AI Paper Atlas

Research map / Reinforcement learning / Visual and representation RL

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

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

CVPR 2026 · Reinforcement learning › Visual and representation RL · approach: Dexterous manipulation transfer · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add mask-invariant KL loss (lambda 1.0) to masked base-policy training (ablation)MaskAdapt base policy · LAFAN1 coveragenormalized entropy: 0.7448 -> 0.9025Improved
use weaker MI loss weight (0.1 vs 1.0) (ablation)MaskAdapt base policy · LAFAN1 coveragenormalized entropy: 0.8968 vs 0.9025No clear effect
apply stochastic body-part masking without MI loss (ablation)base policy · LAFAN1 coveragenormalized entropy: 0.9124 -> 0.7448Hurt
use MaskAdapt residual policy instead of CMLmotion composition · Case (a) preserved regionnormalized entropy: 0.7835 -> 0.8580Improved
use MaskAdapt instead of CMLmotion composition · Case (b) adapted regionnormalized entropy: 0.7478 -> 0.8710Improved
use MaskAdapt instead of CMLmotion composition · Case (c) preserved regionnormalized entropy: 0.8461 -> 0.8980Improved
use MaskAdapt instead of CML on Strike taskgoal-driven task · Strikesuccess rate (%): 46.5 -> 64.2Improved
use MaskAdapt instead of CML on Location taskgoal-driven task · Locationsuccess rate (%): 100.0 -> 99.5No clear effect
use MaskAdapt instead of CML on Heading taskgoal-driven task · Headingsuccess rate (%): 75.2 -> 76.9No clear effect
residual on mask-invariant base instead of from-scratch policypartial goal tracking · Partial tracking (Arms)tracking error (m): 0.076 -> 0.055Improved
use MaskAdapt instead of CMLpartial goal tracking · Partial tracking (Arms)DTW pose error (m): 0.079 -> 0.047Improved
use arm-only masked residual vs MaskedMimic arm trackingpartial goal tracking · Arm trackingtracking error (m): 0.0849 -> 0.0423Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy