Skip to content

Commit 107cc43

Browse files
github-actions[bot]github-actions[bot]
committed
Refresh publication metadata [skip ci]
1 parent 3b8f550 commit 107cc43

1 file changed

Lines changed: 223 additions & 0 deletions

File tree

‎site/published.xml‎

Lines changed: 223 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,223 @@
1+
<?xml version="1.0" encoding="UTF-8"?>
2+
<rss version="2.0">
3+
<channel>
4+
<title>Computo Journal - Recent Articles</title>
5+
<link>https://computo-journal.org/</link>
6+
<description>Latest published articles from Computo Journal</description>
7+
<item>
8+
<title>Should we correct the bias in Confidence Bands for Repeated Functional Data? - Emilie Devijver and Adeline Samson</title>
9+
<link>https://computo-journal.org/published-202606-devijver-scb/</link>
10+
<guid>https://computo-journal.org/published-202606-devijver-scb/</guid>
11+
<pubDate>Mon, 20 Jul 2026 00:00:00 GMT</pubDate>
12+
<description>While confidence intervals for finite quantities are
13+
well-established, constructing confidence bands for objects of
14+
infinite dimension, such as functions, is challenging. In this
15+
paper, we explore the concept of parametric confidence bands for
16+
functional data with an orthonormal basis. Specifically, we revisit
17+
the method proposed by Sun and Loader, which yields confidence bands
18+
for the projection of the regression function in a fixed-dimensional
19+
space. This approach can introduce bias into the confidence bands
20+
when the dimension of the basis is misspecified. Building on this
21+
insight, we introduce a corrected, unbiased confidence band.
22+
Surprisingly, our corrected band tends to be wider than that
23+
suggested by a naive approach. To address this, we propose a model
24+
selection criterion that allows for data-driven estimation of the
25+
basis dimension. The bias is then automatically corrected after
26+
dimension selection. We illustrate these strategies through an
27+
extensive simulation study. We conclude with an application to real
28+
data.</description>
29+
</item>
30+
<item>
31+
<title>Variational inference for approximate objective priors using neural networks - Nils Baillie, Antoine Van Biesbroeck and Clément Gauchy</title>
32+
<link>https://computo-journal.org/published-202512-baillie-varp/</link>
33+
<guid>https://computo-journal.org/published-202512-baillie-varp/</guid>
34+
<pubDate>Mon, 01 Dec 2025 00:00:00 GMT</pubDate>
35+
<description>In Bayesian statistics, the choice of the prior can have
36+
an important influence on the posterior and the parameter
37+
estimation, especially when few data samples are available. To limit
38+
the added subjectivity from a priori information, one can use the
39+
framework of objective priors, more particularly, we focus on
40+
reference priors in this work. However, computing such priors is a
41+
difficult task in general. Hence, we consider cases where the
42+
reference prior simplifies to the Jeffreys prior. We develop in this
43+
paper a flexible algorithm based on variational inference which
44+
computes approximations of priors from a set of parametric
45+
distributions using neural networks. We also show that our algorithm
46+
can retrieve modified Jeffreys priors when constraints are specified
47+
in the optimization problem to ensure the solution is proper. We
48+
propose a simple method to recover a relevant approximation of the
49+
parametric posterior distribution using Markov Chain Monte Carlo
50+
(MCMC) methods even if the density function of the parametric prior
51+
is not known in general. Numerical experiments on several
52+
statistical models of increasing complexity are presented. We show
53+
the usefulness of this approach by recovering the target
54+
distribution. The performance of the algorithm is evaluated on both
55+
prior and posterior distributions, jointly using variational
56+
inference and MCMC sampling.</description>
57+
</item>
58+
<item>
59+
<title>`regMMD`: an `R` package for parametric estimation and regression with maximum mean discrepancy - Pierre Alquier and Mathieu Gerber</title>
60+
<link>https://computo-journal.org/published-202511-alquier-regmmd/</link>
61+
<guid>https://computo-journal.org/published-202511-alquier-regmmd/</guid>
62+
<pubDate>Tue, 18 Nov 2025 00:00:00 GMT</pubDate>
63+
<description>The Maximum Mean Discrepancy (MMD) is a kernel-based
64+
metric widely used for nonparametric tests and estimation. Recently,
65+
it has also been studied as an objective function for parametric
66+
estimation, as it has been shown to yield robust estimators. We have
67+
implemented MMD minimization for parameter inference in a wide range
68+
of statistical models, including various regression models, within
69+
an `R` package called `regMMD`. This paper provides an introduction
70+
to the `regMMD` package. We describe the available kernels and
71+
optimization procedures, as well as the default settings. Detailed
72+
applications to simulated and real data are provided.</description>
73+
</item>
74+
<item>
75+
<title>Fast confidence bounds for the false discovery proportion over a path of hypotheses - Guillermo Durand</title>
76+
<link>https://computo-journal.org/published-202510-durand-fast/</link>
77+
<guid>https://computo-journal.org/published-202510-durand-fast/</guid>
78+
<pubDate>Thu, 09 Oct 2025 00:00:00 GMT</pubDate>
79+
<description>This paper presents a new algorithm (and an additional
80+
trick) that allows to compute fastly an entire curve of post hoc
81+
bounds for the False Discovery Proportion when the underlying bound
82+
\$V\_\{\textbackslash mathfrak\{R\}\}\^{}\{\textbackslash ast\}\$
83+
construction is based on a reference family \$\textbackslash
84+
mathfrak\{R\}\$ with a forest structure à la @MR4178188. By an
85+
entire curve, we mean the values \$V\_\{\textbackslash
86+
mathfrak\{R\}\}\^{}\{\textbackslash ast\}(S\_1),\textbackslash
87+
dotsc,V\_\{\textbackslash mathfrak\{R\}\}\^{}\{\textbackslash
88+
ast\}(S\_m)\$ computed on a path of increasing selection sets
89+
\$S\_1\textbackslash subsetneq\textbackslash dotsb\textbackslash
90+
subsetneq S\_m\$, \$\textbar S\_t\textbar=t\$. The new algorithm
91+
leverages the fact that going from \$S\_t\$ to \$S\_\{t+1\}\$ is
92+
done by adding only one hypothesis. Compared to a more naive
93+
approach, the new algorithm has a complexity in
94+
\$O(\textbar\textbackslash mathcal K\textbar m)\$ instead of
95+
\$O(\textbar\textbackslash mathcal K\textbar m\^{}2)\$, where
96+
\$\textbar\textbackslash mathcal K\textbar\$ is the cardinality of
97+
the family.</description>
98+
</item>
99+
<item>
100+
<title>Draw Me a Simulator - Sandrine Boulet and Antoine Chambaz</title>
101+
<link>https://computo-journal.org/published-202509-boulet-simulator/</link>
102+
<guid>https://computo-journal.org/published-202509-boulet-simulator/</guid>
103+
<pubDate>Mon, 08 Sep 2025 00:00:00 GMT</pubDate>
104+
<description>This study investigates the use of Variational
105+
Auto-Encoders to build a simulator that approximates the law of
106+
genuine observations. Using both simulated and real data in
107+
scenarios involving counterfactuality, we discuss the general task
108+
of evaluating a simulator’s quality, with a focus on comparisons of
109+
statistical properties and predictive performance. While the
110+
simulator built from simulated data shows minor discrepancies, the
111+
results with real data reveal more substantial challenges. Beyond
112+
the technical analysis, we reflect on the broader implications of
113+
simulator design, and consider its role in modeling reality.</description>
114+
</item>
115+
<item>
116+
<title>Model-Based Clustering and Variable Selection for Multivariate Count Data - Julien Jacques and Thomas Brendan Murphy</title>
117+
<link>https://computo-journal.org/published-202507-jacques-count-data/</link>
118+
<guid>https://computo-journal.org/published-202507-jacques-count-data/</guid>
119+
<pubDate>Tue, 01 Jul 2025 00:00:00 GMT</pubDate>
120+
<description>Model-based clustering provides a principled way of
121+
developing clustering methods. We develop a new model-based
122+
clustering methods for count data. The method combines clustering
123+
and variable selection for improved clustering. The method is based
124+
on conditionally independent Poisson mixture models and Poisson
125+
generalized linear models. The method is demonstrated on simulated
126+
data and data from an ultra running race, where the method yields
127+
excellent clustering and variable selection performance.</description>
128+
</item>
129+
<item>
130+
<title>Reservoir Computing in R: a Tutorial for Using reservoirnet to Predict Complex Time-Series - Thomas Ferté, Kalidou Ba, Dan Dutartre, Pierrick Legrand, Vianney Jouhet, Rodolphe Thiébaut, Xavier Hinaut and Boris P Hejblum</title>
131+
<link>http://computo-journal.org/published-202505-ferte-reservoirnet/</link>
132+
<guid>http://computo-journal.org/published-202505-ferte-reservoirnet/</guid>
133+
<pubDate>Fri, 27 Jun 2025 00:00:00 GMT</pubDate>
134+
<description>Reservoir Computing (RC) is a machine learning method
135+
based on neural networks that efficiently process information
136+
generated by dynamical systems. It has been successful in solving
137+
various tasks including time series forecasting, language processing
138+
or voice processing. RC is implemented in `Python` and `Julia` but
139+
not in `R`. This article introduces `reservoirnet`, an `R` package
140+
providing access to the `Python` API `ReservoirPy`, allowing `R`
141+
users to harness the power of reservoir computing. This article
142+
provides an introduction to the fundamentals of RC and showcases its
143+
real-world applicability through three distinct sections. First, we
144+
cover the foundational concepts of RC, setting the stage for
145+
understanding its capabilities. Next, we delve into the practical
146+
usage of `reservoirnet` through two illustrative examples. These
147+
examples demonstrate how it can be applied to real-world problems,
148+
specifically, regression of COVID-19 hospitalizations and
149+
classification of Japanese vowels. Finally, we present a
150+
comprehensive analysis of a real-world application of
151+
`reservoirnet`, where it was used to forecast COVID-19
152+
hospitalizations at Bordeaux University Hospital using public data
153+
and electronic health records.</description>
154+
</item>
155+
<item>
156+
<title>Efficient simulation of individual-based population models - Daphné Giorgi, Sarah Kaakai and Vincent Lemaire</title>
157+
<link>http://computo-journal.org/published-202412-giorgi-efficient/</link>
158+
<guid>http://computo-journal.org/published-202412-giorgi-efficient/</guid>
159+
<pubDate>Mon, 27 Jan 2025 00:00:00 GMT</pubDate>
160+
<description>The `R` Package `IBMPopSim` facilitates the simulation of
161+
the random evolution of heterogeneous populations using stochastic
162+
Individual-Based Models (IBMs). The package enables users to
163+
simulate population evolution, in which individuals are
164+
characterized by their age and some characteristics, and the
165+
population is modified by different types of events, including
166+
births/arrivals, death/exit events, or changes of characteristics.
167+
The frequency at which an event can occur to an individual can
168+
depend on their age and characteristics, but also on the
169+
characteristics of other individuals (interactions). Such models
170+
have a wide range of applications in fields including actuarial
171+
science, biology, ecology or epidemiology. `IBMPopSim` overcomes the
172+
limitations of time-consuming IBMs simulations by implementing new
173+
efficient algorithms based on thinning methods, which are compiled
174+
using the `Rcpp` package while providing a user-friendly interface.</description>
175+
</item>
176+
<item>
177+
<title>Spectral Bridges - Félix Laplante and Christophe Ambroise</title>
178+
<link>http://computo-journal.org/published-202412-ambroise-spectral/</link>
179+
<guid>http://computo-journal.org/published-202412-ambroise-spectral/</guid>
180+
<pubDate>Fri, 13 Dec 2024 00:00:00 GMT</pubDate>
181+
<description>In this paper, Spectral Bridges, a novel clustering
182+
algorithm, is introduced. This algorithm builds upon the traditional
183+
k-means and spectral clustering frameworks by subdividing data into
184+
small Voronoï regions, which are subsequently merged according to a
185+
connectivity measure. Drawing inspiration from Support Vector
186+
Machine’s margin concept, a non-parametric clustering approach is
187+
proposed, building an affinity margin between each pair of Voronoï
188+
regions. This approach delineates intricate, non-convex cluster
189+
structures and is robust to hyperparameter choice. The numerical
190+
experiments underscore Spectral Bridges as a fast, robust, and
191+
versatile tool for clustering tasks spanning diverse domains. Its
192+
efficacy extends to large-scale scenarios encompassing both
193+
real-world and synthetic datasets. The Spectral Bridge algorithm is
194+
implemented both in Python (\textless
195+
https://pypi.org/project/spectral-bridges\textgreater) and R
196+
\textless
197+
https://github.com/cambroise/spectral-bridges-Rpackage\textgreater).</description>
198+
</item>
199+
<item>
200+
<title>AdaptiveConformal: An `R` Package for Adaptive Conformal Inference - Herbert Susmann, Antoine Chambaz and Julie Josse</title>
201+
<link>http://computo-journal.org/published-202407-susmann-adaptive-conformal/</link>
202+
<guid>http://computo-journal.org/published-202407-susmann-adaptive-conformal/</guid>
203+
<pubDate>Thu, 18 Jul 2024 00:00:00 GMT</pubDate>
204+
<description>Conformal Inference (CI) is a popular approach for
205+
generating finite sample prediction intervals based on the output of
206+
any point prediction method when data are exchangeable. Adaptive
207+
Conformal Inference (ACI) algorithms extend CI to the case of
208+
sequentially observed data, such as time series, and exhibit strong
209+
theoretical guarantees without having to assume exchangeability of
210+
the observed data. The common thread that unites algorithms in the
211+
ACI family is that they adaptively adjust the width of the generated
212+
prediction intervals in response to the observed data. We provide a
213+
detailed description of five ACI algorithms and their theoretical
214+
guarantees, and test their performance in simulation studies. We
215+
then present a case study of producing prediction intervals for
216+
influenza incidence in the United States based on black-box point
217+
forecasts. Implementations of all the algorithms are released as an
218+
open-source `R` package, `AdaptiveConformal`, which also includes
219+
tools for visualizing and summarizing conformal prediction
220+
intervals.</description>
221+
</item>
222+
</channel>
223+
</rss>

0 commit comments

Comments
 (0)