English

Diversifying Joint Vision-Language Tokenization Learning

Computer Vision and Pattern Recognition 2023-06-19 v2

Abstract

Building joint representations across images and text is an essential step for tasks such as Visual Question Answering and Video Question Answering. In this work, we find that the representations must not only jointly capture features from both modalities but should also be diverse for better generalization performance. To this end, we propose joint vision-language representation learning by diversifying the tokenization learning process, enabling tokens that are sufficiently disentangled from each other to be learned from both modalities. We observe that our approach outperforms the baseline models in a majority of settings and is competitive with state-of-the-art methods.

Keywords

Cite

@article{arxiv.2306.03421,
  title  = {Diversifying Joint Vision-Language Tokenization Learning},
  author = {Vardaan Pahuja and AJ Piergiovanni and Anelia Angelova},
  journal= {arXiv preprint arXiv:2306.03421},
  year   = {2023}
}

Comments

Accepted to Transformers for Vision (T4V) workshop, CVPR 2023; 7 pages, 5 figures

R2 v1 2026-06-28T10:57:27.885Z