视觉-语言模型具备时长常识推理能力吗？人机协同标注的推特语料研究
Leaving, Staying, or Living There? Duration Commonsense Reasoning with Human-LLM Co-annotations
AI 导读
一项研究探究视觉-语言模型是否具备“时长常识”——如判断发帖人是在某地居住还是短暂停留。研究者构建了由人类与大语言模型（LLM）协同标注的地理标记推特语料库作为测试集。结果显示，即使采用多轮 Chain-of-Thought 提示，SOTA 模型在时长推断上仍表现不佳，常常无法整合关键上下文信息，或过度依赖显著视觉细节而忽视隐晦的话题标签与语义线索。
正文 · 原文
Abstract:Gardening in the backyard implies living there, while checking into an Airbnb more likely means staying for a few days. Such duration judgments are important for understanding human mobility, travel behavior, and social context. However, duration is rarely stated explicitly in social media posts and must instead be inferred from noisy textual, visual, and contextual cues. In this paper, we investigate whether vision-language models possess such duration commonsense. We create a new corpus of geotagged tweets co-annotated by humans and LLMs to serve as a testbed for this task. We find that state-of-the-art models struggle with duration inference, even when guided by multi-turn Chain-of-Thought prompting. Error analyses further reveal that models often fail to integrate critical contextual information, or they overly rely on prominent visual details while overlooking subtle hashtags and semantic nuances. These findings underscore the significant challenges remaining in context-aware multimodal understanding.
| Subjects: | Artificial Intelligence (cs.AI); Information Retrieval (cs.IR) |
| Cite as: | arXiv:2610.38755 [cs.AI] |
| (or arXiv:2610.38755v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.38755 arXiv-issued DOI via DataCite |
Submission history
From: Chenyang Yu [view email]
[v1]
Wed, 28 Oct 2026 19:35:59 UTC (36,591 KB)
来源：arXiv cs.AI · arxiv.org