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 601–700 of 1,023
VRAG: Retrieval-Augmented Video Question Answering for Long-Form Videos
Gia, B. T., Le, K., Do, T., Mai, T.-D., Ngo, T. D., Le, D.-D., & Satoh, S. (2025) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw67362.2025.00354 · Cited by 8
“Dense retrieval techniques, such as Dense Passage Retrieval (DPR) [15] and ColBERT [17], leverage neural embeddings to retrieve semantically relevant documents, demonstrating superior performance over traditional sparse retrieval methods like BM25.”
A Survey of Long-Document Retrieval in the PLM and LLM Era
Li, M., Zhang, Y., Lv, T., Zhao, S., Luo, M., Nie, E., & Zhou, G. (2026) · ACM Transactions on Information Systems · doi:10.1145/3849382 · Cited by 8
“We emphasize benchmarks where document length, dispersed evidence, or structure/layout are intrinsic to the task.”
Query-driven Segment Selection for Ranking Long Documents
Kim, Y., Rahimi, R., Bonab, H., & Allan, J. (2021) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3459637.3482101 · Cited by 3
“For the models where the score is based on cosine similarity of embeddings between a query and documents, the maximum cosine similarity scores could be pooled and fed to final score generating layers [4, 6, 10].”
MethodA Question Answering Tool for Website Privacy Policy Comprehension
Mazzola, L., Shankar, A., Bless, C., Rodriguez, M. A., Waldis, A., Denzler, A., & Roey, M. V. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-35822-7_14 · Cited by 3
Understanding the Multi-vector Dense Retrieval Models
Liu, Q., & Mao, J. (2023) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3583780.3615282 · Cited by 3
KGPR: Knowledge Graph Enhanced Passage Ranking
Fang, J., Meng, Z., & Macdonald, C. (2023) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3583780.3615252 · Cited by 3
BERT-LBIA: A BERT-Based Late Bidirectional Interaction Attention Model for Legal Case Retrieval
Yang, B., Zhu, J., Luo, X., & Zhang, X. (2023) · Communications in computer and information science · doi:10.1007/978-981-99-8184-7_21 · Cited by 3
Context Retrieval via Normalized Contextual Latent Interaction for Conversational Agent
Liu, J., Mei, Z., Peng, K., & Vatsavai, R. R. (2023) · IEEE ... International Conference on Data Mining workshops · doi:10.1109/icdmw60847.2023.00196 · Cited by 4
“The state-of-the-art retrieval method, ColBERT [10], leverages contextual latent interactions to better capture the low-level word similarities between the query and documents.”
BackgroundMethodInfluentialEffective Adhoc Retrieval Through Traversal of a Query-Document Graph
Frayling, E., MacAvaney, S., Macdonald, C., & Ounis, I. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-56063-7_6 · Cited by 4
“Indeed, there have been several attempts to reduce the computational costs using a standalone neural ranker, i.e. dense retrieval [15,19,20,36].”
PLAID SHIRTTT for Large-Scale Streaming Dense Retrieval
Lawrie, D., Kayi, E. S., Yang, E., Mayfield, J., & Oard, D. W. (2024) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3626772.3657964 · Cited by 3
“But in multilingual tasks such as Cross-Language Information Retrieval (CLIR), it is outperformed by ColBERT [15] (our focus in this paper), where each token is represented by a dense vector.”
BackgroundOn Embedding Implementations in Text Ranking and Classification Employing Graphs
Kalogeropoulos, N.-R., Ioannou, D., Stathopoulos, D., & Makris, C. H. (2024) · Electronics · doi:10.3390/electronics13101897 · Cited by 3
Using Large Language Models for Math Information Retrieval
Mansouri, B., & Maarefdoust, R. (2024) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3626772.3657907 · Cited by 4
Hakim, Z. I. A., Sarker, N. H., Singh, R., Paul, B., Dabouei, A., & Xu, M. G. (2024) · IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW) · doi:10.1109/cvprw63382.2024.00203 · Cited by 3
Application of Multi-Way Recall Fusion Reranking Based on Tensor and ColBERT in RAG
Wang, W., Ma, J., Zhang, P., Hu, Z., Jiang, Q., & Liu, Y. (2024) · IEEE International Conference on Information Systems and Computer Aided Education (ICISCAE) · doi:10.1109/iciscae62304.2024.10761558 · Cited by 3
“ColBERT (Contextualized Late Interaction over BERT) [9] introduces a delayed interaction similarity function to calculate the similarity (MaxSim) between queries and documents.”
Open-Domain question classification and completion in conversational information search
Kia, O. M., Neshati, M., & Alamdari, M. S. (2020) · International Conference on Information and Knowledge Technology (IKT) · doi:10.1109/ikt51791.2020.9345613 · Cited by 5
“However, because such methods process each pair (query - documents) along with a vast neural network, they dramatically increase the cost of calculating the query-document relevance score [8], [9].”
MethodKalyan, V., Witteveen, S., & Andrews, M. (2021) · Workshop on Graph Based Methods for Natural Language Processing · doi:10.18653/v1/2021.textgraphs-1.20 · Cited by 3
Expanded Lattice Embeddings for Spoken Document Retrieval on Informal Meetings
Villatoro-Tello, E., Madikeri, S. R., Motlíček, P., Ganapathiraju, A., & Ivanov, A. V. (2022) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3477495.3531921 · Cited by 3
“Followingmethods serve as comparative evaluations:COLBERT [9, 16] - we use the checkpoint trained by the University of Glasgow on the MSMARCO passage ranking dataset.”
MethodMunarko, Y., Rampadarath, A., & Nickerson, D. P. (2023) · F1000Research · doi:10.12688/f1000research.128982.1 · Cited by 6
“Then, ColBERT uses a polling approach by calculating the similarity of each query token embedding with each sentence token embedding.”
BackgroundMethodInfluentialRobustness of Neural Rankers to Typos: A Comparative Study
Zhuang, S., Mao, X., & Zuccon, G. (2022) · Australasian Document Computing Symposium · doi:10.1145/3572960.3572981 · Cited by 3
“• TCT-ColBERTv2 [17]: A BERT-based dense retriever trained with ranking knowledge distillation from both a cross-encoder teacher model and a ColBERT [15] (a strong multi-vector dense retriever) teacher model.”
MethodOFAR: A Multimodal Evidence Retrieval Framework for Illegal Live-streaming Identification
Lin, D., Ma, Y., Li, Y., Song, X., Wu, J., & Nie, L. (2023) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3539618.3591864 · Cited by 3
“Later, Khattab and Zaharia [12] adopted the dual encoder architecture and introduced a new dual-attention mechanism, named MaxSim-based late interaction, for efficiently conducting the query-document matching.”
MethodInfluentialWang, J., Chen, Z., Zhang, Y., He, D., & Lin, F. (2023) · Neurocomputing · doi:10.1016/j.neucom.2023.126299 · Cited by 3
Mahendra, R., Spina, D., & Verspoor, K. M. (2023) · International Workshop on Semantic Evaluation (SemEval) · doi:10.18653/v1/2023.semeval-1.316 · Cited by 6
“, ColBERT (Khattab and Zaharia, 2020) and Dense Passage Retrieval (Karpukhin et al.”
Hybrid and Collaborative Passage Reranking
Zhang, Z., Zhou, W., Shi, J., & Li, H. (2023) · Findings of the Association for Computational Linguistics: ACL (Findings of ACL) · doi:10.18653/v1/2023.findings-acl.880 · Cited by 9
“Different from prior approaches, we extract the collaborative feature from the retrieval list, represent the query and each passages as hybrid and collaborative sequences, and measure the relevance between query and passages using these sequences from the perspective of collaboration.”
BackgroundMethodInfluentialEntity Retrieval from Multilingual Knowledge Graphs
Esmeir, S., Câmara, A., & Meij, E. (2022) · Workshop on Multilingual Representation Learning (MRL) · doi:10.18653/v1/2022.mrl-1.1 · Cited by 4
Zhang, Z., Yu, W., Zheng, N., Ju, M., & Jiang, M. (2023) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl_a_00591 · Cited by 5
Resources for Combining Teaching and Research in Information Retrieval Coursework
Fröbe, M., Scells, H., Elstner, T., Akiki, C., Gienapp, L., Reimer, J. H., MacAvaney, S., Stein, B., Hagen, M., & Potthast, M. (2024) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3626772.3657886 · Cited by 5
Contextualized Offline Relevance Weighting for Efficient and Effective Neural Retrieval
Chen, X., He, B., Hui, K., Wang, Y., Sun, L., & Sun, Y. (2021) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3404835.3463073 · Cited by 3
QAnswer: Towards Question Answering Search over Websites
Guo, K., Defretiere, C., Diefenbach, D., Gravier, C., & Gourru, A. (2022) · International World Wide Web Conference (WWW) · doi:10.1145/3487553.3524250 · Cited by 3
SPSD: Similarity-preserving self-distillation for video–text retrieval
Wang, J., Yan, H., Yang, Y., & Kou, H. (2023) · International Journal of Multimedia Information Retrieval · doi:10.1007/s13735-023-00298-1 · Cited by 3
“Inaddition,alateinteraction architecture is proposed to firstly compute the similarities between tokens of video and text elements, and then, the summation [16] or mean [17] of token-wise similarities is calculated as the relevance score for the video and text.”
BackgroundMethodAn in-depth analysis of passage-level label transfer for contextual document ranking
Rudra, K., Fernando, Z. T., & Anand, A. (2023) · Information Retrieval · doi:10.1007/s10791-023-09430-5 · Cited by 4
Graph-based Multilingual Product Retrieval in E-commerce SearchPreprint
Lu, H., Hu, Y., Zhao, T., Wu, T., Song, Y., & Yin, B. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2105.02978 · Cited by 7
“, 2018) becoming the state-of-the-art embedding method, it is adopted for various applications (Yang et al., 2019a; Yu et al., 2020; Khattab and Zaharia, 2020; Humeau et al., 2020; Chang et al., 2020).”
BackgroundLLM as HPC Expert: Extending RAG Architecture for HPC DataPreprint
Miyashita, Y., Tung, P. K. M., & Barthélemy, J. (2025) · Qeios · doi:10.32388/imkkvr · Cited by 11
Generalized Weak Supervision for Neural Information Retrieval
Lien, Y.-C., Zamani, H., & Croft, W. B. (2024) · ACM Transactions on Information Systems · doi:10.1145/3647639 · Cited by 4
“Several works [19, 33, 44, 52] were proposed based on the dense retrieval setting to move from re-ranking to ranking.”
BackgroundEvaluating the Explainability of Neural Rankers
Pandian, S., Ganguly, D., & MacAvaney, S. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-56066-8_28 · Cited by 3
“A neural ranking model (NRM) involves learning a data-driven parameterised similarity function between queries and documents [8,19,26,5,16].”
MethodWhich Neurons Matter in IR? Applying Integrated Gradients-based Methods to Understand Cross-Encoders
Vast, M., Cooten, B. V., Soulier, L., & Piwowarski, B. (2024) · ACM SIGIR International Conference on Theory of Information Retrieval (ICTIR) · doi:10.1145/3664190.3672528 · Cited by 6
“As detailed in Section 3.2, we care-fully select our baseline to compute NIG attributions by empirically validating that it erased most of the relevance signal in the original input embeddings.”
InfluentialTemporal Extraction and Retrieval
Alonso, O., & Spaniol, M. (2024) · ACM eBooks · doi:10.1145/3674127.3674137 · Cited by 2
Entity and Relationship Extraction with High-quality Spans and Enhanced Marker Strategies
Li, Q., Yao, N., & Zhou, N. (2025) · Machine Intelligence Research · doi:10.1007/s11633-024-1515-3 · Cited by 2
Dual-Aspect Noise-Based Regularization for Multi-Modal Relation Extraction in Media Posts
Sun, K., Shi, B., Mensah, S., Liu, W., & Dong, B. (2025) · IEEE Transactions on Audio Speech and Language Processing · doi:10.1109/taslpro.2025.3546851 · Cited by 2
“ThismechanismaddsaSum-of-Max(SOM)[12]based filtergatetothetraditionalattentionmechanism,enablingthe exclusionofnoisyimagesthatdonotcontributetotheaccurate prediction of relations.”
Wang, Y., Xiong, F. Q., Han, Z., Song, Q., Zhan, K., & Wang, B. (2025) · International World Wide Web Conference (WWW) · doi:10.1145/3696410.3714829 · Cited by 2
“Since the two-tower models trained independently, they cannot leverage cross-features or interactions between user and item features until the very end, which is referred to as "Late Interaction" [17].”
TITE: Token-Independent Text Encoder for Information Retrieval
Schlatt, F., Hagen, T., Potthast, M., & Hagen, M. (2025) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3726302.3730094 · Cited by 6
Towards Lossless Token Pruning in Late-Interaction Retrieval Models
Zong, Y., & Piwowarski, B. (2025) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3726302.3730100 · Cited by 6
Retrieving Tables via Inter- and Intra-Content Contrastive Representation Learning
Wu, J., & Hou, M. (2025) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3726302.3730245 · Cited by 2
Learning Query Token Importance for Effective Document Retrieval with Verbose Queries
Podder, D., Paik, J. H., & Mitra, P. (2024) · Annual Meeting of the Forum for Information Retrieval Evaluation (FIRE) · doi:10.1145/3734947.3734954 · Cited by 4
“Since we append document titles for ClueWeb09B and evaluate our model on long queries, many of which exceed 32 terms, we train the ColBERT model on the MSMARCO passage ranking triples [1] for 50,000 batches, using the parameters from [24].”
InfluentialXia, M., & Tong, G. (2025) · International Conference on Electronic Technology and Information Science (ICETIS) · doi:10.1109/icetis66286.2025.11144230 · Cited by 4
“The Requirement Semantic Retrieval Module is designed to construct a high-quality feature matrix to enable efficient matching and retrieval between requirements and knowledge base content [1] , [4] , [5] .”
Šeputis, D., Li, Y., Langerak, K., & Mihailov, S. (2025) · ACM Conference on Recommender Systems (RecSys) · doi:10.1145/3705328.3748155 · Cited by 9
Learning to rank under uncertainty: a robust neural approach
Khodabakhsh, M., & Bagheri, E. (2025) · Knowledge and Information Systems · doi:10.1007/s10115-025-02601-4 · Cited by 2
Multimedia-Aware Question Answering: A Review of Retrieval and Cross-Modal Reasoning Architectures
Raja, R., & Vats, A. (2025) · doi:10.1145/3746274.3760393 · Cited by 3
HF-RAG: Hierarchical Fusion-based RAG with Multiple Sources and Rankers
Santra, P., Ghosh, M., Ganguly, D., Basuchowdhuri, P., & Naskar, S. K. (2025) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3746252.3760942 · Cited by 9
Enhancing Document VQA Models via Retrieval-Augmented Generation
López, E., Llabrés, A., & Valveny, E. (2025) · Lecture notes in computer science · doi:10.1007/978-3-032-09368-4_6 · Cited by 3
C2T-ID: Converting Semantic Codebooks to Textual Document Identifiers for Generative Search
Zhang, Y., Zhang, R., Guo, J., Peng, W., Li, S. X., Lv, F., & Cheng, X. (2025) · Annual International ACM SIGIR Conference on Research and Development in Information Retrieval in the Asia Pacific Region (SIGIR-AP) · doi:10.1145/3767695.3769498 · Cited by 6
S$^{3}$PRank: Toward Satisfaction-Oriented Learning to Rank With Semi-Supervised Pre-Training
Li, Y., Lyu, Z., Zhang, Y., Zhang, H., Peng, T., Xiong, H., Wang, S., Kong, L., Chen, G., & Yin, D. (2025) · IEEE Transactions on Knowledge and Data Engineering · doi:10.1109/tkde.2025.3623607 · Cited by 17
“Representative models include CAR [15], MORES [19], and BERT-based rankers [16], as well as variants such as ColBERT [20].”
Smart Routing for Multimodal Video Retrieval: When to Search What
Rosa, K. D. (2025) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw69036.2025.00434 · Cited by 3
Improving LLM Retrieval with GraphRAG-SF: Semantic Filtering for Graph-Based RAG Systems
Hosseini, M.-P., Shah, A., Miao, C., & Wei, W. (2025) · IEEE International Conference on Big Data (BigData) · doi:10.1109/bigdata66926.2025.11401790 · Cited by 2
TurkColBERT: A Benchmark of Dense and Late-Interaction Models for Turkish Information Retrieval
Ezerceli, Ö., Hussieni, M. E., Taş, S., Bayraktar, R., Terzioğlu, F. B., Çelebi, Y., & Asker, Y. (2026) · Procedia Computer Science · doi:10.1016/j.procs.2026.01.048 · Cited by 3
Arabic Legal Information Retrieval: The Impact of Morphological Segmentation and Semantic Embeddings
AboAsal, R., Montasser, S., Eldin, F. H., Abdelwahab, A., & Abdelazim, H. (2026) · Procedia Computer Science · doi:10.1016/j.procs.2026.01.034 · Cited by 2
LedgerRAG: Governance-Driven Agentic Chain of Retrieval for Dynamic Knowledge Scenarios
Wang, S., Zhang, Y., Guo, Y., & Kang, J. (2026) · Electronics · doi:10.3390/electronics15071376 · Cited by 3
Posokhov, P., Skrylnikov, S., Masliukhin, S. M., Zavgorodniaia, A., Makhnytkina, O. V., & Matveev, Y. N. (2026) · Machine Learning and Knowledge Extraction · doi:10.3390/make8040091 · Cited by 3
Qiu, B., Chen, H., Wu, Y., Zan, C., Wei, C., Zhang, W., & Zeng, X. (2026) · International World Wide Web Conference (WWW) · doi:10.1145/3774904.3792812 · Cited by 7
BERT-TSR: A BERT-based two-stage ranking model for legal case retrieval
Li, Y., Zhu, J., & Luo, X. (2026) · Artificial Intelligence and Law · doi:10.1007/s10506-026-09513-w · Cited by 2
Your LLM Has a Passport: Investigating Brand Origin Geographic Bias in E-Commerce Rankings
Betello, F., Missault, P., Czinczoll, T., & Beliuha, M. (2026) · Conference on Fairness, Accountability and Transparency · doi:10.1145/3805689.3806438 · Cited by 2
“The evolution of deep learning models, from early CNNs and RNNs to Transformer-based architectures, e.g. BERT [9], and efficiency-oriented approaches such as ColBERT [25] and SPLADE [14], has consistently driven major advances in IR.”
Faridi, A. R., Masood, F., Keshvi, & Yunus, T. (2026) · Wiley Interdisciplinary Reviews Data Mining and Knowledge Discovery · doi:10.1002/widm.70122 · Cited by 3
“Late interaction models like ColBERT tend to provide a favorable quality versus latency trade off compared to single vector dense retrieval, but they complicate indexing and serving (Khattab and Zaharia 2020).”
From Rows to Reasoning: A Retrieval-Augmented Multimodal Framework for Spreadsheet Understanding
Gulati, A., Sen, S., Sarguroh, W., & Paul, K. (2026) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3770855.3817962 · Cited by 9
Multi-Layer Contextual Passage Term Embedding for Ad-Hoc Retrieval
Cai, W., Hu, Z., Luo, Y., Liang, D., Feng, Y., & Chen, J. (2022) · Information · doi:10.3390/info13050221 · Cited by 3
“Compared with DRMM, using title queries brings a 13.45% improvement on nDCG@20 metric, and a 9.68% improvement on P@20 metric; using the description queries brings a 22.88% improvement and 21.56% improvement on nDCG@20 and P@20, respectively.”
MethodInfluentialSpaceQA: Answering Questions about the Design of Space Missions and Space Craft Concepts
García-Silva, A., Berrío, C., Gómez-Pérez, J. M., Martinez-Heras, J. A., Donati, A., & Roma, I. (2022) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3477495.3531697 · Cited by 5
“ColBERT [17] adapts BERT for efficient retrieval by adopting a late interaction architecture to compare bag of embeddings representing the question and the passage.”
MethodExploring Retriever-Reader Approaches in Question-Answering on Scientific Documents
Nguyen, D.-H., Le, N.-K., & Nguyen, L.-M. (2022) · Communications in computer and information science · doi:10.1007/978-981-19-8234-7_30 · Cited by 3
MIRS: [MASK] Insertion Based Retrieval Stabilizer for Query Variations
Liu, J., Gong, M., Hu, X., Yang, J., & Guo, Y. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-39847-6_31 · Cited by 2
RePair: An Extensible Toolkit to Generate Large-Scale Datasets for Query Refinement via Transformers
Narayanan, Y. L., & Fani, H. (2023) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3583780.3615129 · Cited by 3
Multi-perspective thought navigation for source-free entity linking
Peng, B., He, W. P., Chen, B., Villavicencio, A., & Wu, C. (2024) · Pattern Recognition Letters · doi:10.1016/j.patrec.2023.12.020 · Cited by 2
Conversational Search with Tail Entities
Tran, H. D., Yates, A., & Weikum, G. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-56060-6_20 · Cited by 3
EASE-DR: Enhanced Sentence Embeddings for Dense Retrieval
Zhou, X., Gao, Y., Jie, X., Cai, X., Bu, J., & Wang, H. (2024) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3626772.3657925 · Cited by 4
“TCT-ColBERT [15], another dense retrieval model, employs knowledge distillation, where a multi-vector dense retrieval model (e.g., Colbert [9]) serves as the teacher model.”
Enhancing Multi-field B2B Cloud Solution Matching via Contrastive Pre-training
Chen, H., Dou, Z., Hao, X., Tao, Y., Song, S., & Sheng, Z. (2024) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3637528.3671513 · Cited by 2
CPLCS: Contrastive Prompt Learning-based Code Search with Cross-modal Interaction Mechanism
Zhang, Y., Liu, Y., Fan, X., & Lu, Y. (2024) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn60899.2024.10650201 · Cited by 4
“Taking inspiration from ColBERT [21], we design a lightweight cross-modal interaction mechanism for code search.”
MethodTRGNN: Text-Rich Graph Neural Network for Few-Shot Document Filtering
Mu, H., Zhang, S., Wang, Y., Sun, Y., & Xu, H. (2024) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn60899.2024.10650066 · Cited by 2
Replicability Measures for Longitudinal Information Retrieval Evaluation
Keller, J., Breuer, T., & Schaer, P. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-71736-9_16 · Cited by 4
“The BM25 [23] ranking function is used as the baseline and first-stage ranker for the advanced systems colBERT [20] and monoT5 [22].”
MethodIntegrating Event Elements for Chinese-Vietnamese Cross-Lingual Event Retrieval
Huang, Y., Yuanlin, Y. A. N. G., Zhu, E., Liang, Y., & Xian, Y. (2024) · IEICE Transactions on Information and Systems · doi:10.1587/transinf.2024edp7055 · Cited by 2
“Omar Khattab et al. [15] propose the colBERT model, which employs the pre-trained language model BERT to represent the query and retrieval texts.”
MethodA Sparsifier Model for Efficient Information Retrieval
Dobrynin, V., Sherman, M., Abramovich, R., & Платонов, А. В. (2024) · IEEE International Conference on Application of Information and Communication Technologies (AICT) · doi:10.1109/aict61888.2024.10740301 · Cited by 2
“There are solutions such as ColBERT(end-to-end) [1] that implement one-step search and significantly improve the quality of the search.”
Ferro, N., & Maistro, M. (2024) · ACM eBooks · doi:10.1145/3674127.3674132 · Cited by 5
Privacy in Information Retrieval
Soboroff, I. M. (2024) · ACM eBooks · doi:10.1145/3674127.3674139 · Cited by 2
Zero-Shot Response-Aware Query Expansion Method for Conversational Retrieval
Wang, J., Chen, X., He, P., Zhang, F., Wang, L., & Sheng, J. (2024) · IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology (WI-IAT) · doi:10.1109/wi-iat62293.2024.00101 · Cited by 2
Zhang, X., Hu, B., Liu, S., Sun, Q., & Chen, L. (2025) · Applied Sciences · doi:10.3390/app15137517 · Cited by 2
Nguyen, H.-L., Nguyen, D.-M., Nguyen, Q.-A., Chu, H.-Q., Nguyen, H.-D., Nguyen, K.-H., Pham, C.-M., Vuong, T.-H.-Y., & Nguyen, H.-T. (2024) · International Conference on Knowledge and Systems Engineering/Proceedings · doi:10.1109/kse63888.2024.11063556 · Cited by 2
“Therefore, we adopted ColBERTv2 [16] based on ColBERT [17], a model that employs several advanced techniques to achieve state-of-the-art performance in retrieval tasks.”
SentireCache: Accelerate Sentiment Classification With Saliency-Based Caching
Zhu, Y., Jia, J., Dong, M., & Qi, J. (2025) · IEEE Transactions on Affective Computing · doi:10.1109/taffc.2025.3578574 · Cited by 3
Colflor: Towards Bert-Size Vision-Language Document Retrieval Models
Masry, A., & Hoque, E. (2025) · Machine learning for signal processing ... · doi:10.1109/mlsp62443.2025.11204231 · Cited by 3
“ColPali also incorporates the late-interaction mechanism MaxSim, originally proposed in ColBERT [2], which enables fine-grained matching between query and document representations.”
InfluentialTREC 2020 Notebook: CAsT Track.
Lin, S.-C., Yang, J.-H., & Lin, J. (2020) · Text REtrieval Conference · Cited by 2
Multi-task Learning Based Keywords Weighted Siamese Model for Semantic Retrieval
Kuang, M., Chen, Z., Wang, W., Kang, L., Yan, Q., Tang, M., & Hao, P. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-33380-4_7 · Cited by 2
Bridging the Training-Inference Gap for Dense Phrase Retrieval
Kim, G., Lee, J., Oğuz, B., Xiong, W., Zhang, Y., Mehdad, Y., & Wang, W. Y. (2022) · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP) · doi:10.18653/v1/2022.findings-emnlp.272 · Cited by 2
Privacy-aware document retrieval with two-level inverted indexing
Qiao, Y., Ji, S., Wang, C., Shao, J., & Yang, T. (2023) · Information Retrieval · doi:10.1007/s10791-023-09428-z · Cited by 2
Tran, N., & Litman, D. J. (2022) · Annual Meeting of the Special Interest Group on Discourse and Dialogue · doi:10.18653/v1/2022.sigdial-1.36 · Cited by 4
Frame-oriented Summarization of Argumentative Discussions
Syed, S., Ziegenbein, T., Heinisch, P., Wachsmuth, H., & Potthast, M. (2023) · Annual Meeting of the Special Interest Group on Discourse and Dialogue · doi:10.18653/v1/2023.sigdial-1.10 · Cited by 14
NAIL: Lexical Retrieval Indices with Efficient Non-Autoregressive Decoders
Soares, L. B., Gillick, D., Cole, J. R., & Kwiatkowski, T. (2023) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2023.emnlp-main.156 · Cited by 5
ECNU-LLM@CHIP-PromptCBLUE: Prompt Optimization and In-Context Learning for Chinese Medical Tasks
Zheng, H., Guan, M., Mei, Y., Li, Y., & Wu, Y. (2024) · Communications in computer and information science · doi:10.1007/978-981-97-1717-0_5 · Cited by 4
Shallow Cross-Encoders for Low-Latency Retrieval
Petrov, A. V., MacAvaney, S., & Macdonald, C. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-56063-7_10 · Cited by 2
“These models separately encode the query and the document, and then estimate relevance score using an inexpensive operation over the encoded representations (e.g. cosine similarity [30] or the MaxSim operation [15]).”
Methodindxr: A Python Library for Indexing File Lines
Bassani, E., & Tonellotto, N. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-56069-9_27 · Cited by 2
Beneath the [MASK]: An Analysis of Structural Query Tokens in ColBERT
Giacalone, B., Paiement, G., Tucker, Q., & Zanibbi, R. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-56063-7_35 · Cited by 2
“The ColBERT [4] retrieval model uses BERT [2] to produce token embeddings for document and query passages.”
BackgroundMethodZhou, Q., Hou, Y., Zhou, R., Li, Y., Wang, J., Wu, Z., Li, H.-W., & Weng, T. (2024) · Connection Science · doi:10.1080/09540091.2024.2325474 · Cited by 3
“Inspired by recent work ColBERT (Khattab & Zaharia, 2020), which design token-wise intersection for text-video retrieval (TVR), we propose an adaptive approach to solve the context matching problem by adjusting the weight magnitude of each token as follows: where e v = g v ( x ) and e t = g w ( y )…”
Special Section on Efficiency in Neural Information Retrieval
Bruch, S., Lucchese, C., Maistro, M., & Nardini, F. M. (2024) · ACM Transactions on Information Systems · doi:10.1145/3641203 · Cited by 2
“Late-interaction multi-vector models, such as ColBERT [10] and COIL [9], achieve state-of-the-art retrieval effectiveness by using all token embeddings to represent documents and queries while modeling their relevance with a sum-of-max operation.”
MethodBagFormer: Better cross-modal retrieval via bag-wise interaction
Hou, H., Yan, X., & Zhang, Y. (2024) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2024.108912 · Cited by 6
Gienapp, L., Deckers, N., Potthast, M., & Scells, H. (2025) · ACM SIGIR International Conference on Theory of Information Retrieval (ICTIR) · doi:10.1145/3731120.3744594 · Cited by 2
Vector Storage Based Long-term Memory Research on LLM
Li, K., Xin, J. L., & Jing, C. (2024) · International Journal of Advanced Network Monitoring and Controls · doi:10.2478/ijanmc-2024-0029 · Cited by 8
“For each text fragment, this research uses CoIBERT [10] as the encoder model to pre-code it into a vector representation, and its vector representation can be expressed in Equation 1: Where E is the function used to convert text segment into vectors, and v i is the first i vector representation of…”
Method
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.