BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//wp-events-plugin.com//7.4.0.1//EN
TZID:Europe/Paris
X-WR-TIMEZONE:Europe/Paris
BEGIN:VEVENT
UID:180@isdm.umontpellier.fr
DTSTART;TZID=Europe/Paris:20260611T140000
DTEND;TZID=Europe/Paris:20260611T140000
DTSTAMP:20260825T124528Z
URL:https://isdm.umontpellier.fr/events/ensembles-in-machine-learning-simp
 le-theory-and-simple-practice/
SUMMARY:Ensembles in machine learning: (simple) theory and (simple) practic
 e
DESCRIPTION:Room 03.124\, Building 5\, St Priest campus\n\nMachine Learning
  in Montpellier\, Theory &amp\; Practice - Pierre-Alexandre Mattei (Inria 
 (Maasai))\n\nEnsemble methods combine predictions from various statistical
  learning models. Their most famous representatives are random forests or 
 deep ensembles. This talk will center around the question: "How many model
 s should I aggregate?" We will see that the answer depends on the chosen p
 erformance metric. Specifically\, in the case of convex losses (such as cr
 oss-entropy in classification or mean squared error in regression)\, the e
 rror is a decreasing function of the number of models. In the case of non-
 convex losses (such as classification error in classification or the Fréc
 het Inception distance in generative modelling)\, things are more nuanced\
 , and the error can sometimes be non-monotonic. These results will be illu
 strated with examples of neural network ensembles\, both for classificatio
 n and generative modelling. This work is notably based on the papers: - Ar
 e Ensembles Getting Better All the Time? (with Damien Garreau)\, JMLR 2025
  - When Are Two Scores Better Than One? Investigating Ensembles of Diffusi
 on Models (with Raphaël Razafindralambo\, Rémy Sun\, Frédéric Precioso
 \, and Damien Garreau)\, TMLR 2026 - Beyond Mixtures and Products for Ense
 mble Aggregation: A Likelihood Perspective on Generalized Means (with Raph
 aël Razafindralambo\, Rémy Sun\, Frédéric Precioso\, and Damien Garrea
 u)\, arXiv 2026\n\nMachine Learning in Montpellier\, Theory &amp\; Practic
 e
ATTACH;FMTTYPE=image/jpeg:https://isdm.umontpellier.fr/wp-content/uploads/
 2026/06/ml-mtp-gC78d5.png
CATEGORIES:ML MTP
END:VEVENT
BEGIN:VTIMEZONE
TZID:Europe/Paris
X-LIC-LOCATION:Europe/Paris
BEGIN:DAYLIGHT
DTSTART:20260329T030000
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
END:DAYLIGHT
END:VTIMEZONE
END:VCALENDAR