Weighted Point Set Embedding for Multimodal Contrastive Learning Toward Optimal Similarity

This repository houses the official implementation of the paper titled "Weighted Point Set Embedding for Multimodal Contrastive Learning Toward Optimal Similarity", which is presented at ICLR 2025. [OpenReview]

The code will be released soon. Please stay tuned.

Contact:

Toshimitsu Uesaka: toshimitsu.uesaka@sony.com

Citation

@inproceedings{
uesaka2025weighted,
title={Weighted Point Set Embedding for Multimodal Contrastive Learning Toward Optimal Similarity Metric},
author={Toshimitsu Uesaka and Taiji Suzuki and Yuhta Takida and Chieh-Hsin Lai and Naoki Murata and Yuki Mitsufuji},
booktitle={The Thirteenth International Conference on Learning Representations},
year={2025},
url={https://openreview.net/forum?id=uSz2K30RRd}
}