RTMC: Step-Level Credit Assignment via Rollout Trees
For multi-step agentic RL, RTMC provides a lightweight alternative to learned value networks that avoids overhead and fragility under sparse rewards.
RTMC introduces a critic-free advantage estimation method that aggregates returns across rollouts sharing intermediate states, improving pass@1 by 3.2 percentage points over GRPO on SWE-bench Verified.
Multi-step agentic reinforcement learning benefits from fine-grained credit assignment, yet existing approaches offer limited options: critic-free methods like GRPO assign a uniform advantage to every action in a trajectory, while learned value networks introduce notable overhead and can be fragile under sparse rewards. We observe that group rollouts targeting the same problem often traverse overlapping intermediate states, implicitly forming a tree whose branches diverge at successive decision points. Building on this insight, we introduce Rollout-Tree Monte Carlo (RTMC) advantage estimation, which aggregates return statistics across rollouts sharing a common state to produce per-step Q-values and advantages--without any learned critic. A state-action signature system compresses raw interaction histories into compact, comparable representations, making cross-rollout state matching tractable. On SWE-bench Verified, RTMC improves pass@1 by 3.2 percentage points over GRPO.