1 h 15 min
Speculations on Test-Time Scaling | Richard M. Karp Distinguished Lecture
Artificial Intelligence, Machine Learning, Natural Language Processing and seminar by Sasha Rush, Cornell University
Hosted by Simons Institute for the Theory of Computing, UC Berkeley
Friday 11:00 Los Angeles (GMT-8)
Recording available
Berkeley, United States
Recording
Abstract
Sasha Rush surveys research on increasing language-model capability through computation at inference time. The lecture places test-time scaling alongside parameter and data scaling, examines model self-improvement and discusses implications for the design of reasoning models and open-source research efforts.
Topics
language-model capabilitycomputation at inference timetest-time scalingData scalingmodel self-improvementreasoning modelsopen-source research
More from Simons Institute for the Theory of Computing, UC Berkeley
All talksOptimal Spectral Gaps | Richard M. Karp Distinguished Lecture
Oct 14, 2026Ramon van HandelMy Favorite Expanders | Richard M. Karp Distinguished Lecture
Sep 29, 2026Ryan O'DonnellEncrypting to a Mathematical Fact | Theoretically Speaking
Video · 1 h 24 min · Captions · Jul 15, 2026Sanjam GargResponsibly Improving AI with Privacy-Sensitive Data: Principles, Theory, and Practice | Richard M. Karp Distinguished Lecture
Video · 1 h 10 min · Captions · Feb 24, 2026Brendan McMahan