BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//wp-events-plugin.com//7.2.3.1//EN
TZID:Asia/Kolkata
X-WR-TIMEZONE:Asia/Kolkata
BEGIN:VEVENT
UID:169@cds.iisc.ac.in
DTSTART;TZID=Asia/Kolkata:20260106T110000
DTEND;TZID=Asia/Kolkata:20260106T120000
DTSTAMP:20251224T073530Z
URL:https://cds.iisc.ac.in/events/seminar-cds-102-january-06th-1100-the-ca
 se-for-limitation-aware-llms/
SUMMARY:{Seminar} @ CDS: #102\, January 06th: 11:00: "The case for Limitati
 on-Aware LLMs."
DESCRIPTION:Department of Computational and Data Sciences\nDepartment Semin
 ar\n\n\n\nSpeaker : Dr. Tanya Goyal\, Cornell University\nTitle : The case
  for Limitation-Aware LLMs\nDate &amp\; Time: January 06th\, 2026 (Tuesday
 )\, 11:00 AM\nVenue : # 102\, CDS Seminar Hall\n\n\n\nABSTRACT\n\nAs we in
 creasingly integrate large language models (LLMs) into our workflows\, it 
 is important for LLMs to act as effective collaborators. This involves a m
 ultitude of skills — estimating their own capability and knowledge bound
 aries to decide when to seek help\, estimating collaborator capability\, e
 ffectively using the signal returned by them\, amongst others. In this tal
 k\, I will discuss recent work from my lab that teaches LLMs to effectivel
 y abstain on questions outside their knowledge boundaries. We train LLMs t
 o learn to seek external help\, e.g. via search tools\, by rewarding accur
 acy but simultaneously penalizing this external help. I will discuss resul
 ts that show that by considering any search invocation as an abstention\, 
 we can effectively turn LLMs trained with such a pay-per-search into abste
 ntion models. Finally\, I will talk about some recent work that benchmarks
  LLMs robustness when collaborating “off-trajectory”\, i.e with out-of
 -distribution help\, and discuss future directions in this space.\n\nBIO: 
 Tanya Goyal is an assistant professor in the Computer Science department a
 t Cornell University. Her research interests include building collaborativ
 e large language models (LLMs)\, reliable and sustainable evaluation frame
 works for such LLMs\, as well as understanding LLM behaviors as a function
  of training data and/or alignment strategies. Tanya completed her Ph.D. i
 n Computer Science at UT Austin in 2023 where her thesis was awarded UTCS'
 s Bert Kay Dissertation award. Her research is supported by NSF and a rese
 arch award from Google.\n\nHost Faculty: Dr. Danish Pruthi\n\n\n\nALL ARE 
 WELCOME
CATEGORIES:Events,Talks
END:VEVENT
BEGIN:VTIMEZONE
TZID:Asia/Kolkata
X-LIC-LOCATION:Asia/Kolkata
BEGIN:STANDARD
DTSTART:20250106T110000
TZOFFSETFROM:+0530
TZOFFSETTO:+0530
TZNAME:IST
END:STANDARD
END:VTIMEZONE
END:VCALENDAR