BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Asia/Tokyo
X-LIC-LOCATION:Asia/Tokyo
BEGIN:STANDARD
TZOFFSETFROM:+0900
TZOFFSETTO:+0900
TZNAME:JST
DTSTART:18871231T000000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20250110T023312Z
LOCATION:Hall B7 (1)\, B Block\, Level 7
DTSTART;TZID=Asia/Tokyo:20241205T130000
DTEND;TZID=Asia/Tokyo:20241205T131400
UID:siggraphasia_SIGGRAPH Asia 2024_sess132_papers_1034@linklings.com
SUMMARY:PDP: Physics-Based Character Animation via Diffusion Policy
DESCRIPTION:Technical Papers\n\nTakara Truong, Michael Piseno, Zhaoming Xi
 e, and Karen Liu (Stanford University)\n\nGenerating diverse and realistic
  human motion that can physically interact with an environment remains a c
 hallenging research area in character animation. Meanwhile, diffusion-base
 d methods, as proposed by the robotics community, have demonstrated the ab
 ility to capture highly diverse and multi-modal skills. However, naively t
 raining a diffusion policy often results in unstable motions for high-freq
 uency, under-actuated control tasks like bipedal locomotion due to rapidly
  accumulating compounding errors, pushing the agent away from optimal trai
 ning trajectories.  The key idea lies in using RL policies not just for pr
 oviding optimal trajectories but for providing corrective actions in sub-o
 ptimal states which gives the policy a chance to correct for errors caused
  by environmental stimulus, model errors, or numerical errors in simulatio
 n. Our method, Physics-Based Character Animation via Diffusion Policy (PDP
 ), combines reinforcement learning (RL) and behavior cloning (BC) to creat
 e a robust diffusion policy for physics-based character animation. We demo
 nstrate PDP on perturbation recovery, universal motion tracking, and physi
 cs-based text-to-motion synthesis.\n\nRegistration Category: Full Access, 
 Full Access Supporter\n\nLanguage Format: English Language\n\nSession Chai
 r: Yi Zhou (Adobe)
URL:https://asia.siggraph.org/2024/program/?id=papers_1034&sess=sess132
END:VEVENT
END:VCALENDAR