This is an implementation of word rotator's distance (WRD), a simple yet effective textual similarity measure based on optimal transport, presented at EMNLP 2020. The paper is available online.
More details coming soon!
Several researchers have published reimplementations in Python:
- Thank @bojone for writing a blog article with Python code in English 🇬🇧, and a blog article with Python code in Chinese 🇨🇳.
- Thank @kenta1984 for writing Python code and a blog article in Japanese 🇯🇵.
Currently, the latest version is available on arXiv (URL) and ACL anthology (v2).
If you use our code for academic work, please cite
@inproceedings{yokoi-emnlp2020-wrd,
title = "Word Rotator{'}s Distance",
author = "Yokoi, Sho and
Takahashi, Ryo and
Akama, Reina and
Suzuki, Jun and
Inui, Kentaro",
booktitle = "Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)",
month = nov,
year = "2020",
address = "Online",
publisher = "Association for Computational Linguistics",
url = "https://www.aclweb.org/anthology/2020.emnlp-main.236",
doi = "10.18653/v1/2020.emnlp-main.236",
pages = "2944--2960",
}