AMP: Adversarial Motion Priors for Stylized Physics-based Character Control
Event Type
Technical Paper
This session WILL be recorded.
Interest Areas
Research & Education
Registration Levels
Ultimate Supporter
Ultimate Attendee
Exhibitor Ultimate
TimeThursday, 12 August 20214pm - 5pm PDT
DescriptionWe present an adversarial imitation learning approach that enables physically simulated characters to perform challenging tasks while imitating behaviors from diverse and unstructured motion datasets.