UOUO: Uncontextualized Uncommon Objects for Measuring Knowledge Horizons of Vision Language Models

Xinyu Pi, Mingyuan Wu, Jize Jiang, Haozhen Zheng, Beitong Tian, ChengXiang Zhai, Klara Nahrstedt, Zhiting Hu


Abstract
Smaller-scale Vision-Language Models (VLMs) often claim to perform on par with larger models in general-domain visual grounding and question-answering benchmarks while offering advantages in computational efficiency and storage. However, their ability to handle rare objects, which fall into the long tail of data distributions, is less understood. To rigorously evaluate this aspect, we introduce the “Uncontextualized Uncommon Objects” (UOUO) benchmark. This benchmark focuses on systematically testing VLMs with both large and small parameter counts on rare and specialized objects. Our comprehensive analysis reveals that while smaller VLMs maintain competitive performance on common datasets, they significantly underperform on tasks involving uncommon objects. We also propose an advanced, scalable pipeline for data collection and cleaning, ensuring the UOUO benchmark provides high-quality, challenging instances. These findings highlight the need to consider long-tail distributions when assessing the true capabilities of VLMs. Code and project details for UOUO can be found at https://zoezheng126.github.io/UOUO-Website/.
Anthology ID:
2024.emnlp-main.369
Volume:
Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing
Month:
November
Year:
2024
Address:
Miami, Florida, USA
Editors:
Yaser Al-Onaizan, Mohit Bansal, Yun-Nung Chen
Venue:
EMNLP
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
6432–6441
Language:
URL:
https://aclanthology.org/2024.emnlp-main.369
DOI:
10.18653/v1/2024.emnlp-main.369
Bibkey:
Cite (ACL):
Xinyu Pi, Mingyuan Wu, Jize Jiang, Haozhen Zheng, Beitong Tian, ChengXiang Zhai, Klara Nahrstedt, and Zhiting Hu. 2024. UOUO: Uncontextualized Uncommon Objects for Measuring Knowledge Horizons of Vision Language Models. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 6432–6441, Miami, Florida, USA. Association for Computational Linguistics.
Cite (Informal):
UOUO: Uncontextualized Uncommon Objects for Measuring Knowledge Horizons of Vision Language Models (Pi et al., EMNLP 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.emnlp-main.369.pdf