No More Stale Feedback: Co-Evolving Critics for Open-World Agent Learning

TL;DR

ECHO introduces a co-evolving critic and policy framework for reinforcement learning, enhancing stability and success in open-world environments by addressing critic staleness.

Contribution

The paper proposes ECHO, a novel co-evolutionary approach that synchronizes critic and policy updates to improve reinforcement learning in dynamic, open-world settings.

Findings

ECHO achieves more stable training compared to static critic methods.

ECHO results in higher long-horizon task success rates.

ECHO effectively mitigates critic staleness in evolving policies.

Abstract

Critique-guided reinforcement learning (RL) has emerged as a powerful paradigm for training LLM agents by augmenting sparse outcome rewards with natural-language feedback. However, current methods often rely on static or offline critic models, which fail to adapt as the policy evolves. In on-policy RL, the agent's error patterns shift over time, causing stationary critics to become stale and providing feedback of diminishing utility. To address this, we introduce ECHO (Evolving Critic for Hindsight-Guided Optimization)}, a framework that jointly optimizes the policy and critic through a synchronized co-evolutionary loop. ECHO utilizes a cascaded rollout mechanism where the critic generates multiple diagnoses for an initial trajectory, followed by policy refinement to enable group-structured advantage estimation. We address the challenge of learning plateaus via a saturation-aware gain shaping objective, which rewards the critic for inducing incremental improvements in high-performing trajectories. By employing dual-track GRPO updates, ECHO ensures the critic's feedback stays synchronized with the evolving policy. Experimental results show that ECHO yields more stable training and higher long-horizon task success across open-world environments.