arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2610.20195cs.CLcs.AIcs.SI
收藏

是什么让社交智能体的交互有效?从配对世界中的可验证结果中学习

What Makes a Social Agent's Interactions Effective? Learning from Verifiable Outcomes in Paired Worlds

Alibaba Group(阿里巴巴集团);The University of Tokyo(东京大学)

Caixin Kang, Ke Wang, Yuchuan Wu, Xiaoqiang Wang, Ling Li, Bingxue Qiu, Yongbin Li, Yoichi Sato

AI总结:

该研究针对现有社交智能体评估未检查计划的问题,构建HICO-Bench基准,提出PaCPO方法,将Qwen3-8B的可验证成功率从20.4%提升至28.3%。

摘要:

有效的社交交互不仅需要流畅的回应:智能体必须达成符合参与者需求和情况的计划。现有对社交智能体的评估主要依赖对对话进行评分的LLM评判器,而未检查最终达成的计划。我们推出HICO-Bench,这是一个包含180个日常协调场景的基准,例如资源分配、日程安排和授权,其中两个拥有私人信息的智能体进行对话,然后各自记录他们达成的计划。一个程序检查可行性(两份记录匹配且满足双方的私人条件),一个LLM评判器检查完整性(对话支持记录的计划);同时通过两项即为可验证成功。每个场景有两个版本,共享相同的公共上下文但在一个私人条件上不同,因此正确计划也不同。在13个LLM中,36%被评为目标完成度高的对话未通过可验证成功。为了从仅在对话结束时出现的结果中学习,我们提出配对信用策略优化(Paired-Credit Policy Optimization,PaCPO)。在每一轮之后,PaCPO要求智能体给出一个暂定计划,在场景的两个版本中对其进行检查,并对暂定计划符合各自版本的轮次给予奖励。PaCPO将Qwen3-8B的可验证成功率从20.4%提升至28.3%。

英文摘要:

Effective social interaction requires more than fluent responses: agents must reach plans that fit the participants' needs and circumstances. Existing evaluations of social agents mainly rely on LLM judges that rate the conversation, without checking the resulting plan. We introduce HICO-Bench, a benchmark of 180 everyday coordination scenarios, such as resource allocation, scheduling, and authorization, in which two agents with private information converse and then independently record the plan they agreed on. A program checks Feasibility (the two records match and satisfy both participants' private conditions), and an LLM judge checks Integrity (the conversation supports the recorded plan); passing both is a Verified Success. Each scenario has two versions that share the same public context but differ in one private condition, so the correct plan differs between them. Across thirteen LLMs, 36% of conversations rated highly for goal completion fail Verified Success. To learn from outcomes that arrive only at the end of a conversation, we propose Paired-Credit Policy Optimization (PaCPO). After each turn, PaCPO asks the agent for a tentative plan, checks it in both versions of the scenario, and rewards the turns after which the tentative plans fit their respective versions. PaCPO raises Qwen3-8B's Verified Success from 20.4% to 28.3%.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑