ColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction over BERT
The DOI matches its Crossref record; 26 of the 26 references it cites were found in a registry or an index.
Abstract
Recent progress in Natural Language Understanding (NLU) is driving fast-paced advances in Information Retrieval (IR), largely owed to fine-tuning deep language models (LMs) for document ranking. While remarkably effective, the ranking models based on these LMs increase computational cost by orders of magnitude over prior approaches, particularly as they must feed each query-document pair through a massive neural network to compute a single relevance score. To tackle this, we present ColBERT, a novel ranking model that adapts deep LMs (in particular, BERT) for efficient retrieval. ColBERT introduces a late interaction architecture that independently encodes the query and the document using BERT and then employs a cheap yet powerful interaction step that models their fine-grained similarity. By delaying and yet retaining this fine-granular interaction, ColBERT can leverage the expressiveness of deep LMs while simultaneously gaining the ability to pre-compute document representations offline, considerably speeding up query processing. Crucially, ColBERT's pruning-friendly interaction mechanism enables leveraging vector-similarity indexes for end-to-end retrieval directly from millions of documents. We extensively evaluate ColBERT using two recent passage search datasets. Results show that ColBERT's effectiveness is competitive with existing BERT-based models (and outperforms every non-BERT baseline), while executing two orders-of-magnitude faster and requiring up to four orders-of-magnitude fewer FLOPs per query.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“The only ML component is a CrossEncoder reranker ( mxbai-rerank-large-v1 , 435M parameters, DeBERTaV3) fused with ColBERT (Khattab & Zaharia, 2020) via Reciprocal Rank Fusion (RRF), applied after retrieval is complete to prioritize passages before token-budget truncation.”
From SmartSearch: How Ranking Beats Structure for Conversational Memory Retrieval (2026) · cites this paper as Khattab & Zaharia, 2020
In the citing paper
“While dense retrievers like Dense Passage Retrieval (DPR) [14] and late-interaction models like ColBERT [15] improve recall on single-hop queries, they often struggle with multi-hop reasoning, where the answer depends on composing information from multiple disjoint documents [20, 6].”
From Replace, Don't Expand: Mitigating Context Dilution in Multi-Hop RAG via Fixed-Budget Evidence Assembly (2025) · cites this paper as [15]
In the citing paper
“Research on retrieval optimization has explored multiple techniques, yet studies specifically focused on prefetching size in RaLM systems are still limited.”
From Exploring Optimal Prefetching Sizes in RaLMSpec to Enhance Retrieval-Augmented Generation Efficiency (2025) · cites this paper as [9]
In the citing paper
“Early RAG implementations, however, often relied on document retrieval methods that introduced noise and excessive context [26, 55, 60, 122].”
From LEGO-GraphRAG: Modularizing Graph-Based Retrieval-Augmented Generation for Design Space Exploration (2025) · cites this paper as [55]
In the citing paper
“Standard retrieval methods like RAG and MemoryBank (Zhong et al., 2024) rely fundamentally on vector similarity (Karpukhin et al., 2020; Khattab and Zaharia, 2020), representing memories as isolated points in embedding space (Hu et al., 2025). Consequently, they struggle with queries requiring causal bridging between semantically dissimilar or distant events (Yang et al., 2018; Qi et al., 2019; Tr”
From Synapse: Empowering LLM Agents with Episodic-Semantic Memory via Spreading Activation (2026) · cites this paper as Khattab and Zaharia, 2020
Citations
Cites
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., & Polosukhin, I. (2017) · arXiv · doi:10.48550/arxiv.1706.03762 · Cited by 120,404 · cited 1×
“At a high level, TK improves the KNRM architecture described earlier: while KNRM employs kernel pooling on top of word-embedding-based interaction, TK uses a Transformer [34] component for contextually encoding queries and documents before kernel pooling.”
MethodBERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2018) · arXiv · doi:10.48550/arxiv.1810.04805 · Cited by 73,173 · cited 2×
“Among them, a recent approach has emerged that ne-tunes deep pre-trained language models (LMs) like ELMo [29] and BERT [5] for estimating relevance.”
BackgroundGoogle's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation
Wu, Y., Schuster, M., Chen, Z., Le, Q. V., Norouzi, M., Macherey, W., Krikun, M., Cao, Y., Gao, Q., Macherey, K., et al. (2016) · arXiv · doi:10.48550/arxiv.1609.08144 · Cited by 5,319
Billion-scale similarity search with GPUs
Johnson, J., Douze, M., & Jégou, H. (2017) · arXiv · doi:10.48550/arxiv.1702.08734 · Cited by 4,140 · cited 2×
“Additionally, it enables ColBERT to leverage vector-similarity search indexes (e.g., [1, 16]) to retrieve the top-k results directly from a large document collection, substantially improving recall over models that only re-rank the output of term-based retrieval.”
MethodMS MARCO: A Human Generated MAchine Reading COmprehension Dataset
Bajaj, P., Campos, D., Craswell, N., Deng, L., Gao, J., Liu, X., Majumder, R., McNamara, A., Mitra, B., Nguyen, T., et al. (2016) · arXiv · doi:10.48550/arxiv.1611.09268 · Cited by 2,473
Robertson, S. E., Walker, S., Jones, S., Hancock‐Beaulieu, M., & Gatford, M. (1994) · Text REtrieval Conference · Cited by 2,212
Learning deep structured semantic models for web search using clickthrough data
Huang, P.-S., He, X., Gao, J., Deng, L., Acero, A., & Heck, L. (2013) · Proceedings of the 22nd ACM international conference on Information & Knowledge Management · doi:10.1145/2505515.2505665 · Cited by 2,009 · cited 1×
“On the le, Figure 2 (a) illustrates representation-focused rankers, which independently compute an embedding for q and another for d and estimate relevance as a single similarity score between two vectors [13, 41].”
BackgroundTinyBERT: Distilling BERT for natural language understanding
Jiao, X., Yin, Y., Shang, L., Jiang, X., Chen, X., Li, L., Wang, F., & Liu, Q. (2020) · Findings of the Association for Computational Linguistics: EMNLP 2020 · doi:10.18653/v1/2020.findings-emnlp.372 · Cited by 1,786
A Deep Relevance Matching Model for Ad-hoc Retrieval
Guo, J., Fan, Y., Ai, Q., & Croft, W. B. (2016) · Proceedings of the 25th ACM International on Conference on Information and Knowledge Management · doi:10.1145/2983323.2983769 · Cited by 806 · cited 1×
“Over the past few years, the Information Retrieval (IR) community has witnessed the introduction of a host of neural ranking models, including DRMM [8], KNRM [5, 36], and Duet [21, 23].”
BackgroundNogueira, R., & Cho, K. (2019) · arXiv · doi:10.48550/arxiv.1901.04085 · Cited by 615 · cited 9×
“Similar to [25], we fine-tune all ColBERT models with learning rate 3 × 10 − 6 with a batch size 32.”
BackgroundMethodInfluential
Research works printed but not linked: 14
- Zhuyun Dai and Jamie Callan . 2019 a. Context-Aware Sentence/Passage Term Importance Estimation For First Stage Retrieval. arXiv preprint arXiv:1910.10687 ( 2019 ). Zhuyun Dai and Jamie Callan. 2019 a. Context-Aware Sentence/Passage Term Importance Estimation For First Stage Retrieval. arXiv preprint arXiv:1910.10687 (2019). · arXiv
- Zhuyun Dai and Jamie Callan . 2019 b. Deeper Text Understanding for IR with Contextual Neural Language Modeling. arXiv preprint arXiv:1905.09217 ( 2019 ). Zhuyun Dai and Jamie Callan. 2019 b. Deeper Text Understanding for IR with Contextual Neural Language Modeling. arXiv preprint arXiv:1905.09217 (2019). · arXiv
- Jiafeng Guo , Yixing Fan , Liang Pang , Liu Yang , Qingyao Ai , Hamed Zamani , Chen Wu , W Bruce Croft , and Xueqi Cheng . 2019. A deep look into neural ranking models for information retrieval. arXiv preprint arXiv:1903.06902 ( 2019 ). Jiafeng Guo, Yixing Fan, Liang Pang, Liu Yang, Qingyao Ai, Hamed Zamani, Chen Wu, W Bruce Croft, and Xueqi Cheng. 2019. A deep look into neural ranking models for information retrieval. arXiv preprint arXiv:1903.06902 (2019). · arXiv
- Sebastian Hofst"atter and Allan Hanbury . 2019. Let's measure run time! Extending the IR replicability infrastructure to include performance aspects. arXiv preprint arXiv:1907.04614 ( 2019 ). Sebastian Hofst"atter and Allan Hanbury. 2019. Let's measure run time! Extending the IR replicability infrastructure to include performance aspects. arXiv preprint arXiv:1907.04614 (2019). · arXiv
- Sebastian Hofst"atter, Markus Zlabinger , and Allan Hanbury . 2019 b. TU Wien@ TREC Deep Learning'19--Simple Contextualization for Re-ranking. arXiv preprint arXiv:1912.01385 ( 2019 ). Sebastian Hofst"atter, Markus Zlabinger, and Allan Hanbury. 2019 b. TU Wien@ TREC Deep Learning'19--Simple Contextualization for Re-ranking. arXiv preprint arXiv:1912.01385 (2019). · arXiv
- Diederik P Kingma and Jimmy Ba . 2014 . Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980 (2014). Diederik P Kingma and Jimmy Ba. 2014. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980 (2014). · arXiv
- Bhaskar Mitra and Nick Craswell . 2019. An Updated Duet Model for Passage Re-ranking. arXiv preprint arXiv:1903.07666 ( 2019 ). Bhaskar Mitra and Nick Craswell. 2019. An Updated Duet Model for Passage Re-ranking. arXiv preprint arXiv:1903.07666 (2019). · arXiv
- Bhaskar Mitra , Corby Rosset , David Hawking , Nick Craswell , Fernando Diaz , and Emine Yilmaz . 2019. Incorporating query term independence assumption for efficient retrieval and ranking using deep neural networks. arXiv preprint arXiv:1907.03693 ( 2019 ). Bhaskar Mitra, Corby Rosset, David Hawking, Nick Craswell, Fernando Diaz, and Emine Yilmaz. 2019. Incorporating query term independence assumption for efficient retrieval and ranking using deep neural networks. arXiv preprint arXiv:1907.03693 (2019). · arXiv
- Rodrigo Nogueira , Wei Yang , Kyunghyun Cho , and Jimmy Lin . 2019 b. Multi-Stage Document Ranking with BERT. arXiv preprint arXiv:1910.14424 ( 2019 ). Rodrigo Nogueira, Wei Yang, Kyunghyun Cho, and Jimmy Lin. 2019 b. Multi-Stage Document Ranking with BERT. arXiv preprint arXiv:1910.14424 (2019). · arXiv
- Rodrigo Nogueira , Wei Yang , Jimmy Lin , and Kyunghyun Cho . 2019 c. Document Expansion by Query Prediction. arXiv preprint arXiv:1904.08375 ( 2019 ). Rodrigo Nogueira, Wei Yang, Jimmy Lin, and Kyunghyun Cho. 2019 c. Document Expansion by Query Prediction. arXiv preprint arXiv:1904.08375 (2019). · arXiv
- Matthew E Peters , Mark Neumann , Mohit Iyyer , Matt Gardner , Christopher Clark , Kenton Lee , and Luke Zettlemoyer . 2018. Deep contextualized word representations. arXiv preprint arXiv:1802.05365 ( 2018 ). Matthew E Peters, Mark Neumann, Mohit Iyyer, Matt Gardner, Christopher Clark, Kenton Lee, and Luke Zettlemoyer. 2018. Deep contextualized word representations. arXiv preprint arXiv:1802.05365 (2018). · arXiv
- Yifan Qiao , Chenyan Xiong , Zhenghao Liu , and Zhiyuan Liu . 2019. Understanding the Behaviors of BERT in Ranking. arXiv preprint arXiv:1904.07531 ( 2019 ). Yifan Qiao, Chenyan Xiong, Zhenghao Liu, and Zhiyuan Liu. 2019. Understanding the Behaviors of BERT in Ranking. arXiv preprint arXiv:1904.07531 (2019). · arXiv
- Colin Raffel , Noam Shazeer , Adam Roberts , Katherine Lee , Sharan Narang , Michael Matena , Yanqi Zhou , Wei Li , and Peter J Liu . 2019. Exploring the limits of transfer learning with a unified text-to-text transformer. arXiv preprint arXiv:1910.10683 ( 2019 ). Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, and Peter J Liu. 2019. Exploring the limits of transfer learning with a unified text-to-text transformer. arXiv preprint arXiv:1910.10683 (2019). · arXiv
- Raphael Tang , Yao Lu , Linqing Liu , Lili Mou , Olga Vechtomova , and Jimmy Lin . 2019. Distilling task-specific knowledge from BERT into simple neural networks. arXiv preprint arXiv:1903.12136 ( 2019 ). Raphael Tang, Yao Lu, Linqing Liu, Lili Mou, Olga Vechtomova, and Jimmy Lin. 2019. Distilling task-specific knowledge from BERT into simple neural networks. arXiv preprint arXiv:1903.12136 (2019). · arXiv
Cited by
Items 301–400 of 1,023
LEGO-GraphRAG: Modularizing Graph-Based Retrieval-Augmented Generation for Design Space Exploration
Cao, Y., Gao, Z., Li, Z., Xie, X., Zhou, S. K., & Xu, J. (2025) · Proceedings of the VLDB Endowment · doi:10.14778/3748191.3748194 · Cited by 35
Lin, W., Blloshmi, R., Byrne, B., de Gispert, A., & Iglesias, G. (2023) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2023.acl-short.133 · Cited by 15
“In particular, we find ColBERT (Khattab and Zaharia, 2020) to be very effective, even if not pretrained for tables.”
BackgroundMethodChang, H.-S., Agarwal, N., & McCallum, A. (2024) · ACM International Conference on Web Search and Data Mining (WSDM) · doi:10.1145/3616855.3635755 · Cited by 17
“Using the grid search results, we can analyze the hyperparameter sensitivity of learning rates, dropouts, and batch sizes.”
BackgroundMethodA Deep Learning Approach for Selective Relevance Feedback
Datta, S., Ganguly, D., MacAvaney, S., & Greene, D. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-56060-6_13 · Cited by 21
ChatDL: An LLM-Based Defect Localization Approach for Software in IIoT Flexible Manufacturing
Yang, H., Zhou, Y. L., Liang, T., & Kuang, L. (2025) · IEEE Internet of Things Journal · doi:10.1109/jiot.2025.3531512 · Cited by 16
“Our model represents an enhancement over ColBERT [26] for defect localization.”
Elmoghazy, S. S., Shouman, M. A., Elminir, H. K., & Selim, G. E. I. (2025) · Artificial Intelligence Review · doi:10.1007/s10462-025-11189-8 · Cited by 14
“…(Cui et al. 2022) was the only one to address this issue through low-latency inference and reducing model size solutions such as late interaction (Khattab and Zaharia 2020), early exiting (Huang et al. 2018; Teerapittayanon et al. 2016), parameter sharing (Lan et al. 2020), and pruning (Zafrir…”
TempRetriever: Fusion-based Temporal Dense Passage Retrieval for Time-Sensitive Questions
Abdallah, A., Piryani, B., Wallat, J., Anand, A., & Jatowt, A. (2026) · ACM International Conference on Web Search and Data Mining (WSDM) · doi:10.1145/3773966.3777938 · Cited by 31
Establishing Strong Baselines For TripClick Health Retrieval
Hofstätter, S., Althammer, S., Sertkan, M., & Hanbury, A. (2022) · Lecture notes in computer science · doi:10.1007/978-3-030-99739-7_17 · Cited by 12
Chu, X., Zhao, J., Zou, L., & Yin, D. (2022) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3477495.3531986 · Cited by 11
“• ColBERT [19]: It introduces a fast BERT-based retrievalmodel under the late interaction-based mechanism.”
BackgroundMethodPersonalized Dynamic Attention Multi-task Learning model for document retrieval and query generation
Zeng, J., Yu, Y., Wen, J., Jiang, W., & Cheng, L. (2022) · Expert Systems with Applications · doi:10.1016/j.eswa.2022.119026 · Cited by 11
Improved relation span detection in question answering systems over extracted knowledge bases
Behmanesh, S., Talebpour, A., Shamsfard, M., & Jafari, M. M. M. (2023) · Expert Systems with Applications · doi:10.1016/j.eswa.2023.119973 · Cited by 12
Beyond Two-Tower: Attribute Guided Representation Learning for Candidate Retrieval
Shan, H., Zhang, Q., Liu, Z., Zhang, G., & Li, C. (2023) · International World Wide Web Conference (WWW) · doi:10.1145/3543507.3583254 · Cited by 16
“Both cross-encoder[7, 43] and late-interaction encoder[21, 21, 34] are often applied in the ranking stage.”
MethodDMFN: A disentangled multi-level fusion network for review helpfulness prediction
Ren, G., Diao, L., & Kim, J. (2023) · Expert Systems with Applications · doi:10.1016/j.eswa.2023.120344 · Cited by 12
Modeling Context in Answer Sentence Selection Systems on a Latency Budget
Han, R., Soldaini, L., & Moschitti, A. (2021) · Conference of the European Chapter of the Association for Computational Linguistics · doi:10.18653/v1/2021.eacl-main.261 · Cited by 14
Semi-Siamese Bi-encoder Neural Ranking Model Using Lightweight Fine-Tuning
Jung, E., Choi, J., & Rhee, W. (2022) · International World Wide Web Conference (WWW) · doi:10.1145/3485447.3511978 · Cited by 26
An Inspection of the Reproducibility and Replicability of TCT-ColBERT
Wang, X., MacAvaney, S., Macdonald, C., & Ounis, I. (2022) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3477495.3531721 · Cited by 20
“For model-specific design of ColBERT encoding process, we refer readers to [10] for full details.”
BackgroundMethodInfluentialRELiC: Retrieving Evidence for Literary Claims
Thai, K., Chang, Y., Krishna, K., & Iyyer, M. (2022) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2022.acl-long.517 · Cited by 28
Learning Passage Impacts for Inverted IndexesPreprint
Mallia, A., Khattab, O., Tonellotto, N., & Suel, T. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2104.12016 · Cited by 11
Self-citationCONFPILOT: A Pilot for Faster Configuration by Learning from Device Manuals
Zhao, J., Sun, H., Wang, J., Qi, Q., Zhuang, Z., Tao, S., & Liao, J. (2023) · IEEE International Conference on Distributed Computing Systems (ICDCS) · doi:10.1109/icdcs57875.2023.00077 · Cited by 13
Improving RAG Quality for Large Language Models with Topic-Enhanced Reranking
Ampazis, N. (2024) · IFIP advances in information and communication technology · doi:10.1007/978-3-031-63215-0_6 · Cited by 13
From Missteps to Mastery: Enhancing Low-Resource Dense Retrieval through Adaptive Query Generation
Tong, Z., Qin, C., Fang, C., Yao, K., Chen, X., Zhang, J., Zhu, C., & Zhu, H. (2025) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3690624.3709225 · Cited by 16
“Compared with traditional bag-of-words-based sparse retrieval models [1, 43, 56], neural network-based dense retrieval [22, 24, 39] has the potential to capture deeper semantic information.”
InfluentialEfficient Constant-Space Multi-vector Retrieval
MacAvaney, S., Mallia, A., & Tonellotto, N. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-88714-7_22 · Cited by 27
“Cross-encoder can be computationally expensive for estimating query-document similarities due to the complexity of the underlying transformer neural network [10, 11, 13, 27].”
A Comprehensive Evaluation of Embedding Models and LLMs for IR and QA Across English and Italian
Oro, E., Granata, F. M., & Ruffolo, M. (2025) · Big Data and Cognitive Computing · doi:10.3390/bdcc9050141 · Cited by 15
“Pioneering dense retrieval methods include DPR [20], ColBERT [21], and ANCE [22].”
In-Storage Acceleration of Retrieval Augmented Generation as a Service
Mahapatra, R., Santhanam, H., Priebe, C., Xu, H., & Esmaeilzadeh, H. (2025) · Proceedings - International Symposium on Computer Architecture/Proceedings · doi:10.1145/3695053.3731032 · Cited by 24
Li, X., Cai, Q., Shu, Y., Guo, C., & Yang, B. (2025) · IEEE International Conference on Data Engineering (ICDE) · doi:10.1109/icde65448.2025.00294 · Cited by 20
Boytsov, L., & Kolter, J. Z. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-72113-8_5 · Cited by 21
Modality-Balanced Embedding for Video Retrieval
Wang, X., Ke, B., Li, X., Liu, F., Zhang, M., Liang, X. M., & Xiao, Q. (2022) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3477495.3531899 · Cited by 14
Implicit Feedback for Dense Passage Retrieval: A Counterfactual Approach
Zhuang, S., Li, H., & Zuccon, G. (2022) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3477495.3531994 · Cited by 11
“Dense Retrievers (DRs) [14, 20, 22, 27, 42] utilise a BERT-based (or variation of) dual-encoder architecture to separately encode queries and passages into a shared embedding space.”
MethodAxiomatic Retrieval Experimentation with ir_axioms
Bondarenko, A., Fröbe, M., Reimer, J. H., Stein, B., Völske, M., & Hagen, M. (2022) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3477495.3531743 · Cited by 14
Scattered or Connected? An Optimized Parameter-efficient Tuning Approach for Information Retrieval
Ma, X., Guo, J., Zhang, R., Fan, Y., & Cheng, X. (2022) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3511808.3557445 · Cited by 18
“Studies have shown that leveraging the existing PTMs can benefit both the retrieval and re-ranking stages significantly [14, 15, 28, 38].”
BackgroundA novel dense retrieval framework for long document retrieval
Wang, J., Zhao, W., Tu, X., & He, T. (2022) · Frontiers of Computer Science · doi:10.1007/s11704-022-2041-5 · Cited by 10
Agarwal, S. (2022) · IEEE Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON) · doi:10.1109/uemcon54665.2022.9965655 · Cited by 10
“Figure 3: Late-Interaction ColBERT architecture Recently, [8] has highlighted the importance of pre-training for the application of language models, particularly transformers, in highly specialized contexts that require niche vocabulary.”
BackgroundMethodInfluentialAn Experimental Study on Pretraining Transformers from Scratch for IR
Lassance, C., Déjean, H., & Clinchant, S. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-28244-7_32 · Cited by 15
LADER: Log-Augmented DEnse Retrieval for Biomedical Literature Search
Jin, Q., Shin, A., & Lu, Z. (2023) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3539618.3592005 · Cited by 11
“We choose to use dense retrievers, where queries and documents are encoded and matched in low-dimensional embedding space, because recent work has shown that dense retrievers based on pre-trained language models such as BERT [6] outperform traditional sparse retrievers on various tasks [19, 21, 37].”
MethodCsFEVER and CTKFacts: acquiring Czech data for fact verification
Ullrich, H., Drchal, J., Rýpar, M., Vincourová, H., & Moravec, V. (2023) · Language Resources and Evaluation · doi:10.1007/s10579-023-09654-3 · Cited by 18
“The last tested model was a recent ColBert, which provides the benefits of both cross-attention and two-tower paradigms [43].”
MethodQuantifying Ranker Coverage of Different Query Subspaces
Arabzadeh, N., Bigdeli, A., Rad, R. H., & Bagheri, E. (2023) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3539618.3592045 · Cited by 11
An efficient long-text semantic retrieval approach via utilizing presentation learning on short-text
Wang, J., Huang, J. X., & Sheng, J. (2023) · Complex & Intelligent Systems · doi:10.1007/s40747-023-01192-3 · Cited by 11
A Fast Attention Network for Joint Intent Detection and Slot Filling on Edge Devices
Huang, L. P., Liang, S., Ye, F., & Gao, N. (2023) · IEEE Transactions on Artificial Intelligence · doi:10.1109/tai.2023.3309272 · Cited by 17
“From our perspective, although large pretrained language models such as BERT bring significant performance gains, these models also have increasing demands on computational resources and memory [21].”
BackgroundDong, Q., Liu, Y., Ai, Q., Li, H., Wang, S., Liu, Y., Yin, D., & Ma, S. (2023) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3583780.3614923 · Cited by 19
“In particular, existing PLM-based IR models can be broadly categorized into cross-encoders [40], dual-encoders [23] and lateinteraction encoders [15, 24], as shown in Figures 1(a), 1(b) and 1(c), respectively.”
BackgroundMethodInfluentialData Augmentation for Sample Efficient and Robust Document Ranking
Anand, A., Leonhardt, J., Singh, J., Rudra, K., & Anand, A. (2023) · ACM Transactions on Information Systems · doi:10.1145/3634911 · Cited by 14
“Recent approaches for ranking documents have focused heavily on contextual transformer-based models for both retrieval [29, 39] and re-ranking [11, 26, 35, 45, 77].”
BackgroundGLEN: Generative Retrieval via Lexical Index Learning
Lee, S., Choi, M., & Lee, J. (2023) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2023.emnlp-main.477 · Cited by 41
Analyzing Adversarial Attacks on Sequence-to-Sequence Relevance Models
Parry, A., Fröbe, M., MacAvaney, S., Potthast, M., & Hagen, M. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-56060-6_19 · Cited by 23
“Neural retrievers are frequently trained with a contrastive approach, where one relevant and one non-relevant document is passed to the model for a given query (either explicit [18, 29] or implicit [22, 32]).”
BackgroundMethodEvaluation of Temporal Change in IR Test Collections
Keller, J., Breuer, T., & Schaer, P. (2024) · ACM SIGIR International Conference on Theory of Information Retrieval (ICTIR) · doi:10.1145/3664190.3672530 · Cited by 15
SPLATE: Sparse Late Interaction Retrieval
Formal, T., Clinchant, S., Déjean, H., & Lassance, C. (2024) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3626772.3657968 · Cited by 14
Pérez, A., Fernández-Pichel, M., Parapar, J., & Losada, D. E. (2025) · Language Resources and Evaluation · doi:10.1007/s10579-025-09831-6 · Cited by 25
“• Late Interaction: ColBERT (Khattab and Zaharia, 2020) employs a late-interaction approach with a bag of contextualised token embeddings.”
Answer Complex Questions: Path Ranker Is All You Need
Zhang, X., Zhan, K., Hu, E., Fu, C., Luo, L., Jiang, H., Jia, Y., Yu, F., Dou, Z., Cao, Z., et al. (2021) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3404835.3462942 · Cited by 14
“However, the two-tower dense retriever possesses weak interactions between query and documents [3, 9, 14] , and suffers from the irrelevance of questions and documents.”
BackgroundLewis, P., Oğuz, B., Xiong, W., Petroni, F., Yih, S., & Riedel, S. (2022) · Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT) · doi:10.18653/v1/2022.naacl-main.226 · Cited by 15
VIRT: Improving Representation-based Text Matching via Virtual Interaction
Li, D., Yang, Y., Tang, H., Liu, J., Wang, Q., Wang, J., Xu, T., Wu, W., & Chen, E. (2022) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2022.emnlp-main.59 · Cited by 13
“Representation-based models (Khattab and Zaharia, 2020; Ni et al., 2022) encode two texts independently with siamese or dual encoders (Cer et al.”
BackgroundHybrid Learning to Rank for Financial Event Ranking
Feng, F., Li, M., Luo, C., Ng, R., & Chua, T. (2021) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3404835.3462969 · Cited by 17
“Labeled data are indispensable for the training of HNR, which largely rely on deep neural networks to encode the query and news [25, 39].”
BackgroundMethodInfluentialCo-BERT: A Context-Aware BERT Retrieval Model Incorporating Local and Query-specific ContextPreprint
Chen, X., Hui, K., He, B., Han, X., Sun, L., & Ye, Z. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2104.08523 · Cited by 15
“Similar to [13], we report FLOPs that includes all computations when using BERT-Base (MS Marco), BERT-QE, and Co-BERT during inference.”
Kazi, S., Khoja, S., & Daud, A. (2025) · Computer Science Review · doi:10.1016/j.cosrev.2025.100756 · Cited by 9
The Open Web Index: Crawling and Indexing the Web for Public Use
Hendriksen, G., Dinzinger, M., Farzana, S. M., Fathima, N. A., Fröbe, M., Schmidt, S., Zerhoudi, S., Granitzer, M., Hagen, M., Hiemstra, D., et al. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-56069-9_10 · Cited by 12
Thakur, N., Bonifacio, L., Fröbe, M., Bondarenko, A., Kamalloo, E., Potthast, M., Hagen, M., & Lin, J. (2024) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3626772.3657861 · Cited by 29
“Substantial progress has been made in developing different types of neural retrieval models, including dense (e.g., [27, 31, 37, 71, 73]), sparse (e.g., [14, 21, 40, 75]), and multi-vector models (e.g., [23, 32, 36, 55]).”
BackgroundCastells, P., & Jannach, D. (2024) · ACM eBooks · doi:10.1145/3674127.3674131 · Cited by 11
MGRAG: MultiModal grid-aware retrieval augmentation generation framework for power grid work tickets
Mao, J., Zheng, C., Liu, W., & Zhang, Z. (2025) · Pattern Recognition · doi:10.1016/j.patcog.2025.111845 · Cited by 16
Efficient Re-ranking with Cross-encoders via Early Exit
Busolin, F., Lucchese, C., Nardini, F. M., Orlando, S., Perego, R., Trani, S., & Veneri, A. (2025) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3726302.3729962 · Cited by 14
Secure Retrieval-Augmented Generation Against Poisoning Attacks
Cheng, Z., Sun, J., Gao, A., Quan, Y., Liu, Z., Hu, X. T., & Fang, M. (2025) · IEEE International Conference on Big Data (BigData) · doi:10.1109/bigdata66926.2025.11402647 · Cited by 39
Faithfulness-Aware Multi-Objective Context Ranking for Retrieval-Augmented Generation
Guan, T., Sun, S., & Chen, B. (2025) · International Conference on Artificial Intelligence, Systems and Network Security (AISNS) · doi:10.1145/3797161.3797180 · Cited by 16
“ColBERT introduced late interaction mechanisms that balance efficiency and effectiveness [9], while recent work has explored LLM-based rerankers that leverage the semantic understanding of large models [10] [11] [12].”
Montenegro, J. L. Z., & da Costa, C. A. (2022) · Journal of Healthcare Informatics Research · doi:10.1007/s41666-022-00115-0 · Cited by 12
“Similar results related to lower learning rate were observed in [49].”
ResultOn Survivorship Bias in MS MARCO
Gupta, P., & MacAvaney, S. (2022) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3477495.3531832 · Cited by 9
A Transformational Biencoder with In-Domain Negative Sampling for Zero-Shot Entity Linking
Sun, K., Zhang, R., Mensah, S., Mao, Y., & Liu, X. (2022) · Findings of the Association for Computational Linguistics: ACL (Findings of ACL) · doi:10.18653/v1/2022.findings-acl.114 · Cited by 13
“DUAL : Recently, Zhang and Stratos (2021) followed the architecture of Wu et al. (2020) and showed that the Sum-Of-Max (SOM) scorer (Khattab and Zaharia, 2020) yields better results in comparison to DUAL.”
MethodResultJoint Optimization of Multi-vector Representation with Product Quantization
Fang, Y., Zhan, J., Liu, Y., Mao, J., Zhang, M., & Ma, S. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-17120-8_49 · Cited by 9
Fine-Tuning BERT-based Language Models for Duplicate Trouble Report Retrieval
Bosch, N., Shalmashi, S., Yaghoubi, F., Holm, H., Gaim, F., & Payberah, A. H. (2022) · IEEE International Conference on Big Data (BigData) · doi:10.1109/bigdata55660.2022.10020825 · Cited by 11
Bridging the Gap between Geometry and User Intent: Retrieval of CAD Models via Regions of Interest
Herzog, V., & Suwelack, S. (2022) · Computer-Aided Design · doi:10.1016/j.cad.2023.103573 · Cited by 10
Typo-Robust Representation Learning for Dense Retrieval
Tasawong, P., Ponwitayarat, W., Limkonchotiwat, P., Udomcharoenchaikit, C., Chuangsuwanich, E., & Nutanong, S. Y. (2023) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2023.acl-short.95 · Cited by 21
Injecting the score of the first-stage retriever as text improves BERT-based re-rankers
Askari, A., Abolghasemi, A., Pasi, G., Kraaij, W., & Verberne, S. (2024) · Discover Computing · doi:10.1007/s10791-024-09435-8 · Cited by 15
“Following prior work on MSMARCO [11, 87–90], we use the dev set ( ∼ 7 k queries) for our empirical evaluation.”
Incorporating Ranking Context for End-to-End BERT Re-ranking
Chen, X., Hui, K., He, B., Han, X., Sun, L., & Ye, Z. (2022) · Lecture notes in computer science · doi:10.1007/978-3-030-99736-6_8 · Cited by 9
“Beyond that, two-tower retrievers [22,47,56], ColBERT [23], EPIC [36], TK [21], as well as PreTTR [35] pre-compute the passage representations to reduce query-time latency, and are further improved by TAS-Balanced [19], PAIR [48] and JPQ [60].”
MethodFrom Easy to Hard: A Dual Curriculum Learning Framework for Context-Aware Document Ranking
Zhu, Y., Nie, J., Su, Y., Chen, H., Zhang, X., & Dou, Z. (2022) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3511808.3557328 · Cited by 15
“Recently, pre-trained language models, such as BERT [14], have achieved promising results on several NLP and IR tasks [33, 34, 37, 38].”
BackgroundMethodStochastic Optimization of Text Set Generation for Learning Multiple Query Intent Representations
Hashemi, H., Zamani, H., & Croft, W. B. (2022) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3511808.3557666 · Cited by 10
“contextual embedding-based models [10, 42].”
One Blade for One Purpose: Advancing Math Information Retrieval using Hybrid Search
Zhong, W., Lin, S.-C., Yang, J.-H., & Lin, J. (2023) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3539618.3591746 · Cited by 21
Multi-CLS BERT: An Efficient Alternative to Traditional Ensembling
Chang, H.-S., Sun, R.-Y., Ricci, K., & McCallum, A. (2023) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2023.acl-long.48 · Cited by 24
Distillation for Multilingual Information Retrieval
Yang, E., Lawrie, D., & Mayfield, J. (2024) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3626772.3657955 · Cited by 31
Zhang, C., Wang, B., & Song, D. (2024) · ACM Transactions on Information Systems · doi:10.1145/3677375 · Cited by 8
“Typically, the initialization can be realized by borrowing structures from the teacher [59] or by complete randomness [68].”
Cap4Video++: Enhancing Video Understanding With Auxiliary Captions
Wu, W., Wang, X., Luo, H., Wang, J., Yang, Y., & Ouyang, W. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2024.3410329 · Cited by 19
“To address this issue, previous works like FILIP [74] and ColBERT [75] have used a Max-Mean pipeline to model the token-wise word-patch alignment for image-text learning.”
Clavié, B. (2025) · Journal of Natural Language Processing · doi:10.5715/jnlp.32.176 · Cited by 26
Towards Brain Passage Retrieval: An Investigation of EEG Query Representations
McGuire, N., & Moshfeghi, Y. (2025) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3726302.3730097 · Cited by 10
“…have significantly advanced the capabilities of IR systems through their ability to capture deep semantic relationships and contextual understanding beyond lexical matching [26, 28, 55, 68], they still rely on users translating their INs into explicit textual queries (as discussed in Section 1).”
Hypencoder: Hypernetworks for Information Retrieval
Killingback, J., Zeng, H., & Zamani, H. (2025) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3726302.3729983 · Cited by 19
“For example, Hypencoder could easily work with multiple vectors similar to existing multi-vector models, e.g., [33], or use training routines popularized in dense retrieval, e.g., [38, 52, 73, 76].”
InfluentialLLM-augmented multi-agent cooperative framework for medical case retrieval in cardiology
Deng, L., Hu, H., Lu, K., & He, P. (2025) · Journal of King Saud University Computer and Information Sciences · doi:10.1007/s44443-025-00311-z · Cited by 8
“ColBERT refines this with late interaction, computing fine-grained term embeddings and achieving efficient yet accurate passage search (Khattab and Zaharia 2020)’.”
MMESGBench: Pioneering Multimodal Understanding and Complex Reasoning Benchmark for ESG Tasks
Zhang, L., Zhou, X., He, C., Wang, D., Wu, Y., Xu, H. W., Liu, W., & Miao, C. (2025) · ACM International Conference on Multimedia (ACM MM) · doi:10.1145/3746027.3758225 · Cited by 26
Sharma, P., & Bhattarai, S. (2026) · Journal of Future Artificial Intelligence and Technologies · doi:10.62411/faith.3048-3719-297 · Cited by 13
Explain Like I am BM25: Interpreting a Dense Model's Ranked-List with a Sparse Approximation
Llordes, M., Ganguly, D., Bhatia, S., & Agarwal, C. (2023) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3539618.3591982 · Cited by 17
Improving First-stage Retrieval of Point-of-interest Search by Pre-training Models
Mei, L., Mao, J., Hu, J., Tan, N., Chai, H., & Wen, J.-R. (2023) · ACM Transactions on Information Systems · doi:10.1145/3631937 · Cited by 16
“After fine-tuning the parameters in various downstream tasks [22, 22, 24], it can improve the performance of domain-specific tasks.”
BackgroundMethodEfficient Neural Ranking Using Forward Indexes and Lightweight Encoders
Leonhardt, J., Müller, H., Rudra, K., Khosla, M., Anand, A., & Anand, A. (2023) · ACM Transactions on Information Systems · doi:10.1145/3631939 · Cited by 23
“Hofstätter et al. [26] introduce ColBERTer, an extension of ColBERT [35], which removes irrelevant word representations in order to reduce the number of stored vectors.”
BackgroundMosser, L., Aursand, P., Brakstad, K. S., Lehre, C. N., & Myhre-Bakkevig, J. (2024) · SPE Bergen One Day Seminar · doi:10.2118/218439-ms · Cited by 11
Learning Multi-Stage Multi-Grained Semantic Embeddings for E-Commerce Search
Wang, B., Li, M., Zeng, Z., Zhuo, J., Wang, S., Xu, S., Long, B., & Yan, W. (2023) · International World Wide Web Conference (WWW) · doi:10.1145/3543873.3584638 · Cited by 15
Web-Scale Semantic Product Search with Large Language Models
Muhamed, A., Srinivasan, S., Teo, C.-H., Cui, Q., Zeng, B., Chilimbi, T. M., & Vishwanathan, S. V. N. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-33380-4_6 · Cited by 14
Wang, X., Macdonald, C., Tonellotto, N., & Ounis, I. (2023) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3539618.3591916 · Cited by 13
Personalized Query Expansion with Contextual Word Embeddings
Bassani, E., Tonellotto, N., & Pasi, G. (2023) · ACM Transactions on Information Systems · doi:10.1145/3624988 · Cited by 17
Structured Fine-Tuning of Contextual Embeddings for Effective Biomedical Retrieval
Ueda, A., Santos, R. L. T., Macdonald, C., & Ounis, I. (2021) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3404835.3463075 · Cited by 8
On Single and Multiple Representations in Dense Passage Retrieval
Macdonald, C., Tonellotto, N., & Ounis, I. (2021) · ENLIGHTEN (Jurnal Bimbingan dan Konseling Islam) · doi:10.48550/arxiv.2108.06279 · Cited by 8
“Most BERT-based models are computationally expensive for estimating query-document similarities in ranking, due to the complexity of the underlying transformer neural network [3, 4, 5].”
BackgroundMethodLearning to Rank in the Age of Muppets: Effectiveness–Efficiency Tradeoffs in Multi-Stage Ranking
Zhang, Y., Hu, C., Liu, Y., Fang, H., & Lin, J. (2021) · Workshop on Simple and Efficient Natural Language Processing (SustaiNLP) · doi:10.18653/v1/2021.sustainlp-1.8 · Cited by 23
“For example, according to the ColBERT paper (Khattab and Zaharia, 2020), reranking 1000 hits from the MS MARCO passage dataset takes 32.”
MethodI Know What You Need: Investigating Document Retrieval Effectiveness with Partial Session Contexts
Sen, P., Ganguly, D., & Jones, G. J. F. (2021) · ACM Transactions on Information Systems · doi:10.1145/3488667 · Cited by 12
Continual Learning of Long Topic Sequences in Neural Information Retrieval
Gerald, T., & Soulier, L. (2022) · Lecture notes in computer science · doi:10.1007/978-3-030-99736-6_17 · Cited by 8
In-context Autoencoder for Context Compression in a Large Language Model
Ge, T., Hu, J., Wang, L., Wang, X., Chen, S.-Q., & Wei, F. (2023) · arXiv · doi:10.48550/arxiv.2307.06945 · Cited by 11
Steering Large Language Models for Cross-lingual Information Retrieval
Guo, P., Ren, Y., Hu, Y., Cao, Y., Li, Y., & Huang, H. (2024) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3626772.3657819 · Cited by 25
Yeganova, L. E., Kim, W. B., Tian, S., Comeau, D. C., Wilbur, W. J., & Lu, Z. (2025) · Nucleic Acids Research · doi:10.1093/nar/gkaf417 · Cited by 10
Reproducing NevIR: Negation in Neural Information Retrieval
van den Elsen, C., Barkhof, F., Nijdam, T., Lupart, S., & Aliannejadi, M. (2025) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3726302.3730294 · Cited by 16
“We tested ColBERTv1 [16] and ColBERTv2 [38], as these remain representative of the state of the art in this category.”
Effective Inference-Free Retrieval for Learned Sparse Representations
Nardini, F. M., Nguyen, T. D.-M., Rulli, C., Venturini, R., & Yates, A. (2025) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3726302.3730185 · Cited by 10
“Indeed, the possibility of producing vector representations, i.e., embeddings , of the input text by grasping semantic and contextual information is one of the most important [13, 15, 29].”
Probing Ranking LLMs: A Mechanistic Analysis for Information Retrieval
Chowdhury, T., Nijasure, A., & Allan, J. (2025) · ACM SIGIR International Conference on Theory of Information Retrieval (ICTIR) · doi:10.1145/3731120.3744603 · Cited by 16
Metadata
- Authors
- Khattab, O., & Zaharia, M.
- Affiliations
- Stanford University (US)
- Venue
- Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval
- Publisher
- ACM
- Published
- 2020-07-25
- Pages
- 39-48
- OpenAlex
- W3021397474
- Cited by
- 1,301
- References
- 21
- Citations by year
- 2026: 346 · 2025: 277 · 2024: 197 · 2023: 184 · 2022: 143 · 2021: 135 · 2020: 18
- FWCI
- 90.73
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) (publishedVersion)
- Funders
- NSF (CNS-1651570)
- Topics
- Physical Sciences › Computer Science › Information Systems › Information Retrieval and Search Behavior
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Image and Video Retrieval Techniques
- Keywords
- Colbert, BERT, document ranking, information retrieval, query-document similarity, efficient retrieval
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.