Learning Multilingual Sentence Representations with Cross-lingual Consistency Regularization
This work addresses the need for efficient bitext mining to expand multilingual NMT systems, though it is incremental as it builds on existing cross-lingual regularization techniques.
The paper tackles the problem of learning multilingual sentence representations for bitext mining to scale neural machine translation, introducing MuSR, a model supporting over 220 languages that outperforms LASER3 on similarity search and bitext mining tasks.
Multilingual sentence representations are the foundation for similarity-based bitext mining, which is crucial for scaling multilingual neural machine translation (NMT) system to more languages. In this paper, we introduce MuSR: a one-for-all Multilingual Sentence Representation model that supports more than 220 languages. Leveraging billions of English-centric parallel corpora, we train a multilingual Transformer encoder, coupled with an auxiliary Transformer decoder, by adopting a multilingual NMT framework with CrossConST, a cross-lingual consistency regularization technique proposed in Gao et al. (2023). Experimental results on multilingual similarity search and bitext mining tasks show the effectiveness of our approach. Specifically, MuSR achieves superior performance over LASER3 (Heffernan et al., 2022) which consists of 148 independent multilingual sentence encoders.