The Year in AI Papers
The year/Independent research

Paper 2606.10917

Role-Agent: Bootstrapping LLM Agents via Dual-Role Evolution

Xucong Wang, Ziyu Ma, Shidong Yang, Tongwen Huang, Pengkun Wang, Yong Wang, Xiangxiang Chu

Published
Jun 2026
Research lab
Independent
Citations
0
GitHub
79 stars

01 In brief

Summary

Role-Agent is a framework that enables bootstrapped agent-environment co-evolution using a single LLM in dual roles.

It consists of World-In-Agent (WIA), where the LLM predicts future states after actions, using the alignment between predicted and actual states as a process reward to encourage environment-aware reasoning, and Agent-In-World (AIW), where the LLM analyzes failure modes from failed trajectories and retrieves tasks with similar patterns to reshape the training data distribution.

Experiments on ALFWorld, WebShop, and search-augmented QA benchmarks show consistent improvements, with an average gain of over 4% over strong baselines like GiGPO.

Ablations confirm both components contribute, and sensitivity analysis shows optimal settings for the advantage scaling coefficient and prediction horizon.

The framework adds only about 5.2% extra computation.

02 From the paper

Abstract

Although Large Language Model (LLM) agents have demonstrated strong performance on complex tasks, their learning is often limited by inefficient interaction feedback and static training environments, which hinder broader generalization. To address these limitations, this paper introduces Role-Agent, \textcolor{black}{a framework} that harnesses a single LLM to function concurrently as both the agent and the environment, enabling a bootstrapped co-evolution. Role-Agent comprises two synergistic components: World-In-Agent (WIA) and Agent-In-World (AIW). In WIA, the LLM acts as the agent and predicts future states after each action; the alignment between predicted and actual states is then used as a process reward, encouraging environment-aware reasoning. In AIW, the LLM analyzes failure modes from failed trajectories and retrieves tasks with similar failure patterns, thereby reshaping the training data distribution for targeted practice. Experiments on multiple benchmarks show that Role-Agent consistently improves performance, yielding an average gain of over 4\% over strong baselines.