Aligning Cross-lingual Sentence Representations with Dual Momentum Contrast
This work improves cross-lingual semantic similarity computation for multilingual NLP applications, but it is incremental as it builds on existing methods like MoCo.
The paper tackled aligning cross-lingual sentence representations by fine-tuning pre-trained models with a translation ranking task, adapting MoCo to address easy negatives, and achieved new state-of-the-art results on tasks like Tatoeba en-zh similarity search and BUCC en-zh bitext mining.
In this paper, we propose to align sentence representations from different languages into a unified embedding space, where semantic similarities (both cross-lingual and monolingual) can be computed with a simple dot product. Pre-trained language models are fine-tuned with the translation ranking task. Existing work (Feng et al., 2020) uses sentences within the same batch as negatives, which can suffer from the issue of easy negatives. We adapt MoCo (He et al., 2020) to further improve the quality of alignment. As the experimental results show, the sentence representations produced by our model achieve the new state-of-the-art on several tasks, including Tatoeba en-zh similarity search (Artetxe and Schwenk, 2019b), BUCC en-zh bitext mining, and semantic textual similarity on 7 datasets.