Lessons Learned: A Multi-Agent Framework for Code LLMs to Learn and Improve

TL;DR

This paper introduces a lesson-based multi-agent framework for code LLMs that enables agents to learn from each other's successes and failures, improving overall performance beyond larger models or existing collaboration methods.

Contribution

It proposes a novel lesson-based collaboration framework and mechanisms for LLM agents to learn and share knowledge, enhancing their collective problem-solving abilities.

Findings

Small LLMs with lessons outperform larger LLMs.

The framework improves code optimization performance.

Lessons enable effective knowledge transfer among agents.

Abstract

Recent studies show that LLMs possess different skills and specialize in different tasks. In fact, we observe that their varied performance occur in several levels of granularity. For example, in the code optimization task, code LLMs excel at different optimization categories and no one dominates others. This observation prompts the question of how one leverages multiple LLM agents to solve a coding problem without knowing their complementary strengths a priori. We argue that a team of agents can learn from each other's successes and failures so as to improve their own performance. Thus, a lesson is the knowledge produced by an agent and passed on to other agents in the collective solution process. We propose a lesson-based collaboration framework, design the lesson solicitation--banking--selection mechanism, and demonstrate that a team of small LLMs with lessons learned can outperform a much larger LLM and other multi-LLM collaboration methods.