2605.30159v1 May 28, 2026 cs.AI

Meta-Cognitive Memory Policy Optimization for Long-Horizon LLM Agents

Zhezheng Hao
Zhezheng Hao
Citations: 69
h-index: 4
Hong Wang
Hong Wang
Citations: 62
h-index: 3
Ruiying Ding
Ruiying Ding
Citations: 18
h-index: 2
Jingren Hou
Jingren Hou
Citations: 3
h-index: 1
Feng Liu
Feng Liu
Citations: 2
h-index: 1
Yeqiu Chen
Yeqiu Chen
Citations: 9
h-index: 1
Wence Ji
Wence Ji
Citations: 2
h-index: 1
Weibing Xia
Weibing Xia
Citations: 19
h-index: 1
Yongkang Yang
Yongkang Yang
Citations: 1
h-index: 1
Ziyan Liu
Ziyan Liu
Citations: 597
h-index: 8

Memory-augmented LLM agents tackle complex long-horizon tasks by recursively summarizing interaction trajectories into compact memory. However, existing approaches typically train these memory policies using outcome-based reinforcement learning, failing to localize where intermediate memory quality degrades. As interactions unfold, ambiguous recursive summaries progressively discard task-relevant information and introduce semantic noise. This exacerbates belief deviation, obscuring the agent's estimate of the latent task state and ultimately derailing long-horizon reasoning. We therefore argue that memory optimization should focus not merely on trajectory-level success, but on the clarity of the belief induced by intermediate summaries. To this end, we introduce Belief Entropy, a self-supervised proxy that probes how uncertain the model remains about the latent task state given its current memory. Based on this proxy, we propose Metacognitive Memory Policy Optimization (MMPO). Instead of relying only on sparse outcome-based signals, MMPO provides fine-grained, memory-specific supervision via explicitly penalizing summaries that induce high epistemic uncertainty. Experiments show that MMPO consistently outperforms existing methods on diverse long-horizon tasks, maintaining 97.1% performance even when scaled to 1.75M-token contexts.

1 Citations
0 Influential
4 Altmetric
21.0 Score
Original PDF

No Analysis Report Yet

This paper hasn't been analyzed by Gemini yet.

Log in to request an AI analysis.

댓글

댓글을 작성하려면 로그인하세요.

아직 댓글이 없습니다. 첫 번째 댓글을 남겨보세요!