Unsupervised Multilingual Dense Retrieval via Generative Pseudo Labeling
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916148669317120 |
|---|---|
| author | Huang, Chao-Wei Li, Chen-An Hsu, Tsu-Yuan Hsu, Chen-Yu Chen, Yun-Nung |
| author_facet | Huang, Chao-Wei Li, Chen-An Hsu, Tsu-Yuan Hsu, Chen-Yu Chen, Yun-Nung |
| contents | Dense retrieval methods have demonstrated promising performance in multilingual information retrieval, where queries and documents can be in different languages. However, dense retrievers typically require a substantial amount of paired data, which poses even greater challenges in multilingual scenarios. This paper introduces UMR, an Unsupervised Multilingual dense Retriever trained without any paired data. Our approach leverages the sequence likelihood estimation capabilities of multilingual language models to acquire pseudo labels for training dense retrievers. We propose a two-stage framework which iteratively improves the performance of multilingual dense retrievers. Experimental results on two benchmark datasets show that UMR outperforms supervised baselines, showcasing the potential of training multilingual retrievers without paired data, thereby enhancing their practicality. Our source code, data, and models are publicly available at https://github.com/MiuLab/UMR |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2403_03516 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Unsupervised Multilingual Dense Retrieval via Generative Pseudo Labeling Huang, Chao-Wei Li, Chen-An Hsu, Tsu-Yuan Hsu, Chen-Yu Chen, Yun-Nung Computation and Language Information Retrieval Dense retrieval methods have demonstrated promising performance in multilingual information retrieval, where queries and documents can be in different languages. However, dense retrievers typically require a substantial amount of paired data, which poses even greater challenges in multilingual scenarios. This paper introduces UMR, an Unsupervised Multilingual dense Retriever trained without any paired data. Our approach leverages the sequence likelihood estimation capabilities of multilingual language models to acquire pseudo labels for training dense retrievers. We propose a two-stage framework which iteratively improves the performance of multilingual dense retrievers. Experimental results on two benchmark datasets show that UMR outperforms supervised baselines, showcasing the potential of training multilingual retrievers without paired data, thereby enhancing their practicality. Our source code, data, and models are publicly available at https://github.com/MiuLab/UMR |
| title | Unsupervised Multilingual Dense Retrieval via Generative Pseudo Labeling |
| topic | Computation and Language Information Retrieval |
| url | https://arxiv.org/abs/2403.03516 |