ACL Findings2023
Towards Robust Ranker for Text Retrieval
Yucheng Zhou, Tao Shen, Xiubo Geng, Chongyang Tao, Can Xu, Guodong Long, Binxing Jiao, Daxin Jiang
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 selfadversarial (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 R 2 ANKER), 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.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
3 top-tier papers have cited it since. Know what they found before your reviewers do.
Lune can draw on
the full paper · the top-tier papers that cite it · related work
Free to start. No credit card required.
Lineage
- LexMAE: Lexicon-Bottlenecked Pretraining for Large-Scale Retrieval cites Approximate Nearest Neighbor Negative Contrastive Learning for Dense Text Retrieval
- LexMAE: Lexicon-Bottlenecked Pretraining for Large-Scale Retrieval cites RocketQAv2: A Joint Training Method for Dense Passage Retrieval and Passage Re-ranking
- LexMAE: Lexicon-Bottlenecked Pretraining for Large-Scale Retrieval cites ColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction over BERT
- LexMAE: Lexicon-Bottlenecked Pretraining for Large-Scale Retrieval cites Optimizing Dense Retrieval Model Training with Hard Negatives
- Approximate Nearest Neighbor Negative Contrastive Learning for Dense Text Retrieval cites ColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction over BERT
- RocketQAv2: A Joint Training Method for Dense Passage Retrieval and Passage Re-ranking cites ColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction over BERT
- RocketQAv2: A Joint Training Method for Dense Passage Retrieval and Passage Re-ranking cites Optimizing Dense Retrieval Model Training with Hard Negatives
- Multi-Facet Blending for Faceted Query-by-Example Retrieval cites Approximate Nearest Neighbor Negative Contrastive Learning for Dense Text Retrieval
- Multi-Facet Blending for Faceted Query-by-Example Retrieval cites Optimizing Dense Retrieval Model Training with Hard Negatives
- Optimizing Dense Retrieval Model Training with Hard Negatives cites Approximate Nearest Neighbor Negative Contrastive Learning for Dense Text Retrieval
- Optimizing Dense Retrieval Model Training with Hard Negatives cites ColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction over BERT
Cited by top-tier papers3
- LexMAE: Lexicon-Bottlenecked Pretraining for Large-Scale RetrievalTao Shen, Xiubo Geng, Chongyang Tao, Can Xu et al.ICLR 2023 · 14 citations
- Multi-Facet Blending for Faceted Query-by-Example RetrievalHeejin Do, Sangwon Ryu, Jonghwi Kim, Gary LeeACL 2025
- Incomplete In-context LearningWenqiang Wang, Yujia Wen, Yan Xiao, Zhifeng Chen et al.ACL 2026
Builds on13
- Approximate Nearest Neighbor Negative Contrastive Learning for Dense Text RetrievalLee Xiong, Chenyan Xiong, Ye Li, Kwok-Fung Tang et al.ICLR 2021 · 1,547 citations
- ColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction over BERTOmar Khattab, Matei ZahariaSIGIR 2020 · 1,246 citations
- ELECTRA: Pre-training Text Encoders as Discriminators Rather Than GeneratorsKevin Clark, Minh-Thang Luong, Quoc V. Le, Christopher D. ManningICLR 2020 · 541 citations
- Optimizing Dense Retrieval Model Training with Hard NegativesJingtao Zhan, Jiaxin Mao, Yiqun Liu, Jiafeng Guo et al.SIGIR 2021 · 242 citations
- RocketQAv2: A Joint Training Method for Dense Passage Retrieval and Passage Re-rankingRuiyang Ren, Yingqi Qu, Jing Liu, Wayne Xin Zhao et al.EMNLP 2021 · 147 citations
Related papers
- Adversarial Retriever-Ranker for Dense Text RetrievalHang Zhang, Yeyun Gong, Yelong Shen, Jiancheng Lv et al.ICLR 2022 · 137 citations
- Contrastive Fine-tuning Improves Robustness for Neural RankersXiaofei Ma, Cícero Nogueira dos Santos, Andrew O. ArnoldACL Findings 2021 · 21 citations
- ADAM: Dense Retrieval Distillation with Adaptive Dark ExamplesChongyang Tao, Chang Liu, Tao Shen, Can Xu et al.ACL Findings 2024 · 2 citations
- Training Dense Retrievers with Multiple Positive PassagesBenben Wang, Minghao Tang, Hengran Zhang, Jiafeng Guo et al.KDD 2026 · 1 citation
- Cooperative Retriever and Ranker in Deep RecommendersXu Huang, Defu Lian, Jin Chen, Zheng Liu et al.WWW 2023 · 17 citations
