Unsupervised Multilingual Alignment using Wasserstein Barycenter
Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence
Main track. Pages 3702-3708. https://doi.org/10.24963/ijcai.2020/512
We study unsupervised multilingual alignment, the problem of finding word-to-word translations between multiple languages without using any parallel data. One popular strategy is to reduce multilingual alignment to the much simplified bilingual setting, by picking one of the input languages as the pivot language that we transit through. However, it is well-known that transiting through a poorly chosen pivot language (such as English) may severely degrade the translation quality, since the assumed transitive relations among all pairs of languages may not be enforced in the training process. Instead of going through a rather arbitrarily chosen pivot language, we propose to use the Wasserstein barycenter as a more informative ``mean'' language: it encapsulates information from all languages and minimizes all pairwise transportation costs. We evaluate our method on standard benchmarks and demonstrate state-of-the-art performances.
Natural Language Processing: Machine Translation
Machine Learning: Unsupervised Learning
Machine Learning Applications: Applications of Unsupervised Learning