Tag
This paper proposes action-conditioned bisimulation over an empirical predictive state graph to decide when GUI agent memories of two web pages should be merged, improving success on MiniWoB++ over memoryless baselines without any training.
Introduces Action-Conditioned Predictive Consistency (ACPC), a diagnostic for JEPA world models that measures how clean and perturbed observations diverge under action-conditioned rollouts, with theoretical bounds on prediction error and planner cost. Experiments on visual control tasks validate the diagnostic across models like LeWM and PLDM.
BiPACE introduces a drop-in advantage estimator that fixes state-action credit mismatch in stepwise group-based RL for LLM agents, using bisimulation-guided state clustering and action counterfactual estimation, achieving significant performance gains on ALFWorld, WebShop, and TextCraft with Qwen2.5 models.
This paper introduces CARL, a method for offline hierarchical reinforcement learning that exploits local dynamics regularity to learn reusable skills. The approach clusters state-goal pairs requiring similar action sequences, enabling more effective skill reuse and improved performance on complex humanoid tasks.