@inproceedings{zhou-etal-2023-towards-robust,
title = "Towards Robust Ranker for Text Retrieval",
author = "Zhou, Yucheng and
Shen, Tao and
Geng, Xiubo and
Tao, Chongyang and
Xu, Can and
Long, Guodong and
Jiao, Binxing and
Jiang, Daxin",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Findings of the Association for Computational Linguistics: ACL 2023",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-acl.332",
doi = "10.18653/v1/2023.findings-acl.332",
pages = "5387--5401",
abstract = "A neural ranker plays an indispensable role in the de facto {`}retrieval {\&} rerank{'} pipeline, but its training still lags behind due to the weak negative mining during contrastive learning. Compared to retrievers boosted by self-adversarial (i.e., in-distribution) negative mining, the ranker{'}s heavy structure suffers from query-document combinatorial explosions, so it can only resort to the negative sampled by the fast yet out-of-distribution retriever. Thereby, the moderate negatives compose ineffective contrastive learning samples, becoming the main barrier to learning a robust ranker. To alleviate this, we propose a multi-adversarial training strategy that leverages multiple retrievers as generators to challenge a ranker, where i) diverse hard negatives from a joint distribution are prone to fool the ranker for more effective adversarial learning and ii) involving extensive out-of-distribution label noises renders the ranker against each noise distribution, leading to more challenging and robust contrastive learning. To evaluate our robust ranker (dubbed R2anker), we conduct experiments in various settings on the passage retrieval benchmarks, including BM25-reranking, full-ranking, retriever distillation, etc. The empirical results verify the new state-of-the-art effectiveness of our model.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="zhou-etal-2023-towards-robust">
<titleInfo>
<title>Towards Robust Ranker for Text Retrieval</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yucheng</namePart>
<namePart type="family">Zhou</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tao</namePart>
<namePart type="family">Shen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xiubo</namePart>
<namePart type="family">Geng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Chongyang</namePart>
<namePart type="family">Tao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Can</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Guodong</namePart>
<namePart type="family">Long</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Binxing</namePart>
<namePart type="family">Jiao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Daxin</namePart>
<namePart type="family">Jiang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2023-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2023</title>
</titleInfo>
<name type="personal">
<namePart type="given">Anna</namePart>
<namePart type="family">Rogers</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jordan</namePart>
<namePart type="family">Boyd-Graber</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Naoaki</namePart>
<namePart type="family">Okazaki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Toronto, Canada</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>A neural ranker plays an indispensable role in the de facto ‘retrieval & rerank’ pipeline, but its training still lags behind due to the weak negative mining during contrastive learning. Compared to retrievers boosted by self-adversarial (i.e., in-distribution) negative mining, the ranker’s heavy structure suffers from query-document combinatorial explosions, so it can only resort to the negative sampled by the fast yet out-of-distribution retriever. Thereby, the moderate negatives compose ineffective contrastive learning samples, becoming the main barrier to learning a robust ranker. To alleviate this, we propose a multi-adversarial training strategy that leverages multiple retrievers as generators to challenge a ranker, where i) diverse hard negatives from a joint distribution are prone to fool the ranker for more effective adversarial learning and ii) involving extensive out-of-distribution label noises renders the ranker against each noise distribution, leading to more challenging and robust contrastive learning. To evaluate our robust ranker (dubbed R2anker), we conduct experiments in various settings on the passage retrieval benchmarks, including BM25-reranking, full-ranking, retriever distillation, etc. The empirical results verify the new state-of-the-art effectiveness of our model.</abstract>
<identifier type="citekey">zhou-etal-2023-towards-robust</identifier>
<identifier type="doi">10.18653/v1/2023.findings-acl.332</identifier>
<location>
<url>https://aclanthology.org/2023.findings-acl.332</url>
</location>
<part>
<date>2023-07</date>
<extent unit="page">
<start>5387</start>
<end>5401</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Towards Robust Ranker for Text Retrieval
%A Zhou, Yucheng
%A Shen, Tao
%A Geng, Xiubo
%A Tao, Chongyang
%A Xu, Can
%A Long, Guodong
%A Jiao, Binxing
%A Jiang, Daxin
%Y Rogers, Anna
%Y Boyd-Graber, Jordan
%Y Okazaki, Naoaki
%S Findings of the Association for Computational Linguistics: ACL 2023
%D 2023
%8 July
%I Association for Computational Linguistics
%C Toronto, Canada
%F zhou-etal-2023-towards-robust
%X A neural ranker plays an indispensable role in the de facto ‘retrieval & rerank’ pipeline, but its training still lags behind due to the weak negative mining during contrastive learning. Compared to retrievers boosted by self-adversarial (i.e., in-distribution) negative mining, the ranker’s heavy structure suffers from query-document combinatorial explosions, so it can only resort to the negative sampled by the fast yet out-of-distribution retriever. Thereby, the moderate negatives compose ineffective contrastive learning samples, becoming the main barrier to learning a robust ranker. To alleviate this, we propose a multi-adversarial training strategy that leverages multiple retrievers as generators to challenge a ranker, where i) diverse hard negatives from a joint distribution are prone to fool the ranker for more effective adversarial learning and ii) involving extensive out-of-distribution label noises renders the ranker against each noise distribution, leading to more challenging and robust contrastive learning. To evaluate our robust ranker (dubbed R2anker), we conduct experiments in various settings on the passage retrieval benchmarks, including BM25-reranking, full-ranking, retriever distillation, etc. The empirical results verify the new state-of-the-art effectiveness of our model.
%R 10.18653/v1/2023.findings-acl.332
%U https://aclanthology.org/2023.findings-acl.332
%U https://doi.org/10.18653/v1/2023.findings-acl.332
%P 5387-5401
Markdown (Informal)
[Towards Robust Ranker for Text Retrieval](https://aclanthology.org/2023.findings-acl.332) (Zhou et al., Findings 2023)
ACL
- Yucheng Zhou, Tao Shen, Xiubo Geng, Chongyang Tao, Can Xu, Guodong Long, Binxing Jiao, and Daxin Jiang. 2023. Towards Robust Ranker for Text Retrieval. In Findings of the Association for Computational Linguistics: ACL 2023, pages 5387–5401, Toronto, Canada. Association for Computational Linguistics.