BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//wp-events-plugin.com//7.2.3.1//EN
TZID:Europe/Paris
X-WR-TIMEZONE:Europe/Paris
BEGIN:VEVENT
UID:6480@i2m.univ-amu.fr
DTSTART;TZID=Europe/Paris:20210326T143000
DTEND;TZID=Europe/Paris:20210326T153000
DTSTAMP:20241120T201719Z
URL:https://www.i2m.univ-amu.fr/evenements/hierarchical-bandits-for-quanti
 fying-human-perception/
SUMMARY:Julien Audiffren (University of Fribourg): Hierarchical bandits for
  quantifying human perception
DESCRIPTION:Julien Audiffren: In this presentation we discuss a variant of 
 the continuous multi-armed bandits problem\, called the threshold estimati
 on problem\, which is at the heart of many psychometric experiment. Here\,
  the objective is to estimate the sensitivity threshold for an unknown psy
 chometric function Ψ\, which is assumed to be non decreasing and continuo
 us. Our algorithm\, Dichotomous Optimistic Search (DOS)\, efficiently solv
 es this task by taking inspiration from hierarchical multi-armed bandits a
 nd Black-box optimization. Compared to previous approaches\, DOS is model 
 free and only makes minimal assumption on Ψ smoothness\, while having str
 ong theoretical guarantees that compares favorably to recent methods from 
 both Psychophysics and Global Optimization. We also empirically evaluate D
 OS and show that it significantly outperforms these methods\, both in expe
 riments that mimics the conduct of a psychometric experiment\, and in test
 s with large pulls budgets that illustrate the faster convergence rate.\nh
 ttps://hal.archives-ouvertes.fr/hal-02448282v1\n\n
ATTACH;FMTTYPE=image/jpeg:https://www.i2m.univ-amu.fr/wp-content/uploads/2
 020/01/Julien_Audiffren.jpg
CATEGORIES:Séminaire,Signal et Apprentissage,Virtual event
END:VEVENT
BEGIN:VTIMEZONE
TZID:Europe/Paris
X-LIC-LOCATION:Europe/Paris
BEGIN:STANDARD
DTSTART:20201025T020000
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
END:STANDARD
END:VTIMEZONE
END:VCALENDAR