg/kg. Children's weight should be considered to
anticipate any risk of toxicity.
| stat |
Indirect Gaussian Graph Learning beyond Gaussianity | This paper studies how to capture dependency graph structures from real data
which may not be Gaussian. Starting from marginal loss functions not
necessarily derived from probability distributions, we utilize an additive
over-parametrization with shrinkage to incorporate variable dependencies into
the criterion. An iterative Gaussian graph learning algorithm is proposed with
ease in implementation. Statistical analysis shows that the estimators achieve
satisfactory accuracy with the error measured in terms of a proper Bregman
divergence. Real-life examples in different settings are given to demonstrate
the efficacy of the proposed methodology.
| stat |
The sceptical Bayes factor for the assessment of replication success | There is an urgent need to develop new methodology for the design and
analysis of replication studies. Recently, a reverse-Bayes method called the
sceptical $p$-value has been proposed for this purpose; the inversion of Bayes'
theorem allows us to mathematically formalise the notion of scepticism, which
in turn can be used to assess the agreement between the findings of an original
study and its replication. However, despite its Bayesian nature, the method
relies on tail probabilities as primary inference tools. Here, we present an
extension that uses Bayes factors as an alternative means of quantifying
evidence. This leads to a new measure for evaluating replication success, the
sceptical Bayes factor: Conceptually, the sceptical Bayes factor provides a
bound for the maximum level of evidence at which an advocate of the original
finding can convince a sceptic who does not trust it, in light of the
replication data. While the sceptical $p$-value can only quantify the conflict
between the sceptical prior and the observed replication data, the sceptical
Bayes factor also takes into account how likely the data are under the
posterior distribution of the effect conditional on the original study,
allowing for stronger statements about replication success. Moreover, the
proposed method elegantly combines traditional notions of replication success;
it ensures that both studies need to show evidence against the null, while at
the same time penalising incompatibility of their effect estimates. Case
studies from the Reproducibility Project: Cancer Biology and the Social
Sciences Replication Project show the advantages of the method for the
quantitative assessment of replicability.
| stat |
Scaling Bayesian inference of mixed multinomial logit models to very
large datasets | Variational inference methods have been shown to lead to significant
improvements in the computational efficiency of approximate Bayesian inference
in mixed multinomial logit models when compared to standard Markov-chain Monte
Carlo (MCMC) methods without compromising accuracy. However, despite their
demonstrated efficiency gains, existing methods still suffer from important
limitations that prevent them to scale to very large datasets, while providing
the flexibility to allow for rich prior distributions and to capture complex
posterior distributions. In this paper, we propose an Amortized Variational
Inference approach that leverages stochastic backpropagation, automatic
differentiation and GPU-accelerated computation, for effectively scaling
Bayesian inference in Mixed Multinomial Logit models to very large datasets.
Moreover, we show how normalizing flows can be used to increase the flexibility
of the variational posterior approximations. Through an extensive simulation
study, we empirically show that the proposed approach is able to achieve
computational speedups of multiple orders of magnitude over traditional MSLE
and MCMC approaches for large datasets without compromising estimation
accuracy.
| stat |
A physics-aware, probabilistic machine learning framework for
coarse-graining high-dimensional systems in the Small Data regime | The automated construction of coarse-grained models represents a pivotal
component in computer simulation of physical systems and is a key enabler in
various analysis and design tasks related to uncertainty quantification.
Pertinent methods are severely inhibited by the high-dimension of the
parametric input and the limited number of training input/output pairs that can
be generated when computationally demanding forward models are considered. Such
cases are frequently encountered in the modeling of random heterogeneous media
where the scale of the microstructure necessitates the use of high-dimensional
random vectors and very fine discretizations of the governing equations. The
present paper proposes a probabilistic Machine Learning framework that is
capable of operating in the presence of Small Data by exploiting aspects of the
physical structure of the problem as well as contextual knowledge. As a result,
it can perform comparably well under extrapolative conditions. It unifies the
tasks of dimensionality and model-order reduction through an encoder-decoder
scheme that simultaneously identifies a sparse set of salient lower-dimensional
microstructural features and calibrates an inexpensive, coarse-grained model
which is predictive of the output. Information loss is accounted for and
quantified in the form of probabilistic predictive estimates. The learning
engine is based on Stochastic Variational Inference. We demonstrate how the
variational objectives can be used not only to train the coarse-grained model,
but also to suggest refinements that lead to improved predictions.
| stat |
Functional Regularisation for Continual Learning with Gaussian Processes | We introduce a framework for Continual Learning (CL) based on Bayesian
inference over the function space rather than the parameters of a deep neural
network. This method, referred to as functional regularisation for Continual
Learning, avoids forgetting a previous task by constructing and memorising an
approximate posterior belief over the underlying task-specific function. To
achieve this we rely on a Gaussian process obtained by treating the weights of
the last layer of a neural network as random and Gaussian distributed. Then,
the training algorithm sequentially encounters tasks and constructs posterior
beliefs over the task-specific functions by using inducing point sparse
Gaussian process methods. At each step a new task is first learnt and then a
summary is constructed consisting of (i) inducing inputs -- a fixed-size subset
of the task inputs selected such that it optimally represents the task -- and
(ii) a posterior distribution over the function values at these inputs. This
summary then regularises learning of future tasks, through Kullback-Leibler
regularisation terms. Our method thus unites approaches focused on
(pseudo-)rehearsal with those derived from a sequential Bayesian inference
perspective in a principled way, leading to strong results on accepted
benchmarks.
| stat |
Optimal $2^K$ Paired Comparison Designs for Third-Order Interactions | In psychological research often paired comparisons are used in which either
full or partial profiles of the alternatives described by a common set of
two-level attributes are presented. For this situation the problem of finding
optimal designs is considered in the presence of third-order interactions.
| stat |
Convergence Guarantees for Adaptive Bayesian Quadrature Methods | Adaptive Bayesian quadrature (ABQ) is a powerful approach to numerical
integration that empirically compares favorably with Monte Carlo integration on
problems of medium dimensionality (where non-adaptive quadrature is not
competitive). Its key ingredient is an acquisition function that changes as a
function of previously collected values of the integrand. While this adaptivity
appears to be empirically powerful, it complicates analysis. Consequently,
there are no theoretical guarantees so far for this class of methods. In this
work, for a broad class of adaptive Bayesian quadrature methods, we prove
consistency, deriving non-tight but informative convergence rates. To do so we
introduce a new concept we call weak adaptivity. Our results identify a large
and flexible class of adaptive Bayesian quadrature rules as consistent, within
which practitioners can develop empirically efficient methods.
| stat |
Robust Correction of Sampling Bias Using Cumulative Distribution
Functions | Varying domains and biased datasets can lead to differences between the
training and the target distributions, known as covariate shift. Current
approaches for alleviating this often rely on estimating the ratio of training
and target probability density functions. These techniques require parameter
tuning and can be unstable across different datasets. We present a new method
for handling covariate shift using the empirical cumulative distribution
function estimates of the target distribution by a rigorous generalization of a
recent idea proposed by Vapnik and Izmailov. Further, we show experimentally
that our method is more robust in its predictions, is not reliant on parameter
tuning and shows similar classification performance compared to the current
state-of-the-art techniques on synthetic and real datasets.
| stat |
A simple correction for COVID-19 sampling bias | COVID-19 testing has become a standard approach for estimating prevalence
which then assist in public health decision making to contain and mitigate the
spread of the disease. The sampling designs used are often biased in that they
do not reflect the true underlying populations. For instance, individuals with
strong symptoms are more likely to be tested than those with no symptoms. This
results in biased estimates of prevalence (too high). Typical post-sampling
corrections are not always possible. Here we present a simple bias correction
methodology derived and adapted from a correction for publication bias in meta
analysis studies. The methodology is general enough to allow a wide variety of
customization making it more useful in practice. Implementation is easily done
using already collected information. Via a simulation and two real datasets, we
show that the bias corrections can provide dramatic reductions in estimation
error.
| stat |
Variance decompositions for extensive-form games | Quantitative measures of randomness in games are useful for game design and
have implications for gambling law. We treat the outcome of a game as a random
variable and derive a closed-form expression and estimator for the variance in
the outcome attributable to a player of the game. We analyze poker hands to
show that randomness in the cards dealt has little influence on the outcomes of
each hand. A simple example is given to demonstrate how variance decompositions
can be used to measure other interesting properties of games.
| stat |
Variable selection for transportability | Transportability provides a principled framework to address the problem of
applying study results to new populations. Here, we consider the problem of
selecting variables to include in transport estimators. We provide a brief
overview of the transportability framework and illustrate that while selection
diagrams are a vital first step in variable selection, these graphs alone
identify a sufficient but not strictly necessary set of variables for
generating an unbiased transport estimate. Next, we conduct a simulation
experiment assessing the impact of including unnecessary variables on the
performance of the parametric g-computation transport estimator. Our results
highlight that the types of variables included can affect the bias, variance,
and mean squared error of the estimates. We find that addition of variables
that are not causes of the outcome but whose distributions differ between the
source and target populations can increase the variance and mean squared error
of the transported estimates. On the other hand, inclusion of variables that
are causes of the outcome (regardless of whether they modify the causal
contrast of interest or differ in distribution between the populations) reduces
the variance of the estimates without increasing the bias. Finally, exclusion
of variables that cause the outcome but do not modify the causal contrast of
interest does not increase bias. These findings suggest that variable selection
approaches for transport should prioritize identifying and including all causes
of the outcome in the study population rather than focusing on variables whose
distribution may differ between the study sample and target population.
| stat |
Multi-Decoder RNN Autoencoder Based on Variational Bayes Method | Clustering algorithms have wide applications and play an important role in
data analysis fields including time series data analysis. However, in time
series analysis, most of the algorithms used signal shape features or the
initial value of hidden variable of a neural network. Little has been discussed
on the methods based on the generative model of the time series. In this paper,
we propose a new clustering algorithm focusing on the generative process of the
signal with a recurrent neural network and the variational Bayes method. Our
experiments show that the proposed algorithm not only has a robustness against
for phase shift, amplitude and signal length variations but also provide a
flexible clustering based on the property of the variational Bayes method.
| stat |
Machine Learning vs Statistical Methods for Time Series Forecasting:
Size Matters | Time series forecasting is one of the most active research topics. Machine
learning methods have been increasingly adopted to solve these predictive
tasks. However, in a recent work, these were shown to systematically present a
lower predictive performance relative to simple statistical methods. In this
work, we counter these results. We show that these are only valid under an
extremely low sample size. Using a learning curve method, our results suggest
that machine learning methods improve their relative predictive performance as
the sample size grows. The code to reproduce the experiments is available at
https://github.com/vcerqueira/MLforForecasting.
| stat |
Analyzing CART | Decision trees with binary splits are popularly constructed using
Classification and Regression Trees (CART) methodology. For binary
classification and regression models, this approach recursively divides the
data into two near-homogenous daughter nodes according to a split point that
maximizes the reduction in sum of squares error (the impurity) along a
particular variable. This paper aims to study the bias and adaptive properties
of regression trees constructed with CART. In doing so, we derive an
interesting connection between the bias and the mean decrease in impurity (MDI)
measure of variable importance---a tool widely used for model
interpretability---defined as the sum of impurity reductions over all
non-terminal nodes in the tree. In particular, we show that the probability
content of a terminal subnode for a variable is small when the MDI for that
variable is large and that this relationship is exponential---confirming
theoretically that decision trees with CART have small bias and are adaptive to
signal strength and direction. Finally, we apply these individual tree bounds
to tree ensembles and show consistency of Breiman's random forests. The context
is surprisingly general and applies to a wide variety of multivariable data
generating distributions and regression functions. The main technical tool is
an exact characterization of the conditional probability content of the
daughter nodes arising from an optimal split, in terms of the partial
dependence function and reduction in impurity.
| stat |
Random Partitioning and Distribution-based Thresholding for Iterative
Variable Screening in High Dimensions | In big data analysis, a simple task such as linear regression can become very
challenging as the variable dimension $p$ grows. As a result, variable
screening is inevitable in many scientific studies. In recent years, randomized
algorithms have become a new trend and are playing an increasingly important
role for large scale data analysis. In this article, we combine the ideas of
variable screening and random partitioning to propose a new iterative variable
screening method. For moderate sized $p$ of order $O(n^{2-\delta})$, we propose
a basic algorithm that adopts a distribution-based thresholding rule. For very
large $p$, we further propose a two-stage procedure. This two-stage procedure
first performs a random partitioning to divide predictors into subsets of
manageable size of order $O(n^{2-\delta})$ for variable screening, where
$\delta >0$ can be an arbitrarily small positive number. Random partitioning is
repeated a few times. Next, the final estimate of variable subset is obtained
by integrating results obtained from multiple random partitions. Simulation
studies show that our method works well and outperforms some renowned
competitors. Real data applications are presented. Our algorithms are able to
handle predictors in the size of millions.
| stat |
Time-Varying Coefficient Model Estimation Through Radial Basis Functions | In this paper we estimate the dynamic parameters of a time-varying
coefficient model through radial kernel functions in the context of a
longitudinal study. Our proposal is based on a linear combination of weighted
kernel functions involving a bandwidth, centered around a given set of time
points. In addition, we study different alternatives of estimation and
inference including a Frequentist approach using weighted least squares along
with bootstrap methods, and a Bayesian approach through both Markov chain Monte
Carlo and variational methods. We compare the estimation strategies mention
above with each other, and our radial kernel functions proposal with an
expansion based on regression spline, by means of an extensive simulation study
considering multiples scenarios in terms of sample size, number of repeated
measurements, and subject-specific correlation. Our experiments show that the
capabilities of our proposal based on radial kernel functions are indeed
comparable with or even better than those obtained from regression splines. We
illustrate our methodology by analyzing data from two AIDS clinical studies.
| stat |
Semiparametric Imputation Using Conditional Gaussian Mixture Models
under Item Nonresponse | Imputation is a popular technique for handling item nonresponse in survey
sampling. Parametric imputation is based on a parametric model for imputation
and is less robust against the failure of the imputation model. Nonparametric
imputation is fully robust but is not applicable when the dimension of
covariates is large due to the curse of dimensionality. Semiparametric
imputation is another robust imputation based on a flexible model where the
number of model parameters can increase with the sample size. In this paper, we
propose another semiparametric imputation based on a more flexible model
assumption than the Gaussian mixture model. In the proposed mixture model, we
assume a conditional Gaussian model for the study variable given the auxiliary
variables, but the marginal distribution of the auxiliary variables is not
necessarily Gaussian. We show that the proposed mixture model achieves a lower
approximation error bound to any unknown target density than the Gaussian
mixture model in terms of the Kullback-Leibler divergence. The proposed method
is applicable to high dimensional covariate problem by including a penalty
function in the conditional log-likelihood function. The proposed method is
applied to 2017 Korean Household Income and Expenditure Survey conducted by
Statistics Korea. Supplementary material is available online.
| stat |
A Comparison Study on Nonlinear Dimension Reduction Methods with Kernel
Variations: Visualization, Optimization and Classification | Because of high dimensionality, correlation among covariates, and noise
contained in data, dimension reduction (DR) techniques are often employed to
the application of machine learning algorithms. Principal Component Analysis
(PCA), Linear Discriminant Analysis (LDA), and their kernel variants (KPCA,
KLDA) are among the most popular DR methods. Recently, Supervised Kernel
Principal Component Analysis (SKPCA) has been shown as another successful
alternative. In this paper, brief reviews of these popular techniques are
presented first. We then conduct a comparative performance study based on three
simulated datasets, after which the performance of the techniques are evaluated
through application to a pattern recognition problem in face image analysis.
The gender classification problem is considered on MORPH-II and FG-NET, two
popular longitudinal face aging databases. Several feature extraction methods
are used, including biologically-inspired features (BIF), local binary patterns
(LBP), histogram of oriented gradients (HOG), and the Active Appearance Model
(AAM). After applications of DR methods, a linear support vector machine (SVM)
is deployed with gender classification accuracy rates exceeding 95% on
MORPH-II, competitive with benchmark results. A parallel computational approach
is also proposed, attaining faster processing speeds and similar recognition
rates on MORPH-II. Our computational approach can be applied to practical
gender classification systems and generalized to other face analysis tasks,
such as race classification and age prediction.
| stat |
Learning Output Embeddings in Structured Prediction | A powerful and flexible approach to structured prediction consists in
embedding the structured objects to be predicted into a feature space of
possibly infinite dimension by means of output kernels, and then, solving a
regression problem in this output space. A prediction in the original space is
computed by solving a pre-image problem. In such an approach, the embedding,
linked to the target loss, is defined prior to the learning phase. In this
work, we propose to jointly learn a finite approximation of the output
embedding and the regression function into the new feature space. For that
purpose, we leverage a priori information on the outputs and also unexploited
unsupervised output data, which are both often available in structured
prediction problems. We prove that the resulting structured predictor is a
consistent estimator, and derive an excess risk bound. Moreover, the novel
structured prediction tool enjoys a significantly smaller computational
complexity than former output kernel methods. The approach empirically tested
on various structured prediction problems reveals to be versatile and able to
handle large datasets.
| stat |
Iterative Statistical Linear Regression for Gaussian Smoothing in
Continuous-Time Non-linear Stochastic Dynamic Systems | This paper considers approximate smoothing for discretely observed non-linear
stochastic differential equations. The problem is tackled by developing methods
for linearising stochastic differential equations with respect to an arbitrary
Gaussian process. Two methods are developed based on 1) taking the limit of
statistical linear regression of the discretised process and 2) minimising an
upper bound to a cost functional. Their difference is manifested in the
diffusion of the approximate processes. This in turn gives novel derivations of
pre-existing Gaussian smoothers when Method 1 is used and a new class of
Gaussian smoothers when Method 2 is used. Furthermore, based on the
aforementioned development the iterative Gaussian smoothers in discrete-time
are generalised to the continuous-time setting by iteratively re-linearising
the stochastic differential equation with respect to the current Gaussian
process approximation to the smoothed process. The method is verified in two
challenging tracking problems, a reentry problem and a radar tracked
coordinated turn model with state dependent diffusion. The results show that
the method has better estimation accuracy than state-of-the-art smoothers.
| stat |
Exploration of heterogeneous treatment effects via concave fusion | Understanding treatment heterogeneity is essential to the development of
precision medicine, which seeks to tailor medical treatments to subgroups of
patients with similar characteristics. One of the challenges to achieve this
goal is that we usually do not have a priori knowledge of the grouping
information of patients with respect to treatment. To address this problem, we
consider a heterogeneous regression model by assuming that the coefficient for
treatment variables are subject-dependent and belong to different subgroups
with unknown grouping information. We develop a concave fusion penalized method
for automatically estimating the grouping structure and the subgroup-specific
treatment effects, and derive an alternating direction method of multipliers
algorithm for its implementation. We also study the theoretical properties of
the proposed method and show that under suitable conditions there exists a
local minimizer that equals the oracle least squares estimator with a priori
knowledge of the true grouping information with high probability. This provides
theoretical support for making statistical inference about the
subgroup-specific treatment effects based on the proposed method. We evaluate
the performance of the proposed method by simulation studies and illustrate its
application by analyzing the data from the AIDS Clinical Trials Group Study.
| stat |
An Imputation model by Dirichlet Process Mixture of Elliptical Copulas
for Data of Mixed Type | Copula-based methods provide a flexible approach to build missing data
imputation models of multivariate data of mixed types. However, the choice of
copula function is an open question. We consider a Bayesian nonparametric
approach by using an infinite mixture of elliptical copulas induced by a
Dirichlet process mixture to build a flexible copula function. A slice sampling
algorithm is used to sample from the infinite dimensional parameter space. We
extend the work on prior parallel tempering used in finite mixture models to
the Dirichlet process mixture model to overcome the mixing issue in multimodal
distributions. Using simulations, we demonstrate that the infinite mixture
copula model provides a better overall fit compared to their single component
counterparts, and performs better at capturing tail dependence features of the
data. Simulations further show that our proposed model achieves more accurate
imputation especially for continuous variables and better inferential results
in some analytic models. The proposed model is applied to a medical data set of
acute stroke patients in Australia.
| stat |
Short Communication: Detecting Possibly Frequent Change-points: Wild
Binary Segmentation 2 | This article comments on the new version of wild binary segmentation 2. Wild
Binary Segmentation 2 and Steepest-drop Model Selection has made improvements
on changepoint analysis especially on reducing the computational cost. However,
WBS2 tends to overestimate as WBS and the threshold does not work appropriately
on short sequences without changepoints.
| stat |
Spectral Simulation of Functional Time Series | We develop methodology allowing to simulate a stationary functional time
series defined by means of its spectral density operators. Our framework is
general, in that it encompasses any such stationary functional time series,
whether linear or not. The methodology manifests particularly significant
computational gains if the spectral density operators are specified by means of
their eigendecomposition or as a filtering of white noise. In the special case
of linear processes, we determine the analytical expressions for the spectral
density operators of functional autoregressive (fractionally integrated) moving
average processes, and leverage these as part of our spectral approach, leading
to substantial improvements over time-domain simulation methods in some cases.
The methods are implemented as an R package (specsimfts) accompanied by several
demo files that are easy to modify and can be easily used by researchers aiming
to probe the finite-sample performance of their functional time series
methodology by means of simulation.
| stat |
Being Bayesian, Even Just a Bit, Fixes Overconfidence in ReLU Networks | The point estimates of ReLU classification networks---arguably the most
widely used neural network architecture---have been shown to yield arbitrarily
high confidence far away from the training data. This architecture, in
conjunction with a maximum a posteriori estimation scheme, is thus not
calibrated nor robust. Approximate Bayesian inference has been empirically
demonstrated to improve predictive uncertainty in neural networks, although the
theoretical analysis of such Bayesian approximations is limited. We
theoretically analyze approximate Gaussian distributions on the weights of ReLU
networks and show that they fix the overconfidence problem. Furthermore, we
show that even a simplistic, thus cheap, Bayesian approximation, also fixes
these issues. This indicates that a sufficient condition for a calibrated
uncertainty on a ReLU network is "to be a bit Bayesian". These theoretical
results validate the usage of last-layer Bayesian approximation and motivate a
range of a fidelity-cost trade-off. We further validate these findings
empirically via various standard experiments using common deep ReLU networks
and Laplace approximations.
| stat |
Probabilistic supervised learning | Predictive modelling and supervised learning are central to modern data
science. With predictions from an ever-expanding number of supervised black-box
strategies - e.g., kernel methods, random forests, deep learning aka neural
networks - being employed as a basis for decision making processes, it is
crucial to understand the statistical uncertainty associated with these
predictions.
As a general means to approach the issue, we present an overarching framework
for black-box prediction strategies that not only predict the target but also
their own predictions' uncertainty. Moreover, the framework allows for fair
assessment and comparison of disparate prediction strategies. For this, we
formally consider strategies capable of predicting full distributions from
feature variables, so-called probabilistic supervised learning strategies.
Our work draws from prior work including Bayesian statistics, information
theory, and modern supervised machine learning, and in a novel synthesis leads
to (a) new theoretical insights such as a probabilistic bias-variance
decomposition and an entropic formulation of prediction, as well as to (b) new
algorithms and meta-algorithms, such as composite prediction strategies,
probabilistic boosting and bagging, and a probabilistic predictive independence
test.
Our black-box formulation also leads (c) to a new modular interface view on
probabilistic supervised learning and a modelling workflow API design, which we
have implemented in the newly released skpro machine learning toolbox,
extending the familiar modelling interface and meta-modelling functionality of
sklearn. The skpro package provides interfaces for construction, composition,
and tuning of probabilistic supervised learning strategies, together with
orchestration features for validation and comparison of any such strategy - be
it frequentist, Bayesian, or other.
| stat |
Do NHL goalies get hot in the playoffs? A multilevel logistic regression
analysis | The hot-hand theory posits that an athlete who has performed well in the
recent past will perform better in the present. We use multilevel logistic
regression to test this theory for National Hockey League playoff goaltenders,
controlling for a variety of shot-related and game-related characteristics. Our
data consists of 48,431 shots for 93 goaltenders in the 2008-2016 playoffs.
Using a wide range of shot-based and time-based windows to quantify recent save
performance, we consistently find that good recent save performance has a
negative effect on the next-shot save probability, which contradicts the
hot-hand theory.
| stat |
Global Trends and Predictors of Face Mask Usage During the COVID-19
Pandemic | Background: Guidelines and recommendations from public health authorities
related to face masks have been essential in containing the COVID-19 pandemic.
We assessed the prevalence and correlates of mask usage during the pandemic.
Methods: We examined a total of 13,723,810 responses to a daily
cross-sectional representative online survey in 38 countries who completed from
April 23, 2020 to October 31, 2020 and reported having been in public at least
once during the last seven days. The outcome was individual face mask usage in
public settings, and the predictors were country fixed effects, country-level
mask policy stringency, calendar time, individual sociodemographic factors, and
health prevention behaviors. Associations were modelled using survey-weighted
multivariable logistic regression.
Findings: Mask-wearing varied over time and across the 38 countries. While
some countries consistently showed high prevalence throughout, in other
countries mask usage increased gradually, and a few other countries remained at
low prevalence. Controlling for time and country fixed effects,
sociodemographic factors (older age, female gender, education, urbanicity) and
stricter mask-related policies were significantly associated with higher mask
usage in public settings, while social behaviors considered risky in the
context of the pandemic (going out to large events, restaurants, shopping
centers, and socializing outside of the household) were associated with lower
mask use.
Interpretation: The decision to wear a face mask in public settings is
significantly associated with sociodemographic factors, risky social behaviors,
and mask policies. This has important implications for health prevention
policies and messaging, including the potential need for more targeted policy
and messaging design.
| stat |
On the Interplay Between Exposure Misclassification and Informative
Cluster Size | In this paper we study the impact of exposure misclassification when cluster
size is potentially informative (i.e., related to outcomes) and when
misclassification is differential by cluster size. First, we show that
misclassification in an exposure related to cluster size can induce
informativeness when cluster size would otherwise be non-informative. Second,
we show that misclassification that is differential by informative cluster size
can not only attenuate estimates of exposure effects but even inflate or
reverse the sign of estimates. To correct for bias in estimating marginal
parameters, we propose two frameworks: (i) an observed likelihood approach for
joint marginalized models of cluster size and outcomes and (ii) an expected
estimating equations approach. Although we focus on estimating marginal
parameters, a corollary is that the observed likelihood approach permits valid
inference for conditional parameters as well. Using data from the Nurses Health
Study II, we compare the results of the proposed correction methods when
applied to motivating data on the multigenerational effect of in-utero
diethylstilbestrol exposure on attention-deficit/hyperactivity disorder in
106,198 children of 47,450 nurses.
| stat |
A super scalable algorithm for short segment detection | In many applications such as copy number variant (CNV) detection, the goal is
to identify short segments on which the observations have different means or
medians from the background. Those segments are usually short and hidden in a
long sequence, and hence are very challenging to find. We study a super
scalable short segment (4S) detection algorithm in this paper. This
nonparametric method clusters the locations where the observations exceed a
threshold for segment detection. It is computationally efficient and does not
rely on Gaussian noise assumption. Moreover, we develop a framework to assign
significance levels for detected segments. We demonstrate the advantages of our
proposed method by theoretical, simulation, and real data studies.
| stat |
Restricted Spatial Regression Methods: Implications for Inference | The issue of spatial confounding between the spatial random effect and the
fixed effects in regression analyses has been identified as a concern in the
statistical literature. Multiple authors have offered perspectives and
potential solutions. In this paper, for the areal spatial data setting, we show
that many of the methods designed to alleviate spatial confounding can be
viewed as special cases of a general class of models. We refer to this class as
Restricted Spatial Regression (RSR) models, extending terminology currently in
use. We offer a mathematically based exploration of the impact that RSR methods
have on inference for regression coefficients for the linear model. We then
explore whether these results hold in the generalized linear model setting for
count data using simulations. We show that the use of these methods have
counterintuitive consequences which defy the general expectations in the
literature. In particular, our results and the accompanying simulations suggest
that RSR methods will typically perform worse than non-spatial methods. These
results have important implications for dimension reduction strategies in
spatial regression modeling. Specifically, we demonstrate that the problems
with RSR models cannot be fixed with a selection of "better" spatial basis
vectors or dimension reduction techniques.
| stat |
A posteriori Trading-inspired Model-free Time Series Segmentation | Within the context of multivariate time series segmentation this paper
proposes a method inspired by a posteriori optimal trading. After a
normalization step time series are treated channel-wise as surrogate stock
prices that can be traded optimally a posteriori in a virtual portfolio holding
either stock or cash. Linear transaction costs are interpreted as
hyperparameters for noise filtering. Resulting trading signals as well as
resulting trading signals obtained on the reversed time series are used for
unsupervised labeling, before a consensus over channels is reached that
determines segmentation time instants. The method is model-free such that no
model prescriptions for segments are made. Benefits of proposed approach
include simplicity, computational efficiency and adaptability to a wide range
of different shapes of time series. Performance is demonstrated on synthetic
and real-world data, including a large-scale dataset comprising a multivariate
time series of dimension 1000 and length 2709. Proposed method is compared to a
popular model-based bottom-up approach fitting piecewise affine models and to a
recent model-based top-down approach fitting Gaussian models, and found to be
consistently faster while producing more intuitive results.
| stat |
Variational Bayes method for ODE parameter estimation with application
to time-varying SIR model for COVID-19 epidemic | Ordinary differential equation (ODE) is a mathematical model used in many
application areas such as climatology, bioinformatics, and chemical engineering
with its intuitive appeal to modeling. Despite ODE's wide usage in modeling,
frequent absence of their analytic solutions makes it difficult to estimate ODE
parameters from the data, especially when the model has lots of variables and
parameters. This paper proposes a Bayesian ODE parameter estimating algorithm
which is fast and accurate even for models with many parameters. The proposed
method approximates an ODE model with a state-space model based on equations of
a numeric solver. It allows fast estimation by avoiding computations of a whole
numerical solution in the likelihood. The posterior is obtained by a
variational Bayes method, more specifically, the approximate Riemannian
conjugate gradient method (Honkela et al. 2010), which avoids samplings based
on Markov chain Monte Carlo (MCMC). In simulation studies, we compared the
speed and performance of the proposed method with existing methods. The
proposed method showed the best performance in the reproduction of the true ODE
curve with strong stability as well as the fastest computation, especially in a
large model with more than 30 parameters. As a real-world data application, a
SIR model with time-varying parameters was fitted to the COVID-19 data. Taking
advantage of our proposed algorithm, 30 parameters were adequately fitted for
each country.
| stat |
Firth's logistic regression with rare events: accurate effect estimates
AND predictions? | Firth-type logistic regression has become a standard approach for the
analysis of binary outcomes with small samples. Whereas it reduces the bias in
maximum likelihood estimates of coefficients, bias towards 1/2 is introduced in
the predicted probabilities. The stronger the imbalance of the outcome, the
more severe is the bias in the predicted probabilities. We propose two simple
modifications of Firth-type logistic regression resulting in unbiased predicted
probabilities. The first corrects the predicted probabilities by a post-hoc
adjustment of the intercept. The other is based on an alternative formulation
of Firth-types estimation as an iterative data augmentation procedure. Our
suggested modification consists in introducing an indicator variable which
distinguishes between original and pseudo observations in the augmented data.
In a comprehensive simulation study these approaches are compared to other
attempts to improve predictions based on Firth-type penalization and to other
published penalization strategies intended for routine use. For instance, we
consider a recently suggested compromise between maximum likelihood and
Firth-type logistic regression. Simulation results are scrutinized both with
regard to prediction and regression coefficients. Finally, the methods
considered are illustrated and compared for a study on arterial closure devices
in minimally invasive cardiac surgery.
| stat |
Assume, Augment and Learn: Unsupervised Few-Shot Meta-Learning via
Random Labels and Data Augmentation | The field of few-shot learning has been laboriously explored in the
supervised setting, where per-class labels are available. On the other hand,
the unsupervised few-shot learning setting, where no labels of any kind are
required, has seen little investigation. We propose a method, named Assume,
Augment and Learn or AAL, for generating few-shot tasks using unlabeled data.
We randomly label a random subset of images from an unlabeled dataset to
generate a support set. Then by applying data augmentation on the support set's
images, and reusing the support set's labels, we obtain a target set. The
resulting few-shot tasks can be used to train any standard meta-learning
framework. Once trained, such a model, can be directly applied on small
real-labeled datasets without any changes or fine-tuning required. In our
experiments, the learned models achieve good generalization performance in a
variety of established few-shot learning tasks on Omniglot and Mini-Imagenet.
| stat |
The convergence of the Stochastic Gradient Descent (SGD) : a
self-contained proof | We give here a proof of the convergence of the Stochastic Gradient Descent
(SGD) in a self-contained manner.
| stat |
A semiparametric spatiotemporal Bayesian model for the bulk and extremes
of the Fosberg Fire Weather Index | Large wildfires pose a major environmental concern, and precise maps of fire
risk can improve disaster relief planning. Fosberg Fire Weather Index (FFWI) is
often used to measure wildfire risk; FFWI exhibits non-Gaussian marginal
distributions as well as strong spatiotemporal extremal dependence and thus,
modeling FFWI using geostatistical models like Gaussian processes is
questionable. Extreme value theory (EVT)-driven models like max-stable
processes are theoretically appealing but are computationally demanding and
applicable only for threshold exceedances or block maxima. Disaster management
policies often consider moderate-to-extreme quantiles of climate parameters and
hence, joint modeling of the bulk and the tail of the data is required. In this
paper, we consider a Dirichlet process mixture of spatial skew-t processes that
can flexibly model the bulk as well as the tail. The proposed model has
nonstationary mean and covariance structure, and also nonzero spatiotemporal
extremal dependence. A simulation study demonstrates that the proposed model
has better spatial prediction performance compared to some competing models. We
develop spatial maps of FFWI medians and extremes, and discuss the wildfire
risk throughout the Santa Ana region of California.
| stat |
Normalizing Flows: An Introduction and Review of Current Methods | Normalizing Flows are generative models which produce tractable distributions
where both sampling and density evaluation can be efficient and exact. The goal
of this survey article is to give a coherent and comprehensive review of the
literature around the construction and use of Normalizing Flows for
distribution learning. We aim to provide context and explanation of the models,
review current state-of-the-art literature, and identify open questions and
promising future directions.
| stat |
Uncovering the structure of clinical EEG signals with self-supervised
learning | Objective. Supervised learning paradigms are often limited by the amount of
labeled data that is available. This phenomenon is particularly problematic in
clinically-relevant data, such as electroencephalography (EEG), where labeling
can be costly in terms of specialized expertise and human processing time.
Consequently, deep learning architectures designed to learn on EEG data have
yielded relatively shallow models and performances at best similar to those of
traditional feature-based approaches. However, in most situations, unlabeled
data is available in abundance. By extracting information from this unlabeled
data, it might be possible to reach competitive performance with deep neural
networks despite limited access to labels. Approach. We investigated
self-supervised learning (SSL), a promising technique for discovering structure
in unlabeled data, to learn representations of EEG signals. Specifically, we
explored two tasks based on temporal context prediction as well as contrastive
predictive coding on two clinically-relevant problems: EEG-based sleep staging
and pathology detection. We conducted experiments on two large public datasets
with thousands of recordings and performed baseline comparisons with purely
supervised and hand-engineered approaches. Main results. Linear classifiers
trained on SSL-learned features consistently outperformed purely supervised
deep neural networks in low-labeled data regimes while reaching competitive
performance when all labels were available. Additionally, the embeddings
learned with each method revealed clear latent structures related to
physiological and clinical phenomena, such as age effects. Significance. We
demonstrate the benefit of self-supervised learning approaches on EEG data. Our
results suggest that SSL may pave the way to a wider use of deep learning
models on EEG data.
| stat |
Improving VAEs' Robustness to Adversarial Attack | Variational autoencoders (VAEs) have recently been shown to be vulnerable to
adversarial attacks, wherein they are fooled into reconstructing a chosen
target image. However, how to defend against such attacks remains an open
problem. We make significant advances in addressing this issue by introducing
methods for producing adversarially robust VAEs. Namely, we first demonstrate
that methods proposed to obtain disentangled latent representations produce
VAEs that are more robust to these attacks. However, this robustness comes at
the cost of reducing the quality of the reconstructions. We ameliorate this by
applying disentangling methods to hierarchical VAEs. The resulting models
produce high-fidelity autoencoders that are also adversarially robust. We
confirm their capabilities on several different datasets and with current
state-of-the-art VAE adversarial attacks, and also show that they increase the
robustness of downstream tasks to attack.
| stat |
A Variational Inference Framework for Inverse Problems | We present a framework for fitting inverse problem models via variational
Bayes approximations. This methodology guarantees flexibility to statistical
model specification for a broad range of applications, good accuracy
performances and reduced model fitting times, when compared with standard
Markov chain Monte Carlo methods. The message passing and factor graph fragment
approach to variational Bayes we describe facilitates streamlined
implementation of approximate inference algorithms and forms the basis to
software development. Such approach allows for supple inclusion of numerous
response distributions and penalizations into the inverse problem model. Albeit
our analysis is circumscribed to one- and two-dimensional response variables,
we lay down an infrastructure where streamlining algorithmic steps based on
nullifying weak interactions between variables are extendible to inverse
problems in higher dimensions. Image processing applications motivated by
biomedical and archaeological problems are included as illustrations.
| stat |
Multivariate goodness-of-Fit tests based on Wasserstein distance | Goodness-of-fit tests based on the empirical Wasserstein distance are
proposed for simple and composite null hypotheses involving general
multivariate distributions. For group families, the procedure is to be
implemented after preliminary reduction of the data via invariance.This
property allows for calculation of exact critical values and p-values at finite
sample sizes. Applications include testing for location--scale families and
testing for families arising from affine transformations, such as elliptical
distributions with given standard radial density and unspecified location
vector and scatter matrix. A novel test for multivariate normality with
unspecified mean vector and covariance matrix arises as a special case. For
more general parametric families, we propose a parametric bootstrap procedure
to calculate critical values. The lack of asymptotic distribution theory for
the empirical Wasserstein distance means that the validity of the parametric
bootstrap under the null hypothesis remains a conjecture. Nevertheless, we show
that the test is consistent against fixed alternatives. To this end, we prove a
uniform law of large numbers for the empirical distribution in Wasserstein
distance, where the uniformity is over any class of underlying distributions
satisfying a uniform integrability condition but no additional moment
assumptions. The calculation of test statistics boils down to solving the
well-studied semi-discrete optimal transport problem. Extensive numerical
experiments demonstrate the practical feasibility and the excellent performance
of the proposed tests for the Wasserstein distance of order p = 1 and p = 2 and
for dimensions at least up to d = 5. The simulations also lend support to the
conjecture of the asymptotic validity of the parametric bootstrap.
| stat |
Flexible co-data learning for high-dimensional prediction | Clinical research often focuses on complex traits in which many variables
play a role in mechanisms driving, or curing, diseases. Clinical prediction is
hard when data is high-dimensional, but additional information, like domain
knowledge and previously published studies, may be helpful to improve
predictions. Such complementary data, or co-data, provide information on the
covariates, such as genomic location or p-values from external studies. Our
method enables exploiting multiple and various co-data sources to improve
predictions. We use discrete or continuous co-data to define possibly
overlapping or hierarchically structured groups of covariates. These are then
used to estimate adaptive multi-group ridge penalties for generalised linear
and Cox models. We combine empirical Bayes estimation of group penalty
hyperparameters with an extra level of shrinkage. This renders a uniquely
flexible framework as any type of shrinkage can be used on the group level. The
hyperparameter shrinkage learns how relevant a specific co-data source is,
counters overfitting of hyperparameters for many groups, and accounts for
structured co-data. We describe various types of co-data and propose suitable
forms of hypershrinkage. The method is very versatile, as it allows for
integration and weighting of multiple co-data sets, inclusion of unpenalised
covariates and posterior variable selection. We demonstrate it on two cancer
genomics applications and show that it may improve the performance of other
dense and parsimonious prognostic models substantially, and stabilises variable
selection.
| stat |
Density-based Clustering with Best-scored Random Forest | Single-level density-based approach has long been widely acknowledged to be a
conceptually and mathematically convincing clustering method. In this paper, we
propose an algorithm called "best-scored clustering forest" that can obtain the
optimal level and determine corresponding clusters. The terminology
"best-scored" means to select one random tree with the best empirical
performance out of a certain number of purely random tree candidates. From the
theoretical perspective, we first show that consistency of our proposed
algorithm can be guaranteed. Moreover, under certain mild restrictions on the
underlying density functions and target clusters, even fast convergence rates
can be achieved. Last but not least, comparisons with other state-of-the-art
clustering methods in the numerical experiments demonstrate accuracy of our
algorithm on both synthetic data and several benchmark real data sets.
| stat |
Probabilistic Forecasting of Temporal Trajectories of Regional Power
Production - Part 1: Wind | Renewable energy sources provide a constantly increasing contribution to the
total energy production worldwide. However, the power generation from these
sources is highly variable due to their dependence on meteorological
conditions. Accurate forecasts for the production at various temporal and
spatial scales are thus needed for an efficiently operating electricity market.
In this article - part 1 - we propose fully probabilistic prediction models for
spatially aggregated wind power production at an hourly time scale with lead
times up to several days using weather forecasts from numerical weather
prediction systems as covariates. After an appropriate cubic transformation of
the power production, we build up a multivariate Gaussian prediction model
under a Bayesian inference framework which incorporates the temporal error
correlation. In an application to predict wind production in Germany, the
method provides calibrated and skillful forecasts. Comparison is made between
several formulations of the correlation structure.
| stat |
Approximate Cross-Validation with Low-Rank Data in High Dimensions | Many recent advances in machine learning are driven by a challenging
trifecta: large data size $N$; high dimensions; and expensive algorithms. In
this setting, cross-validation (CV) serves as an important tool for model
assessment. Recent advances in approximate cross validation (ACV) provide
accurate approximations to CV with only a single model fit, avoiding
traditional CV's requirement for repeated runs of expensive algorithms.
Unfortunately, these ACV methods can lose both speed and accuracy in high
dimensions -- unless sparsity structure is present in the data. Fortunately,
there is an alternative type of simplifying structure that is present in most
data: approximate low rank (ALR). Guided by this observation, we develop a new
algorithm for ACV that is fast and accurate in the presence of ALR data. Our
first key insight is that the Hessian matrix -- whose inverse forms the
computational bottleneck of existing ACV methods -- is ALR. We show that,
despite our use of the \emph{inverse} Hessian, a low-rank approximation using
the largest (rather than the smallest) matrix eigenvalues enables fast,
reliable ACV. Our second key insight is that, in the presence of ALR data,
error in existing ACV methods roughly grows with the (approximate, low) rank
rather than with the (full, high) dimension. These insights allow us to prove
theoretical guarantees on the quality of our proposed algorithm -- along with
fast-to-compute upper bounds on its error. We demonstrate the speed and
accuracy of our method, as well as the usefulness of our bounds, on a range of
real and simulated data sets.
| stat |
Flexible nonstationary spatio-temporal modeling of high-frequency
monitoring data | Many physical datasets are generated by collections of instruments that make
measurements at regular time intervals. For such regular monitoring data, we
extend the framework of half-spectral covariance functions to the case of
nonstationarity in space and time and demonstrate that this method provides a
natural and tractable way to incorporate complex behaviors into a covariance
model. Further, we use this method with fully time-domain computations to
obtain bona fide maximum likelihood estimators---as opposed to using
Whittle-type likelihood approximations, for example---that can still be
computed efficiently. We apply this method to very high-frequency Doppler LIDAR
vertical wind velocity measurements, demonstrating that the model can
expressively capture the extreme nonstationarity of dynamics above and below
the atmospheric boundary layer and, more importantly, the interaction of the
process dynamics across it.
| stat |
Unsupervised Space-Time Clustering using Persistent Homology | This paper presents a new clustering algorithm for space-time data based on
the concepts of topological data analysis and in particular, persistent
homology. Employing persistent homology - a flexible mathematical tool from
algebraic topology used to extract topological information from data - in
unsupervised learning is an uncommon and a novel approach. A notable aspect of
this methodology consists in analyzing data at multiple resolutions which
allows to distinguish true features from noise based on the extent of their
persistence. We evaluate the performance of our algorithm on synthetic data and
compare it to other well-known clustering algorithms such as K-means,
hierarchical clustering and DBSCAN. We illustrate its application in the
context of a case study of water quality in the Chesapeake Bay.
| stat |
Classification Logit Two-sample Testing by Neural Networks | The recent success of generative adversarial networks and variational
learning suggests training a classifier network may work well in addressing the
classical two-sample problem. Network-based tests have the computational
advantage that the algorithm scales to large samples. This paper proposes a
two-sample statistic which is the difference of the logit function, provided by
a trained classification neural network, evaluated on the testing set split of
the two datasets. Theoretically, we prove the testing power to differentiate
two sub-exponential densities given that the network is sufficiently
parametrized. When the two densities lie on or near to low-dimensional
manifolds embedded in possibly high-dimensional space, the needed network
complexity is reduced to only scale with the intrinsic dimensionality. Both the
approximation and estimation error analysis are based on a new result of
near-manifold integral approximation. In experiments, the proposed method
demonstrates better performance than previous network-based tests using
classification accuracy as the two-sample statistic, and compares favorably to
certain kernel maximum mean discrepancy tests on synthetic datasets and
hand-written digit datasets.
| stat |
Computing Bayes: Bayesian Computation from 1763 to the 21st Century | The Bayesian statistical paradigm uses the language of probability to express
uncertainty about the phenomena that generate observed data. Probability
distributions thus characterize Bayesian analysis, with the rules of
probability used to transform prior probability distributions for all unknowns
- parameters, latent variables, models - into posterior distributions,
subsequent to the observation of data. Conducting Bayesian analysis requires
the evaluation of integrals in which these probability distributions appear.
Bayesian computation is all about evaluating such integrals in the typical case
where no analytical solution exists. This paper takes the reader on a
chronological tour of Bayesian computation over the past two and a half
centuries. Beginning with the one-dimensional integral first confronted by
Bayes in 1763, through to recent problems in which the unknowns number in the
millions, we place all computational problems into a common framework, and
describe all computational methods using a common notation. The aim is to help
new researchers in particular - and more generally those interested in adopting
a Bayesian approach to empirical work - make sense of the plethora of
computational techniques that are now on offer; understand when and why
different methods are useful; and see the links that do exist, between them
all.
| stat |
Foundations of Structural Causal Models with Cycles and Latent Variables | Structural causal models (SCMs), also known as (nonparametric) structural
equation models (SEMs), are widely used for causal modeling purposes. In
particular, acyclic SCMs, also known as recursive SEMs, form a well-studied
subclass of SCMs that generalize causal Bayesian networks to allow for latent
confounders. In this paper, we investigate SCMs in a more general setting,
allowing for the presence of both latent confounders and cycles. We show that
in the presence of cycles, many of the convenient properties of acyclic SCMs do
not hold in general: they do not always have a solution; they do not always
induce unique observational, interventional and counterfactual distributions; a
marginalization does not always exist, and if it exists the marginal model does
not always respect the latent projection; they do not always satisfy a Markov
property; and their graphs are not always consistent with their causal
semantics. We prove that for SCMs in general each of these properties does hold
under certain solvability conditions. Our work generalizes results for SCMs
with cycles that were only known for certain special cases so far. We introduce
the class of simple SCMs that extends the class of acyclic SCMs to the cyclic
setting, while preserving many of the convenient properties of acyclic SCMs.
With this paper we aim to provide the foundations for a general theory of
statistical causal modeling with SCMs.
| stat |
Bayesian epidemiological modeling over high-resolution network data | Mathematical epidemiological models have a broad use, including both
qualitative and quantitative applications. With the increasing availability of
data, large-scale quantitative disease spread models can nowadays be
formulated. Such models have a great potential, e.g., in risk assessments in
public health. Their main challenge is model parameterization given
surveillance data, a problem which often limits their practical usage.
We offer a solution to this problem by developing a Bayesian methodology
suitable to epidemiological models driven by network data. The greatest
difficulty in obtaining a concentrated parameter posterior is the quality of
surveillance data; disease measurements are often scarce and carry little
information about the parameters. The often overlooked problem of the model's
identifiability therefore needs to be addressed, and we do so using a hierarchy
of increasingly realistic known truth experiments.
Our proposed Bayesian approach performs convincingly across all our synthetic
tests. From pathogen measurements of shiga toxin-producing Escherichia coli
O157 in Swedish cattle, we are able to produce an accurate statistical model of
first-principles confronted with data. Within this model we explore the
potential of a Bayesian public health framework by assessing the efficiency of
disease detection and -intervention scenarios.
| stat |
FFORMPP: Feature-based forecast model performance prediction | This paper introduces a novel meta-learning algorithm for time series
forecast model performance prediction. We model the forecast error as a
function of time series features calculated from the historical time series
with an efficient Bayesian multivariate surface regression approach. The
minimum predicted forecast error is then used to identify an individual model
or a combination of models to produce the final forecasts. It is well-known
that the performance of most meta-learning models depends on the
representativeness of the reference dataset used for training. In such
circumstances, we augment the reference dataset with a feature-based time
series simulation approach, namely GRATIS, in generating a rich and
representative time series collection. The proposed framework is tested using
the M4 competition data and is compared against commonly used forecasting
approaches. Our approach provides provides comparable performances to other
model selection/combination approaches but at lower computational cost and
higher degree of interpretability, which is important for supporting decisions.
We also provide useful insights regarding which forecasting models are expected
to work better for particular types of time series, how the meta-learners work
and how the forecasting performances are affected by various factors.
| stat |
A Bayesian Finite Mixture Model with Variable Selection for Data with
Mixed-type Variables | Finite mixture model is an important branch of clustering methods and can be
applied on data sets with mixed types of variables. However, challenges exist
in its applications. First, it typically relies on the EM algorithm which could
be sensitive to the choice of initial values. Second, biomarkers subject to
limits of detection (LOD) are common to encounter in clinical data, which
brings censored variables into finite mixture model. Additionally, researchers
are recently getting more interest in variable importance due to the increasing
number of variables that become available for clustering.
To address these challenges, we propose a Bayesian finite mixture model to
simultaneously conduct variable selection, account for biomarker LOD and obtain
clustering results. We took a Bayesian approach to obtain parameter estimates
and the cluster membership to bypass the limitation of the EM algorithm. To
account for LOD, we added one more step in Gibbs sampling to iteratively fill
in biomarker values below or above LODs. In addition, we put a spike-and-slab
type of prior on each variable to obtain variable importance. Simulations
across various scenarios were conducted to examine the performance of this
method. Real data application on electronic health records was also conducted.
| stat |
Cost-Effective Incentive Allocation via Structured Counterfactual
Inference | We address a practical problem ubiquitous in modern marketing campaigns, in
which a central agent tries to learn a policy for allocating strategic
financial incentives to customers and observes only bandit feedback. In
contrast to traditional policy optimization frameworks, we take into account
the additional reward structure and budget constraints common in this setting,
and develop a new two-step method for solving this constrained counterfactual
policy optimization problem. Our method first casts the reward estimation
problem as a domain adaptation problem with supplementary structure, and then
subsequently uses the estimators for optimizing the policy with constraints. We
also establish theoretical error bounds for our estimation procedure and we
empirically show that the approach leads to significant improvement on both
synthetic and real datasets.
| stat |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.