Algorithmic game theory, mechanism design
Emanuel Tewolde, Xiao Zhang, David Guzman Piedrahita et al.
For AI safety researchers, it provides the first comparative evaluation of cooperation mechanisms for LLM agents, addressing the critical problem of LLMs' tendency to defect in mixed-motive games.
Mingyang Liu, Yongshan Chen, Zhiyuan Fan et al.
This addresses privacy and human-in-the-loop constraints in sequential decision-making, though it is incremental as it builds on existing online learning frameworks with a new feedback type.
Ander Artola Velasco, Stratis Tsirtsis, Manuel Gomez-Rodriguez
For regulators and product developers in high-stakes domains like drug development, this work addresses the inefficiency of costly trials that may deter socially valuable 'moonshot' products.
Eilam Shapira, Moshe Tennenholtz, Roi Reichart
This reveals a trade-off between optimizing models for human use and using them as proxies for human behavior, which is important for researchers and practitioners in AI alignment and behavioral modeling.
Jiayuan Liu, Tianqin Li, Shiyi Du et al.
For researchers and developers of multi-agent LLM systems, this reveals that longer memory can destabilize cooperation, challenging the assumption that expanded context is always beneficial.
Yan Fang, Jialin Chen, Chun Gan et al.
This work is significant for advertisers and LLM platforms seeking to optimize revenue from native advertising by dynamically determining the best moment to insert an ad within a conversation.
Lingjiao Chen, Chi Zhang, Yeye He et al.
This reveals a critical issue for developers and consumers in selecting models based on cost, highlighting the need for better cost transparency and monitoring.
Hongze Liu, Chang Guo, Yingzeng Li et al.
This addresses the problem of scalable and efficient routing in open agentic web ecosystems, offering a novel co-design approach that is domain-specific to LLM inference systems.
Gonzalo Ballestero, Hadi Hosseini, Samarth Khanna et al.
This addresses coordination challenges in multi-agent AI systems, providing experimental evidence on LLM behavior, but is incremental in extending human studies to AI agents.
Sahaj Singh Maini, Robert L. Goldstone, Zoran Tiganj
This work identifies a coordination gap between humans and LLMs, providing a diagnostic for improving AI in group tasks.
Xuan Qiu
This work addresses the design of healthier online social environments and forecasting collective behavior in AI societies, representing a novel method for a known bottleneck rather than an incremental improvement.
Mehryar Mohri, Clayton Sanford, Jon Schneider et al.
This addresses the challenge of integrating language models into decision-making systems, with implications for AI safety and optimization, though it is incremental in bridging theoretical and practical aspects.
Yanan Cai, Zhongrui Zhao, Zhigang Lu et al.
This survey provides a structured overview for AI alignment researchers, highlighting where game theory can genuinely improve the robustness and adaptability of AI systems.
Ioannis Anagnostides, Constantinos Daskalakis, Gabriele Farina et al.
For game theorists and computer scientists, this work provides the first strong evidence of intractability for correlated equilibria in concave games, while also offering tractable relaxations and algorithms for important subclasses.
Gabriele Farina, Juan Carlos Perdomo
For researchers in online learning and algorithmic fairness, this work unifies and simplifies existing algorithms while enabling new applications in challenging environments and improving regret minimization techniques.
Boning Li, Longbo Huang
This work provides a method for automatically constructing effective abstractions for large imperfect-information games, which is crucial for scaling game-solving algorithms, especially for less-studied games where domain-specific evaluators are unavailable.
Mehryar Mohri, Jon Schneider, Yifan Wu
This work provides a principled framework for optimizing language models based on answer correctness, unifying existing approaches and enabling efficient algorithms.
Keegan Harris, Siddharth Prasad, Asher Trockman
This work addresses the problem of fair credit assignment for AI-generated content, which is important for creators and platforms, but the results are incremental as they apply existing game theory concepts with algorithmic improvements.
Xuanqiang Angelo Huang, Charlie Tharas, Samuele Marro et al.
For AI safety researchers, it demonstrates that ensuring cooperative AI requires not just good rules but also intrinsically prosocial agents.
Eden Saig, Tamar Garbuz, Ariel D. Procaccia et al.
This work addresses cost-effectiveness for organizations delegating AI tasks, offering an incremental improvement in contract design to balance evaluation accuracy and expenses.