Multilingual Universal Sentence Encoder for Semantic Retrieval
This work addresses the need for efficient multilingual semantic retrieval tools, offering a practical solution for cross-lingual applications, though it is incremental as it builds on existing dual-encoder and translation-based methods.
The authors tackled the problem of multilingual semantic retrieval by introducing two pre-trained sentence encoding models based on Transformer and CNN architectures, which embed text from 16 languages into a single semantic space and achieve competitive performance with state-of-the-art on tasks like semantic retrieval, translation pair bitext retrieval, and retrieval question answering.
We introduce two pre-trained retrieval focused multilingual sentence encoding models, respectively based on the Transformer and CNN model architectures. The models embed text from 16 languages into a single semantic space using a multi-task trained dual-encoder that learns tied representations using translation based bridge tasks (Chidambaram al., 2018). The models provide performance that is competitive with the state-of-the-art on: semantic retrieval (SR), translation pair bitext retrieval (BR) and retrieval question answering (ReQA). On English transfer learning tasks, our sentence-level embeddings approach, and in some cases exceed, the performance of monolingual, English only, sentence embedding models. Our models are made available for download on TensorFlow Hub.