BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4//
BEGIN:VEVENT
UID:20260515T212339EDT-72116fxtsS@132.216.98.100
DTSTAMP:20260516T012339Z
DESCRIPTION:Abstract\n\nEfficient learning is a hallmark of human intellige
 nce\; from infancy\, we have the remarkable ability to learn novel concept
 s from very few examples\, using a brain that runs on the energy equivalen
 t to an electric razor. Replicating such behaviour in computers is a long-
 standing challenge of machine learning research with the potential to yiel
 d material benefits for society.\n\nThis thesis improves the efficiency of
  learning by producing algorithmic advances in several directions. The fir
 st part of this thesis presents theoretical and empirical advances in nume
 rical optimization that enable more efficient training of large-scale mach
 ine learning models on distributed computing devices\, while the second pa
 rt of this thesis presents theoretical and empirical advances in represent
 ation learning that improve the label-efficiency of learning. Indeed\, the
  ability of humans to quickly acquire new concepts from few examples depen
 ds greatly on the many previously constructed abstractions and prior exper
 iences\, and one way for an agent to encode prior knowledge and experience
  is by learning to represent data in ways that facilitate processing.\n\nT
 ogether\, Parts I and II provide progress towards learning methods that ca
 n more efficiently utilize distributed training hardware and training data
 \, so as to build more efficient learning machines. The development of mor
 e efficient learning frameworks presents the potential to democratize the 
 practice of machine learning by reducing the computational burden of model
  training and enabling more effective learning in low-resource settings. F
 undamental advances in learning efficiency may also turn out to be chiefly
  important for longer term goals towards advancing machine intelligence.\n
DTSTART:20230323T163000Z
DTEND:20230323T183000Z
LOCATION:Room 603\, McConnell Engineering Building\, CA\, QC\, Montreal\, H
 3A 0E9\, 3480 rue University
SUMMARY:PhD defence of Mido Assran – “Algorithmic Advances Towards Efficien
 t Learning Machines”
URL:https://www.mcgill.ca/ece/channels/event/phd-defence-mido-assran-algori
 thmic-advances-towards-efficient-learning-machines-347190
END:VEVENT
END:VCALENDAR