arXiv：cs.AI· Makoto Fukushima, Hua-Dong Xiong, Ehsan Moradi Pari·· 24 天前
人类与 AI 协作中的「约定鸿沟」：Hanabi 中的隐含沟通诊断指标
The Convention Gap: Beyond Literal Content in Human-AI Cooperation
AI 导读
研究者提出「约定鸿沟」指标，用 Hanabi 游戏中按提示字面内容预测的失败概率减去实际出牌失败率，来衡量伙伴选择是否传递了字面信息之外的含义。
正文 · 原文
Abstract:Cooperative AI agents are often evaluated against other AIs, yet human cooperation is thought to rely on implicit conventions: shared rules for reading meaning beyond the literal message. We propose the convention gap as a diagnostic of such implicit communication. In the card game Hanabi, where hints are the only messages about one's own hand, the convention gap is the failure probability predicted exactly from the literal content of the hints and the visible cards, minus the observed failure rate of the cards played. We prove that the population gap (the value the gap approaches over many games) is zero when the partner's choices convey nothing beyond their literal content, given a fair shuffle and a player who uses only what was said and seen. A nonzero population gap thus indicates that partner choices conveyed more; it does not measure how much. The gap was +26.2 percentage points (pp) in public human-human games (+46 pp on plays of cards that no hint had pointed to), +16.4 pp for human plays with AI partners, and -0.6 pp in AI-AI games we simulated among rule-based agents. With three AI agents of a published human-AI study, the literal failure probability of the cards humans played was similar across partners (38-41%), yet human failure ranged from 14.4% to 34.4% and the gap from +24.1 to +6.2 pp, and the contrast between partners persisted within participants. Across five levels of Off-Belief Learning agents, trained so that the first level reads past actions only for their literal content and each higher level interprets more, the gap rose monotonically from +1.6 to +21.7 pp. These results argue for reporting the convention gap on human-partnered play, not only AI-AI performance, when evaluating agents intended for human partners.
| Subjects: | Artificial Intelligence (cs.AI); Human-Computer Interaction (cs.HC) |
| Cite as: | arXiv:2609.11489 [cs.AI] |
| (or arXiv:2609.11489v4 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2609.11489 arXiv-issued DOI via DataCite |
Submission history
From: Makoto Fukushima [view email]
[v1]
Thu, 10 Sep 2026 12:57:03 UTC (1,220 KB)
[v2]
Fri, 11 Sep 2026 05:18:58 UTC (1,220 KB)
[v3]
Thu, 24 Sep 2026 01:23:10 UTC (1,220 KB)
[v4]
Fri, 9 Oct 2026 08:02:20 UTC (1,325 KB)
来源：arXiv：cs.AI · arxiv.org