标题： 迈向语言和多模态智能体的合理性：综述 显示英文标题

标题： Towards Rationality in Language and Multimodal Agents: A Survey

摘要： 这项工作讨论了如何构建更合理的语言和多模态智能体以及定义智能系统合理性的标准。 合理性是指受理性引导的品质，其特征是与证据和逻辑原则一致的决策。 它在可靠的问题解决中起着关键作用，确保解决方案有坚实的基础并且一致。 尽管取得了进展，大型语言模型（LLMs）由于其有限的知识空间和不一致的输出，常常无法达到合理性。 作为回应，最近的努力转向开发多模态和多智能体系统，以及集成外部工具、编程代码、符号推理器、效用函数和保真风险控制模块，而不是仅仅依赖单一的LLM进行决策。 本文综述了语言和多模态智能体的最新进展，评估了它们在增强合理性方面的作用，并概述了开放挑战和未来研究方向。 我们在https://github.com/bowen-upenn/Agent_Rationality维护了一个开源仓库。 显示英文摘要

摘要： This work discusses how to build more rational language and multimodal agents and what criteria define rationality in intelligent systems. Rationality is the quality of being guided by reason, characterized by decision-making that aligns with evidence and logical principles. It plays a crucial role in reliable problem-solving by ensuring well-grounded and consistent solutions. Despite their progress, large language models (LLMs) often fall short of rationality due to their bounded knowledge space and inconsistent outputs. In response, recent efforts have shifted toward developing multimodal and multi-agent systems, as well as integrating modules like external tools, programming codes, symbolic reasoners, utility function, and conformal risk controls rather than relying solely on a single LLM for decision-making. This paper surveys state-of-the-art advancements in language and multimodal agents, assesses their role in enhancing rationality, and outlines open challenges and future research directions. We maintain an open repository at https://github.com/bowen-upenn/Agent_Rationality.