Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2504.15257 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916699780939776 |
|---|---|
| author | Gao, Hongcheng Liu, Yue He, Yufei Dou, Longxu Du, Chao Deng, Zhijie Hooi, Bryan Lin, Min Pang, Tianyu |
| author_facet | Gao, Hongcheng Liu, Yue He, Yufei Dou, Longxu Du, Chao Deng, Zhijie Hooi, Bryan Lin, Min Pang, Tianyu |
| contents | This paper proposes a query-level meta-agent named FlowReasoner to automate the design of query-level multi-agent systems, i.e., one system per user query. Our core idea is to incentivize a reasoning-based meta-agent via external execution feedback. Concretely, by distilling DeepSeek R1, we first endow the basic reasoning ability regarding the generation of multi-agent systems to FlowReasoner. Then, we further enhance it via reinforcement learning (RL) with external execution feedback. A multi-purpose reward is designed to guide the RL training from aspects of performance, complexity, and efficiency. In this manner, FlowReasoner is enabled to generate a personalized multi-agent system for each user query via deliberative reasoning. Experiments on both engineering and competition code benchmarks demonstrate the superiority of FlowReasoner. Remarkably, it surpasses o1-mini by 10.52% accuracy across three benchmarks. The code is available at https://github.com/sail-sg/FlowReasoner. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_15257 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | FlowReasoner: Reinforcing Query-Level Meta-Agents Gao, Hongcheng Liu, Yue He, Yufei Dou, Longxu Du, Chao Deng, Zhijie Hooi, Bryan Lin, Min Pang, Tianyu Artificial Intelligence This paper proposes a query-level meta-agent named FlowReasoner to automate the design of query-level multi-agent systems, i.e., one system per user query. Our core idea is to incentivize a reasoning-based meta-agent via external execution feedback. Concretely, by distilling DeepSeek R1, we first endow the basic reasoning ability regarding the generation of multi-agent systems to FlowReasoner. Then, we further enhance it via reinforcement learning (RL) with external execution feedback. A multi-purpose reward is designed to guide the RL training from aspects of performance, complexity, and efficiency. In this manner, FlowReasoner is enabled to generate a personalized multi-agent system for each user query via deliberative reasoning. Experiments on both engineering and competition code benchmarks demonstrate the superiority of FlowReasoner. Remarkably, it surpasses o1-mini by 10.52% accuracy across three benchmarks. The code is available at https://github.com/sail-sg/FlowReasoner. |
| title | FlowReasoner: Reinforcing Query-Level Meta-Agents |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2504.15257 |