Joint Learning of Experiential Rules and Policies for Large Language Model Agents
For LLM agents in multi-step interactive environments, a key challenge is to make effective use of accumulated interaction experience. Existing work has typically separated two uses of such experience: keeping it outside the model as natural-language rules for later prompting, or using trajectories and feedback to update the model parameters. The former is easy to interpret but can fall out of sync with the evolving policy; the latter improves the policy more broadly but provides only limited correction for local mistakes in sparse-reward settings. We present Joint Learning of Experiential Rules and Policies for LLM Agents (JERP), which updates a long-term experiential-rule pool and the policy from the same interaction trajectories. At decision time, JERP retrieves task-relevant rules and conditions the agent on them together with the interaction history. After each episode, it uses the collected trajectories both to optimize the policy and to revise the rule pool by comparing current rollouts with reference successful trajectories. This coupling keeps the rule pool aligned with the evolving policy while allowing stable and effective behaviors to be gradually absorbed into the model itself. Experiments on AlfWorld and WebShop show that JERP yields consistent gains in decision performance for complex interactive tasks.
Code (0)
등록된 구현이 없습니다.
Similar Papers 제목 키워드 기반
$\mathcal{X}$-KD: General Experiential Knowledge Distillation for Large Language Models
Knowledge Distillation (KD) for Large Language Models (LLMs) has become increasingly important as models grow in size and complexity. While existing distillation approaches focus on imitating teacher behavior, they often…
Knowledge DistillationReinforcement LearningArithmetic ReasoningMachine TranslationPolishing Every Facet of the GEM: Testing Linguistic Competence of LLMs and Humans in Korean
We introduce the $\underline{Ko}rean \underline{G}rammar \underline{E}valuation Bench\underline{M}ark (KoGEM)$, designed to assess the linguistic competence of LLMs and humans in Korean. KoGEM consists of 1.5k multiple-c…
Multiple-choiceNotes to Self: Can LLMs Benefit from Experiential Abstractions?
Humans distill experience into reusable abstractions, e.g., strategies and cautionary reminders, and apply them to gradually solve problems more effectively. We study whether Large Language Models (LLMs) can similarly be…
Reinforcement LearningLogical ReasoningOnline Experiential Learning for Language Models
The prevailing paradigm for improving large language models relies on offline training with human annotations or simulated environments, leaving the rich experience accumulated during real-world deployment entirely unexp…
A General “Power-of-d” Dispatching Framework for Heterogeneous Systems
Intelligent dispatching is crucial to obtaining low response times in large-scale systems. One common scalable dispatching paradigm is the “power-of-d,” in which the dispatcher queries d servers at…