Multilingual Sentence-T5: Scalable Sentence Encoders for Multilingual Applications
This work addresses the need for scalable and high-performance multilingual sentence embeddings, which is incremental as it extends an existing monolingual model with a known adaptation technique.
The paper tackles the problem of scaling multilingual sentence encoders by introducing Multilingual Sentence T5 (m-ST5), a 5.7-billion-parameter model that outperforms prior NLI-based approaches and shows improved performance, especially for low-resource languages and those less similar to English.
Prior work on multilingual sentence embedding has demonstrated that the efficient use of natural language inference (NLI) data to build high-performance models can outperform conventional methods. However, the potential benefits from the recent ``exponential'' growth of language models with billions of parameters have not yet been fully explored. In this paper, we introduce Multilingual Sentence T5 (m-ST5), as a larger model of NLI-based multilingual sentence embedding, by extending Sentence T5, an existing monolingual model. By employing the low-rank adaptation (LoRA) technique, we have achieved a successful scaling of the model's size to 5.7 billion parameters. We conducted experiments to evaluate the performance of sentence embedding and verified that the method outperforms the NLI-based prior approach. Furthermore, we also have confirmed a positive correlation between the size of the model and its performance. It was particularly noteworthy that languages with fewer resources or those with less linguistic similarity to English benefited more from the parameter increase. Our model is available at https://huggingface.co/pkshatech/m-ST5.