More Victories, Less Cooperation: Assessing Cicero's Diplomacy Play
Wichayaporn Wongkamjan, Feng Gu, Yanze Wang, Ulf Hermjakob, Jonathan May, Brandon M. Stewart, Jonathan K. Kummerfeld, Denis Peskoff, Jordan L. Boyd-Graber
摘要
The boardgame Diplomacy is a challenging setting for communicative and cooperative artificial intelligence. The most prominent communicative Diplomacy AI, Cicero, has excellent strategic abilities, exceeding human players. However, the best Diplomacy players master communication, not just tactics, which is why the game has received attention as an AI challenge. This work seeks to understand the degree to which Cicero succeeds at communication. First, we annotate in-game communication with abstract meaning representation to separate in-game tactics from general language. Second, we run two dozen games with humans and Cicero, totaling over 200 humanplayer hours of competition. While AI can consistently outplay human players, AI-Human communication is still limited because of AI's difficulty with deception and persuasion. This shows that Cicero relies on strategy and has not yet reached the full promise of communicative and cooperative AI.
问问这篇 Paper
智能体会读完全文。
Lune 把这篇 Paper 索引到了每一个公式,引用它的顶会 Paper 也一样。你提问,回答直接引用原文。
引用它的顶会 Paper5
- CoopEval: Benchmarking Cooperation-Sustaining Mechanisms and LLM Agents in Social DilemmasEmanuel Tewolde, Xiao Zhang, David Guzman Piedrahita, Vincent Conitzer 等ICML 2026 · 被引用 15 次
- Explaining Decisions of Agents in Mixed-Motive GamesMaayan Orner, Oleg Maksimov, Akiva Kleinerman, Charles Ortiz 等AAAI 2025 · 被引用 4 次
- NewsInterview: a Dataset and a Playground to Evaluate LLMs' Grounding Gap via Informational InterviewsAlexander Spangher, Michael Lu, Sriya Kalyan, Hyundong Justin Cho 等ACL 2025 · 被引用 2 次
- DipLLM: Fine-Tuning LLM for Strategic Decision-making in DiplomacyKaixuan Xu, Jiajun Chai, Sicheng Li, Yuqian Fu 等ICML 2025
- Mental Model-based Generation of Lies for Insider Threat ModelingBrittany Cates, Sarath SreedharanAAAI 2026
它引用的顶会 Paper14
- Language Models are Few-Shot LearnersTom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah 等NeurIPS 2020 · 被引用 64,255 次
- Chain-of-Thought Prompting Elicits Reasoning in Large Language ModelsJason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma 等NeurIPS 2022 · 被引用 22,562 次
- Inference-Time Intervention: Eliciting Truthful Answers from a Language ModelKenneth Li, Oam Patel, Fernanda B. Viégas, Hanspeter Pfister 等NeurIPS 2023 · 被引用 1,549 次
- AI Chains: Transparent and Controllable Human-AI Interaction by Chaining Large Language Model PromptsTongshuang Wu, Michael Terry, Carrie Jun CaiCHI 2022 · 被引用 465 次
- Formalizing and Benchmarking Prompt Injection Attacks and DefensesYupei Liu, Yuqi Jia, Runpeng Geng, Jinyuan Jia 等USENIX Security 2024 · 被引用 308 次
相关 Paper
- Human-Level Performance in No-Press Diplomacy via Equilibrium SearchJonathan Gray, Adam Lerer, Anton Bakhtin, Noam BrownICLR 2021 · 被引用 61 次
- Learning to Play No-Press Diplomacy with Best Response Policy IterationThomas W. Anthony, Tom Eccles, Andrea Tacchetti, János Kramár 等NeurIPS 2020 · 被引用 50 次
- It Takes Two to Negotiate: Modeling Social Exchange in Online Multiplayer GamesKokil Jaidka, Hansin Ahuja, Lynnette Hui Xian NgCSCW 2024 · 被引用 9 次
- The Hidden Rules of Hanabi: How Humans Outperform AI AgentsMatthew Sidji, Wally Smith, Melissa J. RogersonCHI 2023 · 被引用 9 次
- It Takes Two to Lie: One to Lie, and One to ListenDenis Peskov, Benny Cheng, Ahmed Elgohary, Joe Barrow 等ACL 2020 · 被引用 28 次
