T2RD decouples content and style features then refines task-relevant ones via dynamic prediction to reach SOTA generalization on DeepMind Control Suite and robotic manipulation tasks.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Task-Relevant Representation Decoupling for Visual Reinforcement Learning Generalization
T2RD decouples content and style features then refines task-relevant ones via dynamic prediction to reach SOTA generalization on DeepMind Control Suite and robotic manipulation tasks.