SuperPADL: Scaling Language-Directed Physics-Based Control with Progressive Supervised Distillation
ACM SIGGRAPH 2024
Jordan Juravsky (1, 2)Yunrong Guo (2)Sanja Fidler (2, 3)Xue Bin Peng (2, 4)
(1) Stanford(2) NVIDIA(3) University of Toronto(4) Simon Fraser University
Abstract
Physically-simulated models for human motion can generate highquality
responsive character animations, often in real-time. Natural
language serves as a flexible interface for controlling these models,
allowing expert and non-expert users to quickly create and
edit their animations. Many recent physics-based animation methods,
including those that use text interfaces, train control policies
using reinforcement learning (RL). However, scaling these methods
beyond several hundred motions has remained challenging.
Meanwhile, kinematic animation models are able to successfully
learn from thousands of diverse motions by leveraging supervised
learning methods. Inspired by these successes, in this work we
introduce SuperPADL, a scalable framework for physics-based textto-
motion that leverages both RL and supervised learning to train
controllers on thousands of diverse motion clips. SuperPADL is
trained in stages using progressive distillation, starting with a large
number of specialized experts using RL. These experts are then
iteratively distilled into larger, more robust policies using a combination
of reinforcement learning and supervised learning. Our
final SuperPADL controller is trained on a dataset containing over
5000 skills and runs in real time on a consumer GPU. Moreover, our
policy can naturally transition between skills, allowing for users
to interactively craft multi-stage animations. We experimentally
demonstrate that SuperPADL significantly outperforms RL-based
baselines at this large data scale.
@inproceedings{
juravsky2024superpadl,
title = {SuperPADL: Scaling Language-Directed Physics-Based Control with Progressive Supervised Distillation},
author = {Jordan Juravsky and Yunrong Guo and Sanja Fidler and Xue Bin Peng},
booktitle = {SIGGRAPH 2024 Conference Papers (SIGGRAPH '24 Conference Papers),},
year = {2024}
