BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:AMLD 2021 Workshop – Shedding Light on obscure Graph Deep Learni
 ng
DTSTART:20210929T130000
DTEND:20210929T170000
DTSTAMP:20260511T081145Z
UID:3c1cad3ee9a2a70235ed9aa51b5fd0fb25371dcee2faf9402cbb2328
CATEGORIES:Conferences - Seminars
DESCRIPTION:⚠️ A valid COVID certificate must be presented on site t
 o enter the event. ⚠️\n\nEver wondered why that molecule is hydrophobi
 c? Or why a region is constantly jammed with traffic? Or why certain peopl
 e in a social network might know each other?\n\nAll these problems can be 
 modelled by using powerful data structures: graphs\, made of entities (nod
 es\, such as the atoms of a molecule) and relationships (edges between the
  nodes\, such as the chemical bonds). These can be fed into machine learni
 ng models which leverage relationship information for predictions: for exa
 mple\, a model can be trained to classify whether a molecule will be hydro
 phobic or not. In particular\, deep learning (DL) has recently been extend
 ed to work on graphs\, with the advent of graph neural networks (GNNs). \
 n\nUp until now and like in many other domains\, deep learning on graphs\,
  albeit powerful\, was completely obscure. In fact\, DL models all lack in
 herent interpretability.\nWith the recent introduction of post-hoc interpr
 etability techniques\, light was shed on several DL models. Last year\, th
 is was made possible on graphs too\, thanks to the advent of several new i
 nterpretability techniques.\nThis opens up a world of possibilities to bet
 ter understand how these models leverage complex relationships between ent
 ities during their predictions.\n\nDuring this workshop\, the participants
  will learn how to model a variety of problems using graphs\, train GNNs o
 n them and apply state-of-the-art techniques to interpret the underlying m
 otivations that led to their predictions.
LOCATION:STCC
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR