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:153@isdm.umontpellier.fr
DTSTART;TZID=Europe/Paris:20250516T110000
DTEND;TZID=Europe/Paris:20250516T120000
DTSTAMP:20260602T130233Z
URL:https://isdm.umontpellier.fr/events/unified-breakdown-analysis-for-byz
 antine-robust-gossip-2/
SUMMARY:Unified Breakdown Analysis for Byzantine Robust Gossip
DESCRIPTION:Inria Montpellier\, St-Priest Campus\, Building 5\, Room 02/022
 \nMachine Learning in Montpellier\, Theory &amp\; Practice\n\nIn decentral
 ized machine learning\, different devices communicate in a peer-to-peer ma
 nner to collaboratively learn from each other's data. Such approaches are 
 vulnerable to misbehaving (or Byzantine) devices. We introduce F-RG\, a ge
 neral framework for building robust decentralized algorithms with guarante
 es arising from robust-sum-like aggregation rules F. We then investigate t
 he notion of breakdown point\, and show an upper bound on the number of ad
 versaries that decentralized algorithms can tolerate. We introduce a pract
 ical robust aggregation rule\, coined CSours\, such that CSours-RG has a n
 ear-optimal breakdown. Other choices of aggregation rules lead to existing
  algorithms such as ClippedGossip or NNA. We give experimental evidence to
  validate the effectiveness of CSours-RG and highlight the gap with NNA\, 
 in particular against a novel attack tailored to decentralized communicati
 ons.\n\nMachineLearning\, LabéliséHallesIA\, IA&amp\;Expert
ATTACH;FMTTYPE=image/jpeg:https://isdm.umontpellier.fr/wp-content/uploads/
 2025/06/ml-mpt-ZV3BFD.png
END:VEVENT
BEGIN:VTIMEZONE
TZID:Europe/Paris
X-LIC-LOCATION:Europe/Paris
BEGIN:DAYLIGHT
DTSTART:20250330T030000
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
END:DAYLIGHT
END:VTIMEZONE
END:VCALENDAR