从因果视角审视缓解知识 grounded 对话生成中的幻觉问题
A Cause-Effect Look at Alleviating Hallucination of Knowledge-grounded Dialogue Generation
AI总结：
本文用反事实推理分析知识 grounded 对话中知识噪声导致幻觉的因果机制，并提出利用对话—知识交互的轻量缓解方法，在不影响对话性能且适配多模型的前提下减少幻觉。
摘要：
在大规模预训练语言模型的赋能下，现有对话系统已在进行流畅且听起来自然的对话方面展现出令人瞩目的性能。然而，它们仍受到幻觉问题的困扰，导致生成回复中出现不可预测的事实性错误。近来，知识 grounded 对话生成模型会有意调用外部知识资源以生成信息量更高的回复，这类模型也被证明能够有效减少幻觉。沿着获取高质量知识这一思路，已有若干工作在该问题上取得了相当好的性能。由于某些不可避免的知识噪声同样可能引发幻觉，因此亟需研究其原因，并探讨在 KGD 任务中构建噪声容忍方法的未来方向。本文中，我们采用反事实推理方法分析该问题背后的因果脉络。基于因果效应分析，我们提出一种可能的解决方案，通过利用对话与知识之间的交互来缓解 KGD 中的幻觉。我们示例实现的实验结果表明，该方法能够在不损害其他对话性能的情况下减少幻觉，同时保持对不同生成模型的适应性。我们希望这些努力能够支持并呼吁更多关注，以开发面向鲁棒、可信对话系统的轻量级技术。
英文摘要：
Empowered by the large-scale pretrained language models, existing dialogue systems have demonstrated impressive performance conducting fluent and natural-sounding conversations. However, they are still plagued by the hallucination problem, causing unpredictable factual errors in the generated responses. Recently, knowledge-grounded dialogue generation models, that intentionally invoke external knowledge resources to more informative responses, are also proven to be effective in reducing hallucination. Following the idea of getting high-quality knowledge, a few efforts have achieved pretty good performance on this issue. As some inevitable knowledge noises may also lead to hallucinations, it is emergent to investigate the reason and future directions for building noise-tolerant methods in KGD tasks. In this paper, we analyze the causal story behind this problem with counterfactual reasoning methods. Based on the causal effect analysis, we propose a possible solution for alleviating the hallucination in KGD by exploiting the dialogue-knowledge interaction. Experimental results of our example implementation show that this method can reduce hallucination without disrupting other dialogue performance, while keeping adaptive to different generation models. We hope our efforts can support and call for more attention to developing lightweight techniques towards robust and trusty dialogue systems.
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。
