BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//CAIL//Events//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:CAIL Events
BEGIN:VEVENT
UID:2026-10-16-aviral-kumar@cail.columbia.edu
DTSTAMP:20261015T150141Z
DTSTART:20261016T150000Z
DTEND:20261016T160000Z
SUMMARY:ML Seminar: Aviral Kumar - Is Exploration on Hard Problems in LLM 
 RL a Solved Problem?
LOCATION:SSW Room 903
DESCRIPTION:Register: https://www.eventbrite.com/e/mlai-seminar-aviral-kum
 ar-carnegie-mellon-university-tickets-1999106686915\n\nThere has been rema
 rkable progress in using LLMs to solve hard problems\, even just over the 
 past few months\, driven by scaling both pretraining and post-training. Ma
 ny of these problems involve a fundamental search or “exploration” bot
 tleneck\, as it is known in the reinforcement learning literature. But doe
 s recent progress mean that we have solved exploration? In other words\, c
 an we take a sufficiently powerful pretrained model\, scale its test-time 
 compute (TTC)\, and expect it to solve essentially any hard problem? In th
 is talk\, I will take a broader view of exploration in LLMs\, give you a h
 igh-level view on some of our past work on this problem over the last 1 ye
 ar\, and discuss recent efforts from my group to push the limits of explor
 ation using multi-agent systems and weight-update-based internalization. I
  will conclude with several open questions about what remains unsolved in 
 my biased opinion and where progress will come from next.\n\nhttps://cail.
 columbia.edu/events/2026-10-16-aviral-kumar
URL:https://cail.columbia.edu/events/2026-10-16-aviral-kumar
END:VEVENT
END:VCALENDAR