面向可信检索增强生成的同步忠实性监控
Synchronous Faithfulness Monitoring for Trustworthy Retrieval-Augmented Generation
AI总结：
本文提出轻量级监控器 SynCheck，利用细粒度解码动态同步检测检索增强生成中的不忠实句子，并进一步提出忠实性导向解码算法 FOD，在多个任务上显著提升生成忠实性。
摘要：
检索增强语言模型（RALMs）在知识密集型任务中展现出强大性能和广泛适用性。然而，RALMs 存在显著的可信性问题，因为它们容易生成不忠实的输出，包括无根据的信息或与检索到的上下文相矛盾的内容。本文提出 SynCheck，一种轻量级监控器，利用细粒度解码动态——包括序列似然、不确定性量化、上下文影响和语义对齐——来同步检测不忠实的句子。通过整合可高效测量且互补的信号，SynCheck 能够实现准确、即时的反馈与干预，在六个长文本检索增强生成任务中检测忠实性错误时达到 0.85 AUROC，比先前最佳方法提高了 4%。利用 SynCheck，我们进一步引入 FOD，一种以忠实性为导向、由束搜索引导的长文本检索增强生成解码算法。实验结果表明，FOD 在忠实性方面显著优于弃权（不执行）、重排序或对比解码等传统策略，在六个数据集上实现了超过 10% 的提升。
英文摘要：
Retrieval-augmented language models (RALMs) have shown strong performance and wide applicability in knowledge-intensive tasks. However, there are significant trustworthiness concerns as RALMs are prone to generating unfaithful outputs, including baseless information or contradictions with the retrieved context. This paper proposes SynCheck, a lightweight monitor that leverages fine-grained decoding dynamics including sequence likelihood, uncertainty quantification, context influence, and semantic alignment to synchronously detect unfaithful sentences. By integrating efficiently measurable and complementary signals, SynCheck enables accurate and immediate feedback and intervention, achieving 0.85 AUROC in detecting faithfulness errors across six long-form retrieval-augmented generation tasks, improving prior best method by 4%. Leveraging SynCheck, we further introduce FOD, a faithfulness-oriented decoding algorithm guided by beam search for long-form retrieval-augmented generation. Empirical results demonstrate that FOD outperforms traditional strategies such as abstention, reranking, or contrastive decoding significantly in terms of faithfulness, achieving over 10% improvement across six datasets.
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。
