BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:From Coarse to Fine and Back Again: Geometry\, Topology\, and Deep
  Learning
DTSTART:20250416T141500
DTEND:20250416T151500
DTSTAMP:20260417T141054Z
UID:27ecd8fe527919f6b2b855719bdb89d44d82bc3d8f1fceddfa903e4c
CATEGORIES:Conferences - Seminars
DESCRIPTION:Bastian Grossenbacher\, University of Fribourg\nA large driver
  contributing to the success of deep-learning models is their ability to s
 ynthesize task-specific features from data\, which typically outshine hand
 -crafted features. Thus\, for a long time\, the predominant belief was tha
 t "given enough data\, all features can be learned." However\, it turns ou
 t that certain tasks require imbuing models with *inductive biases* such a
 s invariance or equivariance properties that **cannot** be readily gleaned
  from the data!\nThis is particularly true for data sets that model real-w
 orld phenomena\, including those that involve relations beyond the dyadic\
 , creating a crucial need for different approaches.\n\nIn this talk\, I wi
 ll present novel advances in harnessing multi-scale geometrical-topologica
 l characteristics of data\, focusing in particular on how the tandem of ge
 ometry and topology can improve (un)supervised tasks in representation lea
 rning. Underscoring the generality of a hybrid geometrical-topological per
 spective\, I will furthermore showcase applications from a diverse set of 
 data domains\, including point clouds\, graphs\, and higher-order combinat
 orial complexes.\n 
LOCATION:CM 1 517 https://plan.epfl.ch/?room==CM%201%20517
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR