Beyond Frameworks: Unpacking Collaboration Strategies in Multi-Agent Systems
Haochun Wang, Sendong Zhao, Jingbo Wang, Zewen Qiang, Bing Qin, Ting Liu · 2025
Read Beyond Frameworks: Unpacking Collaboration Strategies in Multi-Agent Systems by Haochun Wang, Sendong Zhao, Jingbo Wang, Zewen Qiang, Bing Qin, Ting Liu (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2505-12467.txt.
Description
Haochun Wang and five co-authors tested in 2025 how teams of AI agents should work together, beyond the frameworks that hold them. Multi-agent setups now tackle complex tasks, but most research tweaks the outer frame rather than the day-to-day rules of teamwork. The team studied four such rules: who leads, who takes part, how turns flow, and how past talk is kept. They ran tests in two settings that stress evidence handling.
Central leadership, teacher-led turns, ordered exchanges, and teacher-made summaries of past talk gave the best mix of accuracy and cost. The team also built a score that weighs accuracy against token use. Their message is that smart interaction rules matter as much as system shape. Designers of agent teams can use the findings to cut waste while keeping quality.
Catalog record
| CREATOR | Haochun Wang, Sendong Zhao, Jingbo Wang, Zewen Qiang, Bing Qin, Ting Liu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 50 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2505-12467