arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2408.05767cs.CLcs.AI
收藏

大型视觉语言模型的无参考幻觉检测

Reference-free Hallucination Detection for Large Vision-Language Models

Mohamed bin Zayed University of Artificial Intelligence(穆罕默德·本·扎耶德人工智能大学);Technical University of Munich(慕尼黑工业大学)

Qing Li, Jiahui Geng, Chenyang Lyu, Derui Zhu, Maxim Panov, Fakhri Karray

AI总结:

本文研究不依赖外部工具的无参考方法在大型视觉语言模型幻觉检测中的有效性,通过对比不确定性、一致性和有监督不确定性量化三类方法,发现无参考方法可有效检测非事实响应,且有监督不确定性量化方法表现最佳。

摘要:

大型视觉语言模型(LVLMs)近年来取得了显著进展。尽管LVLMs在语言理解、问答以及视觉输入的对话方面表现出色,但它们容易产生幻觉。虽然已有若干方法被提出用于评估LVLMs中的幻觉,但大多数方法都是基于参考的,并且依赖外部工具,这使得它们的实际应用变得复杂。为了评估替代方法的可行性,关键是要理解不依赖任何外部工具的无参考方法能否有效检测幻觉。因此,我们启动了一项探索性研究,以证明不同无参考解决方案在检测LVLMs幻觉方面的有效性。特别是,我们在两个不同任务上对四种代表性LVLMs进行了广泛研究,涉及三类技术:基于不确定性的方法、基于一致性的方法以及有监督的不确定性量化方法。实证结果表明,无参考方法能够有效检测LVLMs中的非事实性响应,其中有监督的不确定性量化方法优于其他方法,在不同设置下均取得了最佳性能。

英文摘要:

Large vision-language models (LVLMs) have made significant progress in recent years. While LVLMs exhibit excellent ability in language understanding, question answering, and conversations of visual inputs, they are prone to producing hallucinations. While several methods are proposed to evaluate the hallucinations in LVLMs, most are reference-based and depend on external tools, which complicates their practical application. To assess the viability of alternative methods, it is critical to understand whether the reference-free approaches, which do not rely on any external tools, can efficiently detect hallucinations. Therefore, we initiate an exploratory study to demonstrate the effectiveness of different reference-free solutions in detecting hallucinations in LVLMs. In particular, we conduct an extensive study on three kinds of techniques: uncertainty-based, consistency-based, and supervised uncertainty quantification methods on four representative LVLMs across two different tasks. The empirical results show that the reference-free approaches are capable of effectively detecting non-factual responses in LVLMs, with the supervised uncertainty quantification method outperforming the others, achieving the best performance across different settings.

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。

↑