BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:PDE-GCN: Novel Architectures for Graph Neural Networks Motivated b
 y Partial Differential Equations    - Moshe Eliasof
DTSTART:20221201T130000Z
DTEND:20221201T140000Z
UID:TALK193196@talks.cam.ac.uk
CONTACT:AI Aviles-Rivero
DESCRIPTION:Graph neural networks are increasingly becoming the go-to appr
 oach in various fields such as computer vision\, computational biology and
  chemistry\, where data are naturally explained by graphs. However\, unlik
 e traditional convolutional neural networks\, deep graph networks do not n
 ecessarily yield better performance than shallow graph networks. This beha
 vior usually stems from the over-smoothing phenomenon. In this work\, we p
 ropose a family of architectures to control this behavior by design. Our n
 etworks are motivated by numerical methods for solving Partial Differentia
 l Equations (PDEs) on manifolds\, and as such\, their behavior can be expl
 ained by similar analysis. Moreover\, as we demonstrate using an extensive
  set of experiments\, our PDE-motivated networks can generalize and be eff
 ective for various types of problems from different fields. Our architectu
 res obtain better or on par with the current state-of-the-art results for 
 problems that are typically approached using different architectures. 
LOCATION:MR3\,  Centre for Mathematical Sciences\, Wilberforce Road\, Camb
 ridge
END:VEVENT
END:VCALENDAR