Transforming LLMs into Cross-modal and Cross-lingual Retrieval Systems
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_ | 1866911950113341440 |
|---|---|
| author | Gomez, Frank Palma Sanabria, Ramon Sung, Yun-hsuan Cer, Daniel Dalmia, Siddharth Abrego, Gustavo Hernandez |
| author_facet | Gomez, Frank Palma Sanabria, Ramon Sung, Yun-hsuan Cer, Daniel Dalmia, Siddharth Abrego, Gustavo Hernandez |
| contents | Large language models (LLMs) are trained on text-only data that go far beyond the languages with paired speech and text data. At the same time, Dual Encoder (DE) based retrieval systems project queries and documents into the same embedding space and have demonstrated their success in retrieval and bi-text mining. To match speech and text in many languages, we propose using LLMs to initialize multi-modal DE retrieval systems. Unlike traditional methods, our system doesn't require speech data during LLM pre-training and can exploit LLM's multilingual text understanding capabilities to match speech and text in languages unseen during retrieval training. Our multi-modal LLM-based retrieval system is capable of matching speech and text in 102 languages despite only training on 21 languages. Our system outperforms previous systems trained explicitly on all 102 languages. We achieve a 10% absolute improvement in Recall@1 averaged across these languages. Additionally, our model demonstrates cross-lingual speech and text matching, which is further enhanced by readily available machine translation data. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_01616 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Transforming LLMs into Cross-modal and Cross-lingual Retrieval Systems Gomez, Frank Palma Sanabria, Ramon Sung, Yun-hsuan Cer, Daniel Dalmia, Siddharth Abrego, Gustavo Hernandez Computation and Language Information Retrieval Sound Audio and Speech Processing Large language models (LLMs) are trained on text-only data that go far beyond the languages with paired speech and text data. At the same time, Dual Encoder (DE) based retrieval systems project queries and documents into the same embedding space and have demonstrated their success in retrieval and bi-text mining. To match speech and text in many languages, we propose using LLMs to initialize multi-modal DE retrieval systems. Unlike traditional methods, our system doesn't require speech data during LLM pre-training and can exploit LLM's multilingual text understanding capabilities to match speech and text in languages unseen during retrieval training. Our multi-modal LLM-based retrieval system is capable of matching speech and text in 102 languages despite only training on 21 languages. Our system outperforms previous systems trained explicitly on all 102 languages. We achieve a 10% absolute improvement in Recall@1 averaged across these languages. Additionally, our model demonstrates cross-lingual speech and text matching, which is further enhanced by readily available machine translation data. |
| title | Transforming LLMs into Cross-modal and Cross-lingual Retrieval Systems |
| topic | Computation and Language Information Retrieval Sound Audio and Speech Processing |
| url | https://arxiv.org/abs/2404.01616 |