MemPO: Self-Memory Policy Optimization for Long-Horizon Agents
This addresses memory inefficiency in long-horizon agents, offering a novel method for improving performance and reducing computational costs.
The paper tackles the challenge of growing context size degrading performance in long-horizon agents by proposing MemPO, a self-memory policy optimization algorithm that enables autonomous memory management, achieving absolute F1 score gains of 25.98% over the base model and 7.1% over previous SOTA while reducing token usage by 67.58% and 73.12%.
Long-horizon agents face the challenge of growing context size during interaction with environment, which degrades the performance and stability. Existing methods typically introduce the external memory module and look up the relevant information from the stored memory, which prevents the model itself from proactively managing its memory content and aligning with the agent's overarching task objectives. To address these limitations, we propose the self-memory policy optimization algorithm (MemPO), which enables the agent (policy model) to autonomously summarize and manage their memory during interaction with environment. By improving the credit assignment mechanism based on memory effectiveness, the policy model can selectively retain crucial information, significantly reducing token consumption while preserving task performance. Extensive experiments and analyses confirm that MemPO achieves absolute F1 score gains of 25.98% over the base model and 7.1% over the previous SOTA baseline, while reducing token usage by 67.58% and 73.12%.