BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:An Alternative to Backpropagation - Varun Jain
DTSTART:20250514T150000Z
DTEND:20250514T153000Z
UID:TALK231028@talks.cam.ac.uk
CONTACT:Cat Spencer
DESCRIPTION:For decades\, backpropagation has been the backbone of deep le
 arning. However\, it is biologically implausible\, computationally demandi
 ng\, and sometimes inefficient. This has spurred the search for alternativ
 e training paradigms. I will introduce a novel gradient‑free approach in
  which the objective is to align the distribution of a network’s linear 
 regions with that of the training data. On CIFAR‑10\, this approach surp
 asses the performance achieved by Adam\, which demonstrates that carefully
  structuring a model’s linear regimes can provide a powerful alternative
  to traditional gradient‑based methods.
LOCATION: Cambridge University Engineering Department\, CBL Seminar room B
 E4-38.  For directions see http://learning.eng.cam.ac.uk/Public/Directions
END:VEVENT
END:VCALENDAR