Agent Control Flash News·· 3 小时前AI 评分40
阿里 MultiAgent-OPD：层级MAS多智能体系统训练新方法，macro avg@4达38.68%
层级 MAS 别只给 Searcher 结局奖励：阿里 MultiAgent-OPD，macro avg@4 38.68%；outcome-only 检索塌 0
Edge-TTS 广播级有声
全员无门槛畅听
点击合成广播级音频
00:00
00:00
AI 导读
阿里巴巴提出MultiAgent-OPD，结合策略蒸馏与强化学习，实现层级多智能体系统（MAS）的端到端联合训练。在七组数据集上，该系统macro avg@4指标达到38.68%，超越Prompted Critic和C-MAS + RL。
正文
Abstract:In a hierarchical multi-agent system, a Leader assigns tasks and subagents carry them out. Jointly training the Leader and the subagents is difficult because a final answer gives only indirect feedback to the subagents. An LLM judge can score the performance of the subagents, but these scores depend on the judge's own assessment of the quality of subagents' responses. We present MultiAgent-OPD, which combines on-policy distillation with reinforcement learning to jointly train a shared-parameter Leader-Searcher-Answerer team end to end. We first train a Teacher in a Searcher-Answerer workflow using answer correctness. During team training, the frozen Teacher scores student-generated search and summary tokens to provide local process supervision for the Searchers, while the Leader learns from team success and the Answerer from candidate correctness. Across seven datasets, the trained system reaches 38.68% macro avg@4, exceeding Prompted Critic by 4.06 percentage points and C-MAS + RL, the trained cascaded multiagent system, by 2.61 percentage points. The team also solves more questions within four attempts.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.23234 [cs.CL] |
| (or arXiv:2610.23234v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.23234 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yuhan Sun [view email]
[v1]
Fri, 16 Oct 2026 10:01:52 UTC (799 KB)
来源：Agent Control Flash News · arxiv.org
💡量子基态深度拆解 · 关联专栏与产品
原创出品站长/Quantara 团队·约 3 分钟深度解析
Cursor Agent 架构全景拆解：从影子工作区到多文件推演引擎
全面剖析 Cursor 在 Agent 模式下的底层技术实现，解析影子分支隔离、LSP 上下文剪枝以及预测性补全的核心架构细节与工程实践经验。
深入理解底层架构与行业变量阅读完整拆解 →
精选收录Vercel Changelog·约 1 分钟深度解析
Vercel v0 推出全新多 Agent 协作画布与实时全栈部署预览
动态速评：v0 正在从单一组件生成器演进为全栈应用搭建协作台，尤其是与 Neon/Postgres 的直连极大降低了全栈原型验证成本。
深入理解底层架构与行业变量阅读完整拆解 →