Exploratory Memory-Augmented LLM Agent via Hybrid On- and Off-Policy Optimization
This work is significant for researchers developing more exploratory and generalizable LLM-based agents, as it tackles the challenge of discovering novel states in complex environments.
The paper addresses the exploration bottleneck in LLM agents by proposing EMPO², a hybrid RL framework that uses memory for exploration and combines on- and off-policy updates. This approach led to significant performance improvements of 128.6% on ScienceWorld and 11.3% on WebShop compared to GRPO.
Exploration remains the key bottleneck for large language model agents trained with reinforcement learning. While prior methods exploit pretrained knowledge, they fail in environments requiring the discovery of novel states. We propose Exploratory Memory-Augmented On- and Off-Policy Optimization (EMPO$^2$), a hybrid RL framework that leverages memory for exploration and combines on- and off-policy updates to make LLMs perform well with memory while also ensuring robustness without it. On ScienceWorld and WebShop, EMPO$^2$ achieves 128.6% and 11.3% improvements over GRPO, respectively. Moreover, in out-of-distribution tests, EMPO$^2$ demonstrates superior adaptability to new tasks, requiring only a few trials with memory and no parameter updates. These results highlight EMPO$^2$ as a promising framework for building more exploratory and generalizable LLM-based agents.