BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//wp-events-plugin.com//7.2.3.1//EN
TZID:America/New_York
X-WR-TIMEZONE:America/New_York
BEGIN:VEVENT
UID:851@fds.yale.edu
DTSTART;TZID=America/New_York:20250416T113000
DTEND;TZID=America/New_York:20250416T130000
DTSTAMP:20250916T142150Z
URL:https://fds.yale.edu/events/fds-colloquium-elliot-paquette-mcgill-high
 -dimensional-optimization-with-applications-to-compute-optimal-neural-scal
 ing-laws/
SUMMARY:FDS Colloquium: Elliot Paquette (McGill)\, "High-dimensional Optimi
 zation with Applications to Compute-Optimal Neural Scaling Laws"
DESCRIPTION:\nAbstract:&nbsp\;Given the massive scale&nbsp\;of&nbsp\;modern
  ML models\, we now only get a single shot to train them effectively. This
  restricts our ability to test multiple architectures and hyper-parameter 
 configurations. Instead\, we need to understand how these models scale\, a
 llowing us to experiment with smaller problems and then apply those insigh
 ts to larger-scale models. In this talk\, I will present a framework for a
 nalyzing scaling laws in stochastic learning algorithms using a power-law 
 random features model\, leveraging high-dimensional probability and random
  matrix theory. I will then use this scaling law to address the compute-op
 timal question: How should we choose model size and hyper-parameters to ac
 hieve the best possible performance in the most compute-efficient manner?\
 n\n\n\nBased on joint work with Courtney Paquette (McGill &amp\; Google De
 epmind)\, Jeffrey Pennington (Google Deepmind)\, and Lechao Xiao&nbsp\;(Go
 ogle Deepmind).\n\n\n\nSpeaker&nbsp\;Bio: Elliot Paquette is a researcher 
 in probability theory\, and especially theory of high dimensional optimiza
 tion and machine learning. &nbsp\;He received his PhD in 2013 at the Unive
 rsity of Washington\, in random matrix theory.&nbsp\; He was an NSF postdo
 ctoral researcher at Weizmann Institute of Science\, working at the interf
 ace of random matrix theory and branching processes.&nbsp\; In 2016\, he j
 oined Ohio State University as an assistant professor.&nbsp\; In 2020\, he
  moved to McGill University\, department of Mathematics and Statistics\, w
 here he is now associate professor.\n\n\n\nelliotpaquette.github.io\n
CATEGORIES:FDS Events,Colloquium
LOCATION:Yale Institute for Foundations of Data Science\, Kline Tower 13th 
 Floor\, Room 1327\, New Haven\, CT\, 06511\, United States
X-APPLE-STRUCTURED-LOCATION;VALUE=URI;X-ADDRESS=Kline Tower 13th Floor\, Ro
 om 1327\, New Haven\, CT\, 06511\, United States;X-APPLE-RADIUS=100;X-TITL
 E=Yale Institute for Foundations of Data Science:geo:0,0
END:VEVENT
BEGIN:VTIMEZONE
TZID:America/New_York
X-LIC-LOCATION:America/New_York
BEGIN:DAYLIGHT
DTSTART:20250309T030000
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
END:DAYLIGHT
END:VTIMEZONE
END:VCALENDAR