BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Geometric Deep Learning: Grids\, Graphs\, Groups\, Geodesics\, and
  Gauges - Petar Veličković
DTSTART:20210518T121500Z
DTEND:20210518T131500Z
UID:TALK160276@talks.cam.ac.uk
CONTACT:Mateja Jamnik
DESCRIPTION:"Join us on Zoom":https://zoom.us/j/99166955895?pwd=SzI0M3pMVE
 kvNmw3Q0dqNDVRalZvdz09\n\nThe last decade has witnessed an experimental re
 volution in data science and machine learning\, epitomised by deep learnin
 g methods. Indeed\,\nmany high-dimensional learning tasks previously thoug
 ht to be beyond reach –such as computer vision\, playing Go\, or protein
  folding – are in fact feasible with appropriate computational scale. Re
 markably\, the essence of deep learning is built from two simple algorithm
 ic principles: first\, the notion of representation or feature learning\, 
 whereby adapted\, often hierarchical\, features capture the appropriate no
 tion of regularity for each task\, and second\, learning by local gradient
 -descent type methods\, typically implemented as backpropagation.\n\nWhile
  learning generic functions in high dimensions is a cursed estimation prob
 lem\, most tasks of interest are not generic\, and come with essential pre
 -defined regularities arising from the underlying low-dimensionality and s
 tructure of the physical world. This talk is concerned with exposing these
  regularities through unified geometric principles that can be applied thr
 oughout a wide spectrum of applications.\n\nSuch a ‘geometric unificatio
 n’ endeavour in the spirit of Felix Klein's Erlangen Program serves a du
 al purpose: on one hand\, it provides a common mathematical framework to s
 tudy the most successful neural network architectures\, such as CNNs\, RNN
 s\, GNNs\, and Transformers. On the other hand\, it gives a constructive p
 rocedure to incorporate prior physical knowledge into neural architectures
  and provide principled way to build future architectures yet to be invent
 ed.
LOCATION:Zoom
END:VEVENT
END:VCALENDAR