BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Australia/Melbourne
X-LIC-LOCATION:Australia/Melbourne
BEGIN:DAYLIGHT
TZOFFSETFROM:+1000
TZOFFSETTO:+1100
TZNAME:AEDT
DTSTART:19721003T020000
RRULE:FREQ=YEARLY;BYMONTH=4;BYDAY=1SU
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:19721003T020000
TZOFFSETFROM:+1100
TZOFFSETTO:+1000
TZNAME:AEST
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260114T163644Z
LOCATION:Meeting Room C4.11\, Level 4 (Convention Centre)
DTSTART;TZID=Australia/Melbourne:20231212T141000
DTEND;TZID=Australia/Melbourne:20231212T142500
UID:siggraphasia_SIGGRAPH Asia 2023_sess120_papers_543@linklings.com
SUMMARY:AdaptNet: Policy Adaptation for Physics-Based Character Control
DESCRIPTION:Pei Xu (Clemson University, Roblox); Kaixiang Xie (McGill Univ
 ersity); Sheldon Andrews (École de technologie supérieure, Roblox); Paul G
 . Kry (McGill University); Michael Neff (University of California Davis); 
 Morgan McGuire (Roblox, University of Waterloo); Ioannis Karamouzas (Unive
 rsity of California Riverside); and Victor Zordan (Roblox, Clemson Univers
 ity)\n\nMotivated by human’s ability to adapt skills in the learning of ne
 w ones, this paper presents AdaptNet, an approach for modifying the latent
  space of existing policies to allow new behaviors to be quickly learned f
 rom like tasks in comparison to learning from scratch. Building on top of 
 a given reinforcement learning controller, AdaptNet uses a two-tier hierar
 chy that augments the original state embedding to support modest changes i
 n a behavior and further modifies the policy network layers to make more s
 ubstantive changes. The technique is shown to be effective for adapting ex
 isting physics-based controllers to a wide range of new styles for locomot
 ion, new task targets, changes in character morphology, and extensive chan
 ges in environment. Furthermore, it exhibits significant increase in learn
 ing efficiency, as indicated by greatly reduced training times when compar
 ed to training from scratch or using other approaches that modify existing
  policies.\n\nRegistration Category: Full Access\n\nSession Chair: Jungdam
  Won (Seoul National University)\n\n
URL:https://asia.siggraph.org/2023/full-program?id=papers_543&sess=sess120
END:VEVENT
END:VCALENDAR