BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//wp-events-plugin.com//6.4.7.2//EN
TZID:Europe/Paris
X-WR-TIMEZONE:Europe/Paris
BEGIN:VEVENT
UID:0-1079@lptms.universite-paris-saclay.fr
DTSTART;TZID=Europe/Paris:20251212T160000
DTEND;TZID=Europe/Paris:20251212T170000
DTSTAMP:20251204T093309Z
URL:http://www.lptms.universite-paris-saclay.fr/seminars/mlpp-seminars-gia
 nluca-manzan-paris-saclay-university/
SUMMARY:MLP@P seminars : Gianluca Manzan  (LISN) - LISN\, bat 660 salle 201
 4 (2° étage) - 12 Déc 25 16:00
DESCRIPTION:From Hopfield Inference to Federated Learning: challenges and s
 olutions in Teacher–Student models\nGianluca Manzan\nLISN\, Paris Saclay
  University\n\nSeminar of the Series MLP@P (Machine Learning Physics @ Pla
 teau)\, joint with LISN and IPhT.\nWhere: LISN\, bat 660 salle 2014 (2° 
 étage)\n\nIn this work\, we investigate inference in neural networks thro
 ugh the teacher–student framework\, which provides a controlled setting 
 to quantify how a student model learns the underlying signal from data gen
 erated by a teacher. Beginning with the Hopfield model\, interpreted as a 
 dual formulation of associative memory\, we characterize the transition be
 tween non-informative and learning phases as a function of dataset size\, 
 noise level\, and temperature. Extending the analysis to Restricted Boltzm
 ann Machines\, we show how choices in unit priors and regularization shape
  the emergence of the signal-retrieval phase and thus determine learning e
 fficiency.\nWe then address the limitations of the single-student scenario
  by introducing a collective learning strategy in which multiple student n
 etworks are coupled during inference. Recent advances in statistical-physi
 cs of learning show that interactions among students enhance generalizatio
 n\, thereby facilitating the teacher recovery. Our analysis of y-interacti
 ng Hopfield students confirms this cooperative effect\, demonstrating that
  coupling expands the region of successful inference by lowering data requ
 irements.\nThis collective perspective has a natural application to federa
 ted learning (FL)\, a decentralized paradigm where multiple clients collab
 oratively train local models without sharing their private data. In FL\, e
 ach client performs local updates based on its own dataset and communicate
 s only through model parameters. The cooperative mechanism observed in cou
 pled teacher–student systems provides a theoretical analogue to this fra
 mework: just as interacting students benefit from mutual alignment\, feder
 ated clients collectively enrich the global solution while retaining data 
 locality.
CATEGORIES:MLP@P
LOCATION:LISN\, bat 660 salle 2014 (2° étage)\, 660 Av. des Sciences\,  9
 1190 Gif-sur-Yvette\, France\, Gif-sur-Yvette\, 91190 \, France
X-APPLE-STRUCTURED-LOCATION;VALUE=URI;X-ADDRESS=660 Av. des Sciences\,  911
 90 Gif-sur-Yvette\, France\, Gif-sur-Yvette\, 91190 \, France;X-APPLE-RADI
 US=100;X-TITLE=LISN\, bat 660 salle 2014 (2° étage):geo:0,0
END:VEVENT
BEGIN:VTIMEZONE
TZID:Europe/Paris
X-LIC-LOCATION:Europe/Paris
BEGIN:STANDARD
DTSTART:20251026T020000
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
END:STANDARD
END:VTIMEZONE
END:VCALENDAR