IADD-TR factorizes MBRL transitions into an action-intervention stage and an action-free evolution stage using a zero-action anchor, and adds targeted regularization to the critic to reduce policy-gradient bias.
Guided policy search,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
IADD-TR: Intervention-Aware Dynamics Decoupling with Targeted Regularization for Model-Based Reinforcement Learning
IADD-TR factorizes MBRL transitions into an action-intervention stage and an action-free evolution stage using a zero-action anchor, and adds targeted regularization to the critic to reduce policy-gradient bias.