Domain-matched Pre-training Tasks for Dense Retrieval

Barlas Oguz, Kushal Lakhotia, Anchit Gupta, Patrick Lewis, Vladimir Karpukhin, Aleksandra Piktus, Xilun Chen, Sebastian Riedel, Scott Yih, Sonal Gupta, Yashar Mehdad


Abstract
Pre-training on larger datasets with ever increasing model size isnow a proven recipe for increased performance across almost all NLP tasks.A notable exception is information retrieval, where additional pre-traininghas so far failed to produce convincing results. We show that, with theright pre-training setup, this barrier can be overcome. We demonstrate thisby pre-training large bi-encoder models on 1) a recently released set of 65 millionsynthetically generated questions, and 2) 200 million post-comment pairs from a preexisting dataset of Reddit conversations made available by pushshift.io. We evaluate on a set of information retrieval and dialogue retrieval benchmarks, showing substantial improvements over supervised baselines.
Anthology ID:
2022.findings-naacl.114
Volume:
Findings of the Association for Computational Linguistics: NAACL 2022
Month:
July
Year:
2022
Address:
Seattle, United States
Editors:
Marine Carpuat, Marie-Catherine de Marneffe, Ivan Vladimir Meza Ruiz
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
1524–1534
Language:
URL:
https://aclanthology.org/2022.findings-naacl.114
DOI:
10.18653/v1/2022.findings-naacl.114
Bibkey:
Cite (ACL):
Barlas Oguz, Kushal Lakhotia, Anchit Gupta, Patrick Lewis, Vladimir Karpukhin, Aleksandra Piktus, Xilun Chen, Sebastian Riedel, Scott Yih, Sonal Gupta, and Yashar Mehdad. 2022. Domain-matched Pre-training Tasks for Dense Retrieval. In Findings of the Association for Computational Linguistics: NAACL 2022, pages 1524–1534, Seattle, United States. Association for Computational Linguistics.
Cite (Informal):
Domain-matched Pre-training Tasks for Dense Retrieval (Oguz et al., Findings 2022)
Copy Citation:
PDF:
https://aclanthology.org/2022.findings-naacl.114.pdf
Video:
 https://aclanthology.org/2022.findings-naacl.114.mp4
Code
 facebookresearch/dpr-scale
Data
ConvAI2DSTC7 Task 1KILTMS MARCONatural QuestionsPAQ