arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2407.06192cs.CVcs.AIcs.CL
收藏

视觉-语言模型中的多对象幻觉

Multi-Object Hallucination in Vision-Language Models

University of Michigan(密歇根大学);University of Virginia(弗吉尼亚大学);New York University(纽约大学)

Xuweiyi Chen, Ziqiao Ma, Xuejun Zhang, Sihan Xu, Shengyi Qian, Jianing Yang, David F. Fouhey, Joyce Chai

AI总结:

本研究针对视觉语言模型的多对象幻觉问题,提出自动化评估协议ROPE,通过实验揭示多对象场景下幻觉更严重、受类别分布与数据及模型因素影响,为提升模型多对象推理能力提供参考。

摘要:

大型视觉语言模型(LVLM)常存在对象幻觉问题,会生成给定图像中不存在的对象。当前针对对象幻觉的基准测试主要关注单类别对象的存在性,而非单个实体,本研究系统探究了多对象幻觉,考察模型在需要同时关注多个对象的任务中出现的感知错误(例如凭空捏造不存在的对象或注意力分散)。我们提出了基于识别的对象探测评估(ROPE),这是一种自动化评估协议,测试时会考虑单张图像内的对象类别分布,并使用视觉指代提示消除歧义。通过全面的实证研究以及对多对象幻觉潜在影响因素的分析,我们发现:(1)与关注单个对象相比,LVLM在关注多个对象时会产生更多幻觉。(2)测试的对象类别分布会影响幻觉行为,表明LVLM可能依赖捷径和虚假相关性。(3)幻觉行为受数据特定因素(显著性与频率)以及模型内在行为的影响。我们希望能让LVLM具备识别和推理现实视觉场景中常见的多对象的能力,为缓解该问题提供洞见并量化我们的进展。

英文摘要:

Large vision language models (LVLMs) often suffer from object hallucination, producing objects not present in the given images. While current benchmarks for object hallucination primarily concentrate on the presence of a single object class rather than individual entities, this work systematically investigates multi-object hallucination, examining how models misperceive (e.g., invent nonexistent objects or become distracted) when tasked with focusing on multiple objects simultaneously. We introduce Recognition-based Object Probing Evaluation (ROPE), an automated evaluation protocol that considers the distribution of object classes within a single image during testing and uses visual referring prompts to eliminate ambiguity. With comprehensive empirical studies and analysis of potential factors leading to multi-object hallucination, we found that (1). LVLMs suffer more hallucinations when focusing on multiple objects compared to a single object. (2). The tested object class distribution affects hallucination behaviors, indicating that LVLMs may follow shortcuts and spurious correlations. (3). Hallucinatory behaviors are influenced by data-specific factors, salience and frequency, and model intrinsic behaviors. We hope to enable LVLMs to recognize and reason about multiple objects that often occur in realistic visual scenes, provide insights, and quantify our progress towards mitigating the issues.

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。

↑