[ABE-L] Fwd: ISBA-BNP October 2025 webinar: Guillaume Kon Kam King

[Hedibert Lopes]

---------- Forwarded message ---------
From: ISBA-BNP Program Chair Juhee Lee <juheelee em soe.ucsc.edu>
Date: Tue, Oct 14, 2025 at 1:12 PM
Subject: ISBA-BNP October 2025 webinar: Guillaume Kon Kam King
To: Prof. Hedibert Lopes <hedibert em gmail.com>


Dear ISBA members,

We are pleased to announce the launch of the BNP-ISBA Webinar Series for
the new academic year, organized by the Bayesian Nonparametric Section of
ISBA. Last year, we introduced a new format for the series, and we will
continue with this format this year. Webinars will be held approximately
every two months. Details about upcoming webinars, including the Zoom link,
are available at: https://bnp-isba.github.io/webinars.html. We warmly
invite you to join us for the first webinar of the series and for the
events to follow throughout the year.

DATE & TIME: 16:00 UTC on October 29th, 2025. Note that 16:00 UTC
corresponds to 12:00 US Eastern and *17:00* Central European (ended
daylight saving time on Oct 26, 2025).

SPEAKER: Guillaume Kon Kam King (Senior Research Fellow, INRAE, France)

TITLE: Bayesian nonparametric mixture models and the posterior number of
clusters.

ABSTRACT: Bayesian nonparametric (BNP) mixture models, such as Dirichlet
and Pitman–Yor processes, are powerful tools for clustering because they
allow an unbounded number of components. This flexibility is appealing, but
it comes with an implicit prior belief: as the sample size grows, the
number of clusters tends to infinity. While this seems to be a reasonable
prior belief in most applications, there has been recent interest in
examining the misspecified setting where data arise from a finite mixture.
The central question is whether, in this setting, the data can override the
prior and force the posterior distribution on the number of clusters to
concentrate on a finite value.
Miller and Harrison (2014) showed that for Dirichlet and Pitman–Yor
mixtures, the answer is negative—the posterior is inconsistent for the
number of clusters. In this talk, we discuss these findings, generalise
them and examine several solutions. We consider a broad range of BNP
priors, including Gibbs-type processes and overfitted finite mixtures
(Rousseau & Mengersen, 2011). Along the way, we revisit related results on
the consistency of the mixing measure and discuss practical strategies
proposed in the literature, such as the merge–truncate–merge algorithm of
Guha et al. (2019) and the use of hyperpriors on BNP parameters as in
Ascolani et al. (2023).
The talk will provide intuition for why these inconsistency phenomena
occur, explore their theoretical underpinnings, and highlight implications
for clustering practice. The contents of this presentation are based on
Lawless et al. (2023) and Alamichel et al. (2024).

References

   - Alamichel, L., Bystrova, D., Arbel, J., & Kon Kam King, G. (2024).
   Bayesian mixture models (in)consistency for the number of clusters.
   Scandinavian Journal of Statistics, 51(4), 1619–1660.
   https://doi.org/10.1111/sjos.12739
   - Lawless, C., Arbel, J., Alamichel, L., & Kon Kam King, G. (2023).
   Clustering inconsistency for Pitman–Yor mixture models with a prior on the
   precision but fixed discount parameter. In Fifth Symposium on Advances in
   Approximate Bayesian Inference. https://hal.science/hal-04425711
   - Miller, J. W., & Harrison, M. T. (2014). Inconsistency of Pitman–Yor
   process mixtures for the number of components. Journal of Machine Learning
   Research, 15, 3333–3370. https://jmlr.org/papers/v15/miller14a.html
   - Guha, A., Nguyen, X. L., & Ho, N. (2019). Parameter estimation and
   interpretability in Bayesian mixture models. arXiv:1901.05078.
   https://arxiv.org/pdf/1901.05078
   - Ascolani, F., Franzolini, B., Lijoi, A., & Prünster, I. (2023).
   Nonparametric priors with full-range borrowing of information.
   https://beatricefranzolini.github.io/AscolaniFranzoliniLijoiPruenster2023.pdf
   - Rousseau, J., & Mengersen, K. (2011). Asymptotic behaviour of the
   posterior distribution in overfitted mixture models. Journal of the Royal
   Statistical Society: Series B, 73(5), 689–710.
   https://doi.org/10.1111/j.1467-9868.2011.00781.x

Best regards,

--
Juhee Lee
ISBA - BNP Section
Program Chair 2024-2025
e-mail: jle297 em ucsc.edu
------------------------------
Sent from the International Society for Bayesian Analysis
members-info em bayesian.org
United States
To customize your greeting or update your name, email or mailing
subscriptions login at http://bayesian.org and go to My Account. ISBA
protects the privacy of its members and does not distribute its mailing
list to third parties. To opt out of future mailings go to
https://bayesian.org/civicrm/?civiwp=CiviCRM&q=civicrm%2Fmailing%2Foptout&reset=1&jid=6314&qid=1338059&h=kns13wtnycpoaqrg


-- 
Hedibert Freitas Lopes, PhD
Professor of Statistics and Econometrics
INSPER - Institute of Education and Research
Rua Quatá, 300 - São Paulo, SP 04546-042 Brazil
Phone: +55 11 4504-2343
www.hedibert.org
-------------- Próxima Parte ----------
Um anexo em HTML foi limpo...
URL: <http://lists.ime.usp.br/pipermail/abe/attachments/20251014/a8674f9b/attachment.htm>