Skill transfer in multi-agent teams reduces training cost and boosts task performance
Evo2Team: When Do Evolved Skills Transfer? From Selection to Deployment
Artificial Intelligence
Summary
Skills that work well in one team of agents don't always help another team perform better. The authors studied how transferring skills for routing and communication between teams can reduce the effort to train new teams. Their method Evo2Team selects, adapts, and tests transferred skills to save time and reduce costs. The study shows that just choosing skills isn't enough; evaluating how agents actually use them and their impact on tasks is essential.
What this means in practice
- •For multi-agent system engineers: Cut training time and computational cost by selectively reusing evolved skills when deploying teams for tasks like communication and routing.
- •For robotics development teams: Improve multi-robot team coordination by adapting and confirming transferred behavioral rules rather than evolving from scratch.
Authors
Renxiang Wang, Jiaming Cui
Abstract
A skill bank that helps one multi-agent system may leave another's behavior unchanged. A transferred rule helps only when target agents act on it successfully. We study this path for routing and communication skills in Count-Frequency and AgentsNet, using teams of 4--32 agents and GPT and Qwen model ladders. Source evolution meets a joint quality, cost, model-tier, and confirmation goal in 14 of 16 settings. We then evaluate Evo2Team, which selects, adapts, and confirms source skills for the target team, alongside six frozen selectors across 28 transfer directions. Evo2Team's target-side exploration cost is below that of evolving a new target bank in every direction, even when reused reference evaluations are charged once. Twenty of 28 held-out outcomes meet the positive-transfer criterion, including three saved diagnostic tests. Selection alone does not explain these outcomes: KNN and CORAL choose different banks in two AgentsNet directions but produce identical recorded executions. When Evo2Team changes execution, gains can reach many tasks, as in a Count-Frequency direction that improves 28 of 32 tasks over KNN. Seven positive AgentsNet outcomes save 6.1--14.6\% in deployment cost while using transferred skills on only three to six of fifteen tasks. In five earlier accepted directions, all 22 task records using transferred skills pass three fixed-graph confirmations, but four fail in recorded executions on new graphs. Graphs and model responses change together in this comparison. These results show that skill transfer must be assessed through the actions agents take, the tasks those actions reach, and the quality and cost of the final deployment.