CausalGPT: Illuminating Faithfulness and Causality for Knowledge Reasoning with Foundation Models
Ziyi Tang, Ruilin Wang, Weixing Chen, Yongsen Zheng, Zechuan Chen, Yang Liu, Keze Wang, Tianshui Chen, Liang Lin Ziyi Tang, Ruilin Wang, Weixing Chen, Yongsen Zheng, Zechuan Chen, Yang Liu, Keze Wang, and Liang Lin are with the Sun Yat-sen University, Guangzhou, Guangdong 510275, China (email:
Abstract
Despite the progress of foundation models, knowledge-based reasoning remains a persistent challenge due to their limited capacity for knowledge recall and inference. Existing methods primarily focus on encouraging these models to plan and solve problems or extensively sample reasoning chains independently. However, these methods often overlook conceptual errors and inferential fallacies, inevitably leading to a series of notorious issues such as misleading conclusions, cognitive biases, and reduced decision quality. While explicit modeling of causality is argued to hold promise in addressing these issues, contemporary research efforts have thus far fallen short in achieving causality-based foundation models. Drawing inspiration from the orchestration of diverse specialized agents collaborating to tackle intricate tasks, we propose a framework named Causal-Consistency Chain-of-Thought (CaCo-CoT) that harnesses multi-agent collaboration to bolster the faithfulness and causality of foundation models, involving a set of reasoners and evaluators. These agents collaboratively work within a reasoning-and-consensus paradigm to improve faithfulness. The reasoners are tasked with generating
中文速览
大型基础模型（foundation model）在做知识密集型推理时，常常因为记错事实或推理链出现逻辑跳跃而产生误导性结论。为此，研究者提出了一个名为"因果一致性思维链"（Causal-Consistency Chain-of-Thought，CaCo-CoT）的多智能体协作框架：由一组"忠实推理者"（reasoner）模拟人类因果思维，依次完成概念解释、子问题拆解和子问题回答，再由"因果评估者"（evaluator）从反向非因果视角和反事实视角逐一核查推理链的因果一致性，两类智能体反复协商直到达成共识才输出最终答案。在文本和多模态知识推理基准（包括科学问答和常识推理）上的大量实验表明，该框架显著优于现有最优方法。这项工作是首次将因果一致性原则引入多智能体协作推理体系，为提升基础模型推理的可信度和准确性提供了一条新路径。
原文 arXiv:2308.11914;中英对照 + 大白话阅读 https://aha.fim.ai/paper/2308.11914v4