Réunion

Thematic day on phase transitions in high-dimensional inference

Date : 28 Octobre 2025
Horaire : 08h30 - 18h00
Lieu : Salle de Conférence IV Laboratoire de physique de l'Ecole Normale Supérieure 24 rue Lhomond 75005 Paris France

Axes scientifiques :

Théorie et méthodes
Apprentissage machine

Organisateurs :

- Bruno Loureiro (DI ENS)

Nous vous rappelons que, afin de garantir l'accès de tous les inscrits aux salles de réunion, l'inscription aux réunions est gratuite mais obligatoire.

Inscriptions

1 personnes membres du GdR IASIS, et 5 personnes non membres du GdR, sont inscrits à cette réunion.

Capacité de la salle : 70 personnes. 64 Places restantes

Annonce

Phase transitions have emerged as a unifying theme across probability, statistics, statistical physics, and machine learning. In many high-dimensional inference problems—ranging from community detection in networks and sparse signal recovery to learning with neural networks—one observes sharp thresholds that separate regimes of possibility, impossibility, and computational hardness. Understanding these transitions has become central to both theory and practice: they reveal fundamental limits of what can be inferred from data, clarify the gap between information-theoretic and algorithmic feasibility, and provide new tools to analyse the behavior of algorithms.

This thematic day will bring together researchers from diverse fields to discuss recent progress, open challenges, and emerging connections around phase transitions in high-dimensional inference.

Invited speakers:

Contributed talks

We will have two contributed talks from participants of the thematic day. If you wish to apply for a contributed talk, please register for the thematic day and fill the following form:

https://forms.gle/3vYAi7PfgHuiPTUW6

Note: filling the form does not imply automatic registration.

Organisers: Bruno Loureiro (DI-ENS & CNRS) and Marylou Gabrié (LPENS)