MaskAdapt: Learning Flexible Motion Adaptation via Mask-Invariant Prior for Physics-Based Characters

Soomin Park, Eunseong Lee, Kwang Bin Lee, Sung-Hee Lee · 2026

Read MaskAdapt: Learning Flexible Motion Adaptation via Mask-Invariant Prior for Physics-Based Characters by Soomin Park, Eunseong Lee, Kwang Bin Lee, Sung-Hee Lee (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-29272.txt.

Description

Park and co-authors present MaskAdapt, a framework for adapting motion in physics-based humanoid control. Training runs in two stages. First, a base policy learns with random body parts masked out. A regularizer keeps actions consistent across masking conditions. That yields a stable motion prior under missing observations. Next, a residual policy trains on the frozen base. It changes only targeted body parts and preserves behavior elsewhere. Uses include motion composition and text-driven partial goal tracking.

Catalog record

CREATORSoomin Park, Eunseong Lee, Kwang Bin Lee, Sung-Hee Lee
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE56793
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.29272

Browse the full library · About the Tower of Babel