Despite the impressive advancements achieved through vision-and-language pretraining, it remains unclear whether this joint learning paradigm can help understand each individual modality. In this work, we conduct a comparative analysis of the visual representations in existing vision-and-language models and vision-only models by probing a broad range of tasks, aiming to assess the quality of the learned representations in a nuanced manner. Interestingly, our empirical observations suggest that vision-and-language models are better at label prediction tasks like object and attribute prediction, while vision-only models are stronger at dense prediction tasks that require more localized information. We hope our study sheds light on the role of language in visual learning, and serves as an empirical guide for various pretrained models. Code will be released at https://github.com/Lizw14/visual_probing
@article{arxiv.2212.00281,
title = {Localization vs. Semantics: Visual Representations in Unimodal and Multimodal Models},
author = {Zhuowan Li and Cihang Xie and Benjamin Van Durme and Alan Yuille},
journal= {arXiv preprint arXiv:2212.00281},
year = {2024}
}
Comments
Accepted to EACL 2024. Code is released at https://github.com/Lizw14/visual_probing