Chain-of-Action：基于大语言模型的忠实多模态问答
Chain-of-Action: Faithful and Multimodal Question Answering through Large Language Models
AI总结：
本研究提出Chain-of-Action（CoA）多模态检索增强问答框架，通过推理-检索机制分解复杂问题、设计即插即用动作检索异构信息、提出多参考忠实度分数校验答案，解决现有QA的幻觉与组合推理弱问题，性能优于同类方法。
摘要：
我们提出了一种用于多模态检索增强问答（QA）的Chain-of-Action（CoA，动作链）框架。与现有研究相比，CoA克服了当前QA应用的两大核心挑战：一是与实时或领域事实不符的不忠实幻觉问题，二是对组合信息的推理能力较弱的问题。我们的核心贡献是提出了一种新颖的推理-检索机制，该机制通过系统化提示词工程和预设动作，将复杂问题分解为推理链。在方法层面，我们设计了三类可适配领域的「即插即用」动作，用于从异构数据源中检索实时信息；同时提出了多参考忠实度分数（MRFS），用于验证并解决答案中的冲突。在实验层面，我们结合公开基准数据集与Web3案例研究，验证了CoA的性能优于其他方法。
英文摘要：
We present a Chain-of-Action (CoA) framework for multimodal and retrieval-augmented Question-Answering (QA). Compared to the literature, CoA overcomes two major challenges of current QA applications: (i) unfaithful hallucination that is inconsistent with real-time or domain facts and (ii) weak reasoning performance over compositional information. Our key contribution is a novel reasoning-retrieval mechanism that decomposes a complex question into a reasoning chain via systematic prompting and pre-designed actions. Methodologically, we propose three types of domain-adaptable `Plug-and-Play' actions for retrieving real-time information from heterogeneous sources. We also propose a multi-reference faith score (MRFS) to verify and resolve conflicts in the answers. Empirically, we exploit both public benchmarks and a Web3 case study to demonstrate the capability of CoA over other methods.
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。
