BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Iowa State University CALS LAS Web Team//sites.iastate.edu//EN
BEGIN:VEVENT
UID:20250306T221500-4671-www.cs.iastate.edu
DTSTART:20250306T221500Z
SEQUENCE:0
TRANSP:OPAQUE
DTEND:20250306T231500Z
LOCATION:2200 Marston
SUMMARY:CS Colloquium: Dr. Vaneet Aggarwal\, Purdue University
CLASS:PUBLIC
DESCRIPTION:Order-Optimal Sample Complexity for Reinforcement LearningDeep 
 Reinforcement Learning (DRL) has seen tremendous advancements\, yet achiev
 ing optimal sample complexity remains a fundamental challenge. This talk p
 resents recent progress in developing order-optimal sample complexity guar
 antees for RL algorithms with general parametrization. We first discuss th
 e key difficulties in achieving optimal sample complexity and introduce an
  accelerated natural policy gradient (ANPG) approach. We will further summ
 arize some extensions of the approach. Our results bridge crucial gaps in 
 RL theory\, offering practical implications for scalable and efficient dec
 ision-making systems.\n\nMore information at: https://www.cs.iastate.edu/e
 vent/2025/cs-colloquium-dr-vaneet-aggarwal-purdue-university
DTSTAMP:20260510T170534Z
END:VEVENT
END:VCALENDAR