BEGIN:VCALENDAR
CALSCALE:GREGORIAN
VERSION:2.0
METHOD:PUBLISH
PRODID:-//Drupal iCal API//EN
X-WR-TIMEZONE:America/New_York
BEGIN:VTIMEZONE
TZID:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
DTSTART:20070311T020000
TZNAME:EDT
TZOFFSETTO:-0400
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
DTSTART:20071104T020000
TZNAME:EST
TZOFFSETTO:-0500
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
SEQUENCE:1
X-APPLE-TRAVEL-ADVISORY-BEHAVIOR:AUTOMATIC
234016
20260402T135235Z
DTSTART;TZID=America/New_York:20260409T100000
DTEND;TZID=America/New_York:2
 0260409T110000
URL;TYPE=URI:https://www.wpi.edu/news/calendar/events/compu
 ter-science-department-phd-defense-yang-wu-advancing-proactive-capacity-ll
 ms-cost-sensitive
Computer Science Department , PhD Defense  Yang Wu "   Advancing the Proactive Capacity of LLMs in Cost-Sensitive Domains"
\nYang Wu\nPhD Candidate\nWPI – Computer Science Department\nThursday, April9th, 2026\nTime:
 10:00 AM-11:30AM\nLocation: Innovation Studio 105\nZoom Link:https://wpi.z
 oom.us/j/6065536440\nCommittee members :\nProf. Xiaozhong Liu (Advisor), W
 PI-Computer Science\nProf. Kyumin Lee, WPI-Computer Science\nProf. Frank Z
 ou, WPI-Mathematical Science\nProf. Haixu Tang, Indiana University Bloomin
 gton-Data Science\n\nAbstract:\nRecent advances in large language models (
 LLMs) have demonstrated remarkable capabilities across a wide range of tas
 ks; however, their effectiveness remains limited in cost-sensitive domains
  such as biomedical research, legal reasoning, and personalized education,
  where high-quality data is scarce, expert knowledge is expensive, and rea
 l-time interaction constraints are critical.\nThis dissertation addresses 
 how to advance the proactive capacity of LLMs under such constraints, enab
 ling them not only to respond to user queries but also to actively acquire
  knowledge, adapt to domain-specific needs, and collaborate effectively wi
 th humans. To this end, it presents a unified framework spanning three dim
 ensions: cost-efficient data acquisition, proactive domain assistance, and
  proactive team-AI collaboration. For data efficiency, we develop PUtree, 
 PU-ADKA, and ROSE to improve model performance under limited supervision a
 nd constrained expert budgets.\nFor domain assistance, we design systems s
 uch as D3LM, TASA, and CamPilot that enable LLMs to actively elicit missin
 g information, personalize interactions, and generate structured outputs a
 cross legal, educational, and multimodal settings. For collaborative intel
 ligence, we introduce CoLabScience and DTR, which enable LLMs to identify 
 when and how to intervene in ongoing scientific discussions and achieve lo
 w-latency, high-quality interaction through offline rehearsal. Together, t
 hese contributions transform LLMs from passive responders into proactive c
 ollaborators and provide a principled approach for deploying LLM systems i
 n real-world, high-stakes, and cost-sensitive environments.\n
END:VEVENT
END:VCALENDAR