MEML-GRPO: Heterogeneous Multi-Expert Mutual Learning for RLVR Advancement
Weitao Jia, Jinghui Lu, Haiyang Yu, Siqi Wang, Guozhi Tang, An-Lan Wang, Weijie Yin, Dingkang Yang
Abstract
Recent advances demonstrate that reinforcement learning with verifiable rewards (RLVR) significantly enhances the reasoning capabilities of large language models (LLMs). However, standard RLVR faces challenges with reward sparsity, where zero rewards from consistently incorrect candidate answers provide no learning signal, particularly in challenging tasks. To address this,we propose Multi-Expert Mutual Learning GRPO (MEML-GRPO), an innovative framework that utilizes diverse expert prompts as system prompts to generate a broader range of responses, substantially increasing the likelihood of identifying correct solutions. Additionally, we introduce an inter-expert mutual learning mechanism that facilitates knowledge sharing and transfer among experts, further boosting the model’s performance through RLVR. Extensive experiments across multiple reasoning benchmarks show that MEML-GRPO delivers significant improvements, achieving an average performance gain of 4.89% with Qwen and 11.33% with Llama, effectively overcoming the core limitations of traditional RLVR methods.
BibTeX
@inproceedings{aaai2026_memlgrpoheteroge,
title = {MEML-GRPO: Heterogeneous Multi-Expert Mutual Learning for RLVR Advancement},
author = {Weitao Jia and Jinghui Lu and Haiyang Yu and Siqi Wang and Guozhi Tang and An-Lan Wang and Weijie Yin and Dingkang Yang and Yuxiang Nie and Bin Shan and Hao Feng and Irene Li and Kun Yang and Han Wang and Jingqun Tang and Teng Fu and Changhong Jin and Chao Feng and Xiaohui Lv and Can Huang},
booktitle = {AAAI 2026},
year = {2026}
}