Presentation / Installation
PMP: Learning to Physically Interact With Environments Using Part-wise Motion Priors
DescriptionWe propose a reinforcement learning framework to enable complex levels of physical interaction for a whole-body character. Our method leverages part-wise skills from multiple sources and combines them to generate novel skills. Experiments show our approach encourages agents to exhibit natural motions while performing in challenging scenarios like rope climbing.
Event Type
Technical Paper
TimeThursday, 10 August 202311:27am - 11:37am PDT
LocationPetree Hall C
ACM Digital Library
Technical Paper PDF
Session Time & Location
Research & Education
Livestreamed
Recorded
Animation/Simulation
Artificial Intelligence/Machine Learning
FC
FCS
V
VS
EFC