Pyserini: A Python Toolkit for Reproducible Information Retrieval Research with Sparse and Dense Representations
Open accessThe DOI matches its Crossref record; 39 of the 41 references it cites were found in a registry or an index.
Abstract
Pyserini is a Python toolkit for reproducible information retrieval research with sparse and dense representations. It aims to provide effective, reproducible, and easy-to-use first-stage retrieval in a multi-stage ranking architecture. Our toolkit is self-contained as a standard Python package and comes with queries, relevance judgments, pre-built indexes, and evaluation scripts for many commonly used IR test collections. We aim to support, out of the box, the entire research lifecycle of efforts aimed at improving ranking with modern neural approaches. In particular, Pyserini supports sparse retrieval (e.g., BM25 scoring using bag-of-words representations), dense retrieval (e.g., nearest-neighbor search on transformer-encoded representations), as well as hybrid retrieval that integrates both approaches. This paper provides an overview of toolkit features and presents empirical results that illustrate its effectiveness on two popular ranking tasks. Around this toolkit, our group has built a culture of reproducibility through shared norms and tools that enable rigorous automated testing.
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
“Interestingly, the classic BM25 model [18] remains competitive, surpassing some semantic models trained on search behavior datasets.”
From Generative Agents Navigating Digital Libraries (2025) · cites this paper as [18]
In the citing paper
“Sparse lexical methods, from classical BM25 [Lin et al., 2021] to lightweight full-text indexes such as SQLite FTS5, complement dense retrieval by matching exact terms.”
From MemX: A Local-First Long-Term Memory System for AI Assistants (2026) · cites this paper as Lin et al., 2021
Citations
Cites
PyTorch: An Imperative Style, High-Performance Deep Learning Library
Paszke, A., Gross, S., Massa, F., Lerer, A., Bradbury, J., Chanan, G., Killeen, T., Lin, Z., Gimelshein, N., Antiga, L., et al. (2019) · arXiv · doi:10.48550/arxiv.1912.01703 · Cited by 31,473 · cited 1×
“This is particularly the case for work on deep learning today, since the major toolkits (PyTorch [29] and Tensorflow [1]) have both adopted Python as their front-end This work is licensed under a Creative Commons Attribution International 4.”
BackgroundTensorFlow: A system for large-scale machine learning
Abadi, M., Barham, P., Chen, J., Chen, Z., Davis, A., Dean, J., Devin, M., Ghemawat, S., Irving, G., Isard, M., et al. (2016) · arXiv (Cornell University) · doi:10.48550/arxiv.1605.08695 · Cited by 11,822
Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
Raffel, C., Shazeer, N., Roberts, A., Lee, K., Narang, S., Matena, M., Zhou, Y., Li, W., & Liu, P. J. (2019) · arXiv · doi:10.48550/arxiv.1910.10683 · Cited by 10,929 · cited 1×
“also provides document expansion baselines using our doc2query method [28]; the latest model uses T5 [33] as described in Nogueira and Lin [26].”
MethodTransformers: State-of-the-Art Natural Language Processing
Wolf, T., Debut, L., Sanh, V., Chaumond, J., Delangue, C., Moi, A., Cistac, P., Rault, T., Louf, R., Funtowicz, M., et al. (2020) · Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing: System Demonstrations · doi:10.18653/v1/2020.emnlp-demos.6 · Cited by 8,741
Billion-scale similarity search with GPUs
Johnson, J., Douze, M., & Jegou, H. (2021) · IEEE Transactions on Big Data · doi:10.1109/tbdata.2019.2921572 · Cited by 4,228
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 3×
“Row (2a) refers to brute-force scans over the document vectors in Faiss [12], which provides exact nearest-neighbor search.”
BackgroundMethodDense passage retrieval for open-domain question answering
Karpukhin, V., Oguz, B., Min, S., Lewis, P., Wu, L., Edunov, S., Chen, D., & Yih, W.-T. (2020) · Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2020.emnlp-main.550 · Cited by 3,575
MS 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
Malkov, Y. A., & Yashunin, D. A. (2020) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2018.2889473 · Cited by 1,904 · cited 3×
“To support dense and hybrid retrieval, Pyserini integrates Facebook’s Faiss library for efficient similarity search over dense vectors [12], which in turns integrates the HNSW library [23] to support low-latency querying.”
MethodColBERT: Efficient and Effective Passage Search via Contextualized Late Interaction over BERT
Khattab, O., & Zaharia, M. (2020) · Proceedings of the 43rd International ACM SIGIR Conference on Research and Development in Information Retrieval · doi:10.1145/3397271.3401075 · Cited by 1,301 · cited 8×
“Row (3a) shows the results of combining TCT-ColBERT with BM25 bag-of-words search over the original texts, while row (3b) shows results that combine document expansion using doc2query with the T5 model.”
MethodInfluential
Research works printed but not linked: 6
- Sebastian Hofstatter, Sophia Althammer, Michael Schröder, Mete Sertkan, and Allan Hanbury. 2020. Improving Efficient Neural Ranking Models with Cross-Architecture Knowledge Distillation. arXiv:2010.02666 (2020). · arXiv
- Sebastian Hofstätter, Sophia Althammer, Michael Schröder, Mete Sertkan, and Allan Hanbury. 2021. Improving Efficient Neural Ranking Models with Cross-Architecture Knowledge Distillation. arXiv:2010.02666 (2021). · arXiv
- Sean MacAvaney, Andrew Yates, Sergey Feldman, Doug Downey, Arman Cohan, and Nazli Goharian. 2021. Simplified Data Wrangling with textttir_datasets. arXiv:2103.02280 (2021). · arXiv
- Ronak Pradeep, Rodrigo Nogueira, and Jimmy Lin. 2021 c. The Expando-Mono-Duo Design Pattern for Text Ranking with Pretrained Sequence-to-Sequence Models. arXiv:2101.05667 (2021). · arXiv
- Chenyan Xiong, Zhenghao Liu, Si Sun, Zhuyun Dai, Kaitao Zhang, Shi Yu, Zhiyuan Liu, Hoifung Poon, Jianfeng Gao, and Paul Bennett. 2020 a. CMT in TREC-COVID Round 2: Mitigating the Generalization Gaps from Web to Special Domain Search. arXiv:2011.01580 (2020). · arXiv
- Yongze Yu, Jussi Karlgren, Hamed Bonab, Ann Clifton, Md Iftekhar Tanveer, and Rosie Jones. 2020. Spotify at the TREC 2020 Podcasts Track: Segment Retrieval. In Proceedings of the Twenty-Ninth Text REtrieval Conference (TREC 2020).
Cited by
Items 501–600 of 809
AgAsk: A Conversational Search Agent for Answering Agricultural Questions
Li, H., Koopman, B., Mourad, A., & Zuccon, G. (2023) · ACM International Conference on Web Search and Data Mining (WSDM) · doi:10.1145/3539597.3573034 · Cited by 2
“For the BM25 index, we used Pyserini [8] to index one passage at a time in a standard inverted file index.”
InfluentialEfficient In-Memory Inverted Indexes: Theory and Practice
Mackenzie, J., MacAvaney, S., Mallia, A., & Siedlaczek, M. (2025) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3726302.3731688 · Cited by 2
“…advances in retrieval – particularly the rise of learned sparse retrieval (LSR) models [1, 4, 9, 14, 19, 26, 36], hybrid search pipelines [5, 7, 11, 15], and retrieval-augmented generation (RAG) applications [10] – have renewed interest in efficient sparse retrieval techniques that can be…”
Impact of Shallow vs. Deep Relevance Judgments on BERT-based Reranking Models
Iturra-Bocaz, G., Vo, D., & Galuščáková, P. (2025) · ACM SIGIR International Conference on Theory of Information Retrieval (ICTIR) · doi:10.1145/3731120.3744602 · Cited by 2
Liu, L., & Zhang, M. (2025) · Findings of the Association for Computational Linguistics: ACL (Findings of ACL) · doi:10.18653/v1/2025.findings-acl.8 · Cited by 2
Study on LLMs for Promptagator-Style Dense Retriever Training
Gwon, D., Jedidi, N., & Lin, J. (2025) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3746252.3760960 · Cited by 2
Self-citationUpcycling Candidate Tokens of Large Language Models for Query Expansion
Kim, J., Cho, S., Jeong, S., Kim, S., & Cho, S. (2025) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3746252.3760895 · Cited by 2
DispatchQA: A Benchmark for Small Function Calling Language Models in E-Commerce Applications
Daiber, J., Maricato, V., Sinha, A., & Rabinovich, A. (2025) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2025.emnlp-industry.154 · Cited by 2
“To simulate realistic search systems, we use the Pyserini toolkit (Lin et al., 2021), which provides the ability to create search systems using repeatable explicit recipes.”
RankSteer: Can Pointwise LLM Rankers Be Calibrated at the Representation Level?
Wang, Y., Chen, C., & Verberne, S. (2026) · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP) · doi:10.18653/2026.findings-emnlp.1189 · Cited by 2
Improving Document Retrieval Coherence for Semantically Equivalent Queries
Campese, S., Moschitti, A., & Lauriola, I. (2025) · International Joint Conference on Natural Language Processing and Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics (IJCNLP-AACL) · doi:10.18653/v1/2025.ijcnlp-long.182 · Cited by 2
RoutIR: Fast Serving of Retrieval Pipelines for Retrieval-Augmented Generation
Yang, E., Yates, A., Lawrie, D., Mayfield, J., & Adriaanse, T. V. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-21321-1_60 · Cited by 2
RAGTurk: Best Practices for Retrieval Augmented Generation in Turkish
Köse, S. K., Baytekin, M. C., Aktaş, B., Görür, B. K., Munis, E. A., Yılmaz, D., Kartal, M. Y., & Toraman, Ç. (2026) · Workshop on Natural Language Processing for Turkic Languages (SIGTURK) · doi:10.18653/v1/2026.sigturk-1.15 · Cited by 2
Controlling Gender Bias in Retrieval via a Backpack Architecture
Afzali, A., Velae, A., Ahmadi, I., & Aliannejadi, M. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-21300-6_47 · Cited by 2
A critical evaluation of generative query expansion on biomedical literature retrieval
Fang, Y., Zhang, G., Chen, F., Peng, Y., & Weng, C. (2026) · Journal of the American Medical Informatics Association · doi:10.1093/jamia/ocag037 · Cited by 2
“We used the BM25 implementation provided by Pyserini 39 with default parameter settings.”
InfluentialEnhancing conversational recommendations through contrastive user preference modeling
Kumar, B., Raj, S., Anand, M., Verma, V. K., Tripathi, P., & Jha, A. (2026) · Discover Artificial Intelligence · doi:10.1007/s44163-026-00912-1 · Cited by 2
“In these approaches, the dialogue is treated as a query while items are represented as documents, enabling the use of text-matching techniques such as BM25 [26], thereby ensuring high scalability and strong generalizability.”
LongRanker: Efficient One-Pass Document Reranking with Long-Context Large Language Models
Zhou, C., Zhang, R., Guo, J., de Rijke, M., Yixing, F., & Cheng, X. (2026) · International World Wide Web Conference (WWW) · doi:10.1145/3774904.3792179 · Cited by 2
“We employ Pyserini [14] to implement BM25.”
Natural language processing and computational linguistics for Vietnamese: A comprehensive review
Tran, K. V., Thai, T. M., Luu, S. T., Nguyen, K. V., & Nguyen, N. L. T. (2026) · Expert Systems with Applications · doi:10.1016/j.eswa.2026.132696 · Cited by 2
Medical multimodal entity linking under modality missingness
Zang, S., Bu, C., Hong, Y., Ren, H., & Ding, W. (2026) · Artificial Intelligence in Health · doi:10.36922/aih026070015 · Cited by 2
Nazi, Z. A., Hristidis, V., McLean, A. L., Meem, J. A., & Chowdhury, M. T. A. (2026) · Data Mining and Knowledge Discovery · doi:10.1007/s10618-026-01220-z · Cited by 2
“As a lexical baseline, we use the BM25 retriever implemented via the Pyserini toolkit (Lin et al. 2021).”
NanoKnow: How to Know What Your Language Model Knows
Gu, L., Jedidi, N., & Lin, J. (2026) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3805712.3808604 · Cited by 2
“The FineWeb-Edu corpus was chosen in our experiments primarily due to the recent release of nanochat [10], a family of small LLMs that utilized it for pre-training.”
Self-citationMala, C. S., Gezici, G., & Giannotti, F. (2026) · Machine Learning · doi:10.1007/s10994-026-07121-y · Cited by 2
“Retrieval methods can be broadly divided into two paradigms: sparse and dense retrievers (Lin et al., 2021).”
CELI: Simple yet Effective Approach to Enhance Out-of-Domain Generalization of Cross-Encoders.
Zhang, C., Li, M., & Lin, J. (2024) · Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL-HLT) · doi:10.18653/v1/2024.naacl-short.16 · Cited by 2
“7 The MS MARCO results reranks the top-1k candidates from TCT-ColBERT, and the BEIR results reranks the top-1k candidates from BM25.”
MethodSelf-citationZero-shot Document Retrieval with Hybrid Pseudo-document Retriever
Sun, D. S., Guo, W., Liu, X., Zhang, Y., Hou, Z., & Li, Z. (2025) · IEEE International Conference on Acoustics Speech and Signal Processing · doi:10.1109/icassp49660.2025.10889897 · Cited by 2
“Our experiments were based on the Pyserini [12] retrieval toolkit, using Java 11 and Ubuntu 20.04.”
InfluentialDUNKS: Chunking and Summarizing Large and Heterogeneous Data for Dataset Search
Chen, Q., Zhou, X., Zhang, Z., & Cheng, G. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-77850-6_5 · Cited by 2
Enhancing RAG’s Retrieval via Query Backtranslations
Rajaei, D., Taheri, Z., & Fani, H. (2024) · Lecture notes in computer science · doi:10.1007/978-981-96-0579-8_20 · Cited by 2
“We have applied two information retrieval methods, namely bm25 [56] and qld [50], using pyserini [37] to retrieve relevant content for the original queries as well as the backtranslated versions and evaluate the retrieval performances based on two metrics, i.e., map , mrr , and ndcg , using…”
MethodSplits! Flexible Sociocultural Linguistic Investigation at Scale
Caplan, E., Chakraborty, T., & Goldwasser, D. (2026) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2026.acl-long.117 · Cited by 2
“Then, for every topic t , and two demographics A, B , the two post sets C A,t and C B,t were combined and indexed together using the pyserini BM25 implementation (Lin et al., 2021).”
Enhancing queries for code generation with reinforcement learning
Yuan, D., Liang, G., Li, T., & Liu, S. (2025) · Scientific Reports · doi:10.1038/s41598-025-21271-4 · Cited by 2
Another Look at Information Retrieval as Statistical Translation
Liu, Y., Hu, C., & Lin, J. (2022) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3477495.3531717 · Cited by 2
“In our implementation, k candidates in the first stage are retrieved using the Pyserini IR toolkit [18], which is built on the open-source Lucene search library.”
MethodSelf-citationZhang, X., Li, M., & Lin, J. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2302.06589 · Cited by 2
“Retrievers results are produced using one of Pyserini [10], BEIR [18] repository, and ColBERT [8] repository.”
MethodSelf-citationHLTCOE at TREC 2023 NeuCLIR TrackPreprint
Yang, E., Lawrie, D., & Mayfield, J. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2404.08118 · Cited by 2
“The Patapsco framework [1] supports CLIR lexical retrieval through Pyserini [10].”
MethodYoon, S.-Y., Kim, J., & Hwang, S.-W. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2407.06716 · Cited by 2
“For the implementaiton, we utilized python library castorini/pyserini [30].”
MethodInteractive Information Need Prediction with Intent and ContextPreprint
Ros, K., Pandya, D., & Zhai, C. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2501.02635 · Cited by 2
“To simulate the "Source" of each query, we performed a keyword-based search of the query over the corpus using BM25 [37] via Pyserini [24], and found a set of paragraphs that matched the query in topic but did not answer the query.”
Exploring new Approaches for Information Retrieval through Natural Language ProcessingPreprint
Raj, M. K., & Mishra, N. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2505.02199 · Cited by 2
ModelTables: A Corpus of Tables about ModelsPreprint
Dong, Z., Zhong, V., & Miller, R. J. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2512.16106 · Cited by 2
“For sparese retrieval, we use Pyserini [34] sparse-term index with 1024-clause truncation, and Hybrid Retrieval, which first retrieves the top-100 candidates via sparse search and then applies dense Sentence-BERT re-ranking.”
User Intent and State Modeling in \nConversational Systems
Ye, F. (2024) · UCL Discovery (University College London) · Cited by 2
Designing accurate retrieval systems using language models
Sachan, D. S. (2024) · eScholarship@McGill (McGill) · doi:10.82308/38920 · Cited by 2
Chang, Y.-C., Yeo, G.-W., Eugene, Q., Shih, F.-J., Kuo, Y.-C., Yu, T.-E., Hsu, H.-C., Tsai, M.-F., & Wang, C.-J. (2025) · arXiv (Cornell University) · doi:10.48550/arxiv.2509.15588 · Cited by 2
“Similar retrievers have been widely used in prior works [13].”
Domain-Specific Text Embedding Models for Information Retrieval
Shiraee Kasmaee, A. (2025) · MacSphere (McMaster University) · doi:10.71548/20804 · Cited by 2
A Systematic Study of Pseudo-Relevance Feedback with LLMsPreprint
Jedidi, N., & Lin, J. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2603.11008 · Cited by 2
Self-citationLi, M., & Zhou, G. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2603.13776 · Cited by 2
“For the TREC DL series (DL19/DL20/DL21), we follow the default evaluation protocol of Pyserini [23]: nDCG @ 10 is computed using graded relevance; while MAP and MRR adopt binary relevance, with the relevance threshold set to 𝑙=2 (i.e., qrels with relevance ≥ 2 are regarded as relevant).”
Paramés-Estévez, S., Filloy-Montesino, N., Fernández-Fabeiro, J., & Mouriño, J. (2026) · arXiv (Cornell University) · doi:10.5281/zenodo.19591020 · Cited by 2
“Consequently, hybrid approaches that combine BM25, a classical retrieval algorithm based on term frequencies, with dense retrieval [9] have become the practical state of the art.”
DeepRefine: Agentic Knowledge Refinement via Reinforcement LearningPreprint
Huang, H., Bai, J., Liu, S., Wei, Y., Jing, H., Tsang, H. T., Gao, Y., Xie, Z., Li, Y., & Song, Y. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2605.10488 · Cited by 2
“The GenAcc is defined as follows: which is a combination of span-matching test [Ma et al., 2021, Lin et al., 2021] and a LLM-based correctness check with DeepSeek-V3 [Liu et al., 2024] as a judge.”
AgentIR: A Workload-Adaptive Cascade Retrieval Substrate for Long-Term Conversational MemoryPreprint
Yuan, A., Zhang, H., & Nazarian, S. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2605.25092 · Cited by 2
“(1) CPU evaluation (Tables 4, 6 The primary baseline is Pyserini 0.22.1 [15] (Java Lucene 9 with English analyzer), the de-facto BM25 reference in the IR community.”
InfluentialCMASS: A ClimbMix Agentic Search Suite for Evaluating Search Agents Across Heterogeneous TasksPreprint
Sharifymoghaddam, S., Gu, L., Ge, Y., & Lin, J. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2608.20317 · Cited by 2
“We index ClimbMix with Pyserini [Lin et al., 2021] and serve its BM25 index ( climbmix-400b ) from a Pyserini HTTP server.”
Self-citationDhara Bench: An Evidence-Oriented Retrieval Benchmark for Indian Knowledge SystemsPreprint
Nagendra, A. H., Prasad, A. K. S., Sridharan, K. S., & Raghavan, M. (2026) · arXiv (Cornell University) · doi:10.48550/arxiv.2611.03196 · Cited by 2
Generative Agents Navigating Digital Libraries
Zerhoudi, S., & Granitzer, M. (2025) · Lecture notes in computer science · doi:10.1007/978-981-96-0865-2_14 · Cited by 1
“Interestingly, the classic BM25 model [18] remains competitive, surpassing some semantic models trained on search behavior datasets.”
Mobile-Env: Building Qualified Evaluation Benchmarks for LLM-GUI Interaction
Zhang, D., Shen, Z., Xie, R., Zhang, S., Xie, T., Zhao, Z., Chen, S., Chen, L., Xu, H., Cao, R., et al. (2023) · arXiv · doi:10.48550/arxiv.2305.08144 · Cited by 1
“Consequently, an open-source information retrieval engine, Pyserini [9], is adopted following Yao et al. [31] to mimic the WikiHow search engine.”
MethodAccelerating Retrieval-Augmented Language Model Serving with Speculation
Zhang, Z., Zhu, A., Yang, L., Xu, Y., Li, L., Phothilimthana, P. M., & Jia, Z. (2024) · arXiv · doi:10.48550/arxiv.2401.14021 · Cited by 1
“We use the implementation from Pyserini (Lin et al., 2021) for all dense and sparse retrievers, where the dense retrievers are built on top of the standard FAISS library (Johnson et al., 2019).”
MethodSPARe: Supercharged Lexical Retrievers on GPU with Sparse Kernels
de Almeida, T. D., & Matos, S. (2024) · Lecture notes in computer science · doi:10.1007/978-3-031-56063-7_33 · Cited by 1
A Counterfactual Explanation Framework for Retrieval Models
Chandna, B., & Sen, P. (2026) · Findings of the Association for Computational Linguistics: ACL (Findings of ACL) · doi:10.18653/v1/2026.findings-acl.1917 · Cited by 1
Bridging Preferences: Multi-Stakeholder Insights on Ideal News Recommendations
Kolb, T. E., Nalis, I., & Neidhardt, J. (2025) · ACM Conference on User Modeling, Adaptation and Personalization (UMAP) · doi:10.1145/3699682.3728355 · Cited by 1
QPP-RA: Aggregating Large Language Model Rankings
Betello, F., Russo, M., Dütting, P., Leonardi, S., & Silvestri, F. (2025) · ACM SIGIR International Conference on Theory of Information Retrieval (ICTIR) · doi:10.1145/3731120.3744575 · Cited by 1
“In particular, we employed the SPLADE++ [23] retriever, as implemented in the version available through the Pyserini library [32].”
Towards Fair Rankings: Leveraging LLMs for Gender Bias Detection and Measurement
Mousavian, M., Abbasiantaeb, Z., Aliannejadi, M., & Crestani, F. (2025) · ACM SIGIR International Conference on Theory of Information Retrieval (ICTIR) · doi:10.1145/3731120.3744620 · Cited by 1
HiPC-QR: Hierarchical Prompt Chaining for Query Reformulation
Yang, H., Li, H., & Gonçalves, T. (2025) · Information · doi:10.3390/info16090790 · Cited by 1
A Large-Scale Web Search Dataset for Federated Online Learning to Rank
Gregoriadis, M., Kang, J., & Pouwelse, J. A. (2025) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3746252.3761651 · Cited by 1
Rethinking Data Use in Large Language Models
Min, S., Hajishirzi, H., & Zettlemoyer, L. (2025) · Computational Linguistics · doi:10.1162/coli.a.573 · Cited by 1
Wang, Y., Zhang, Y., Xu, K., Xu, Y., Wang, Y., & Feng, Q. (2025) · International Conference on Computer Communication and Artificial Intelligence (CCAI) · doi:10.1109/ccai65422.2025.11189813 · Cited by 1
Diversification and Fairness in Search: Two Sides of the Same Coin?
Cherumanal, S. P., Scholer, F., & Spina, D. (2025) · ACM Transactions on Information Systems · doi:10.1145/3774320 · Cited by 1
“For the initial unfair ranking, most teams used retrieval systems like BM25 [66] from the Pyserini implementation [53], Divergence from Randomness from the Terrier platform [55] and LambdaMART [11] from the LightGBM [44] implementation.”
MessIRve: A Large-Scale Spanish Information Retrieval Dataset
Valentini, F., Cotik, V., Furman, D., Bercovich, I., Altszyler, E., & Pérez, J. M. (2025) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2025.emnlp-main.1412 · Cited by 1
“We index the corpus and run retrieval using Pyserini with default parameters and Spanish analyzer (Lin et al., 2021).”
QuackIR: Retrieval in DuckDB and Other Relational Database Management Systems
Ge, Y., Chen, Z., & Lin, J. (2025) · Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2025.emnlp-industry.33 · Cited by 1
“QuackIR supports sparse, dense, and hybrid retrieval (Lin, 2021).”
InfluentialSelf-citationLearning Context-aware Term Importance for Query Performance Prediction
Saleminezhad, A., Arabzadeh, N., Beheshti, S., & Bagheri, E. (2025) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3779428 · Cited by 1
“Looking at the baselines, among the traditional ones, we observe that while score-based methods such as NQC and SMV perform reasonably well on easier datasets, their performance diminishes on DL-Hard and varies considerably on TREC DL 2021 and 2022, showcasing their limitations in handling diverse…”
RAG-LER: Ranking adapted generation with language-model enabled regulation
Zhai, F., Tang, W., & Jin, J. (2025) · Neurocomputing · doi:10.1016/j.neucom.2025.131514 · Cited by 1
Impact of LLM-Modified Queries and Documents in Training Data on Neural Retrieval Models
Nakachi, Y., & Kato, M. P. (2025) · Annual International ACM SIGIR Conference on Research and Development in Information Retrieval in the Asia Pacific Region (SIGIR-AP) · doi:10.1145/3767695.3769485 · Cited by 1
A retrieval model with contextual correlation analysis for verbose queries
Podder, D., Paik, J. H., & Mitra, P. (2026) · Journal of Intelligent Information Systems · doi:10.1007/s10844-025-01009-4 · Cited by 1
Agrawal, K., Nargund, N., & Banerjee, O. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-16632-6_24 · Cited by 1
“It becomes feasible to strategically maximize their complementary strengths for better retrieval results when the interaction between these two approaches is viewed as a competitive process [19].”
Beyond Pointwise: Exploring Comparative Approaches for Predicting Query Performance Using LLMs
Golecki, C., Racki, A., Król, J., Zawistowski, P., & Wawer, A. (2026) · ACM Transactions on Information Systems · doi:10.1145/3793556 · Cited by 1
Legal Case Entailment via Optimal Transport–Enhanced Retrieval and Large Language Model Ranking
Tran, T. N., Nguyen, P. M., & Nguyen, M. L. (2026) · ACM Transactions on Knowledge Discovery from Data · doi:10.1145/3789254 · Cited by 1
“We use the implementation of BM25 provided by the Pyserini toolkit [32].”
Practical, Efficient, In-Memory Inverted Indexes
Mackenzie, J., MacAvaney, S., Mallia, A., & Siedlaczek, M. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-21321-1_1 · Cited by 1
Wang, R. T., Ding, Y., Li, M., Lin, S., & Qi, Y. (2026) · Knowledge-Based Systems · doi:10.1016/j.knosys.2026.116004 · Cited by 1
Self-citationReaLiC: Reward-aligned latent reasoning for conversational query rewriting
Xu, L., Qin, G., Yi, Z., Cao, Y., Luo, Z., Shi, L., & Kou, F. (2026) · Information Fusion · doi:10.1016/j.inffus.2026.104409 · Cited by 1
Krishnasamy, S., & Wihardjo, I. (2026) · International Workshop on Semantic Evaluation (SemEval) · doi:10.18653/v1/2026.semeval-1.389 · Cited by 1
“We evaluated four approaches: BM25 using Pyserini (Lin et al., 2021) ( k 1 =0 .”
Improving Retrieval-Augmented Generation without Taxonomy-based Error Categorization
Zhang, G., Peng, Y., & Weng, C. (2026) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/2026.acl-short.14 · Cited by 1
“(1)) 15: end for 16: return π θ baselines supported by pyserini and BM25 utilities (Lin et al., 2021; Douze et al., 2024).”
Reproducibility in Information Retrieval
Maistro, M., Breuer, T., & Ferro, N. (2026) · Foundations and Trends® in Information Retrieval · doi:10.1108/ftinr-05-2025-0087 · Cited by 1
Abbasiantaeb, Z., Lupart, S., Gohsen, M., Mirzakhmedova, N., Kiesel, J., Dalton, J., & Aliannejadi, M. (2026) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3805712.3808591 · Cited by 1
“In addition to the raw passages, we also release pre-built Pyserini [28] (BM25) and SPLADE indices for the collection to facilitate fast and effective retrieval experiments.”
CRED: Calibrated Relational Enhanced Distillation for LLM-Based Pointwise Reranking
Yang, Q., Zhang, W., Wang, Y., Li, X., & Nie, L. (2026) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3805712.3809850 · Cited by 1
Zhang, D., & Smucker, M. D. (2026) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3805712.3809973 · Cited by 2
“For a given query, we first perform lexical retrieval using Pyserini’s implementation of BM25 with RM3 pseudo-relevance feedback [9].”
MCP Servers for Pyserini and RankLLM: Enabling Agentic Retrieval-Augmented Generation
Ge, Y., Guo, Z., Sharifymoghaddam, S., & Lin, J. (2026) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3805712.3808376 · Cited by 1
“To better address this new use case of search, agentic RAG, we introduce MCP servers to Pyserini [23] and RankLLM [36], widely-used Python toolkits for information retrieval (IR) research; we have been building toward this release since June 2025, 1 and we now share the full system with the…”
Self-citationDoes LLM Relevance Labelling Work for Arabic?
Alaofi, M., & Haouari, F. (2026) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3805712.3809855 · Cited by 1
“We indexed the HTML web pages – including their titles and main content – via Pyserini [19] with Arabic language processing enabled.”
Zerhoudi, S., Roegiest, A., Mitrović, J., & Granitzer, M. (2026) · ACM SIGIR International Conference on Theory of Information Retrieval (ICTIR) · doi:10.1145/3805713.3820402 · Cited by 1
Revisiting BM25 Feedback Models using HyDE
Jedidi, N., & Lin, J. (2026) · International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) · doi:10.1145/3805712.3809889 · Cited by 1
“Pyserini provides convenient APIs for accessing index statistics used to compute feedback document vectors, ˜ 𝑓 ( ˆ 𝑑 𝑖 ) , via its IndexReader API, as well as for constructing custom Lucene queries through its Query Builder API.”
Self-citationReproducibility in the Era of AI Agents
Lin, J. (2026) · ACM SIGIR Forum · doi:10.1145/3841298.3841310 · Cited by 1
“• Pyserini provides Python bindings to the capabilities offered in Anserini and integration with neural retrieval models built on industry-standard packages such as Hugging Face Transformers, PyTorch, and Faiss.”
InfluentialSelf-citationCHALLENGES AND BARRIERS IN ICT-BASED LEARNING AND INFORMATION RETRIEVAL
Sumbal*, ISMAIL, M., & Awan, A. K. (2025) · Journal for Current Sign · Cited by 1
Security Camera Implementation System on Smart Home Using Web-Based Frame Difference Methods
Irawanto, M. F., Imammuddien, A. M., & Wijayanti, R. A. (2023) · Jurnal Jartel Jurnal Jaringan Telekomunikasi · doi:10.33795/jartel.v13i1.537 · Cited by 1
Research on Dynamic Mathematical Resource Screening Methods Based on Machine Learning
Zhou, H. (2023) · Journal of Applied Mathematics and Physics · doi:10.4236/jamp.2023.1111227 · Cited by 1
Detecting Semantic Errors in Tables using Textual Evidence
Pham, T. M., Knoblock, C. A., & Chen, M. (2023) · IEEE International Conference on Big Data (BigData) · doi:10.1109/bigdata59044.2023.10386632 · Cited by 1
“First, we use the segmented Wikipedia provided in KILT [22] and index the segments in KILT using a traditional Lucene search engine, Pyserini [23].”
MethodShodiev, D., Kopanichuk, I., Chashchin, A., Grabovoy, A., Kildyakov, A., & Chekhovich, Y. V. (2023) · Ivannikov ISPRAS Open Conference (ISPRAS) · doi:10.1109/ispras60948.2023.10508170 · Cited by 1
“We choose Pyserini [23], a library for working with functions provided by Lucene and Anserini (search engines) in Python.”
MethodThe Information Retrieval Experiment Platform (Extended Abstract)
Fröbe, M., Reimer, J. H., MacAvaney, S., Deckers, N., Reich, S., Bevendorff, J., Stein, B., Hagen, M., & Potthast, M. (2024) · International Joint Conference on Artificial Intelligence · doi:10.24963/ijcai.2024/931 · Cited by 1
“…degree the advanced system is still better than the baseline on the different task (1 indicates a perfect reproducibility, values between 0 and 1 indicate reproducibility with diminished improvements on the different task, and 0 indicates failed reproducibility), while the delta relative…”
BackgroundMethodLesavourey, M., & Hubert, G. (2023) · HAL · Cited by 1
LEq: Large Language Models Generate Expanded Queries for Searching
Sinhababu, N., & Khatun, R. (2024) · International Conference on Computing, Communication and Networking Technologies (ICCCNT) · doi:10.1109/icccnt61001.2024.10725314 · Cited by 1
“For retrieval and evaluation, we used libraries provided by PyTerrier [21] and Pyserini [20].”
MethodEfficient Session Retrieval Using Topical Index Shards
Hendriksen, G., Hiemstra, D., & de Vries, A. P. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-88711-6_3 · Cited by 1
D3: Dynamic Docid Decoding for Multi-Intent Generative Retrieval
Kim, J., Lee, D., Hong, S., & Hwang, S.-W. (2026) · Conference of the European Chapter of the Association for Computational Linguistics · doi:10.18653/v1/2026.eacl-industry.58 · Cited by 1
THUIR at the NTCIR-18 FairWeb-2 Task
Su, H., Li, H., Tu, Y., Ai, Q., & Liu, Y. (2025) · Institutional Repositories DataBase (IRDB) · doi:10.20736/0002002035 · Cited by 1
AMS42 at the NTCIR-18 FairWeb-2 Task
Rus, C., Kareem, J., Xu, C., Liu, Y., Deng, Z., & Heuss, M. (2025) · Institutional Repositories DataBase (IRDB) · doi:10.20736/0002002031 · Cited by 1
University of Amsterdam at the CLEF 2024 SimpleText Track
Bakker, J., Yüksel, G., & Kamps, J. (2024) · UvA-DARE (University of Amsterdam) · Cited by 1
Adaptive Latent Entity Expansion for Document RetrievalPreprint
Mackie, I., Chatterjee, S., MacAvaney, S., & Dalton, J. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2306.17082 · Cited by 1
“We use Pyserini [23] version 0.16.0 for indexing the corpora and datasets for terms and entities.”
MethodInfluentialXWalk: Random Walk Based Candidate Retrieval for Product SearchPreprint
Eskreis-Winkler, J. M., Kim, Y., & Stanton, A. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2307.12019 · Cited by 1
“First is lexical retrieval using BM25 scoring (BM25 Pyserini [9] to build a Lucene index based on listing titles in our dataset and then retrieve candidates using BM25 rankings using bag-of-words representations.”
MethodBuilding Interpretable and Reliable Open Information Retriever for New Domains OvernightPreprint
Yu, X., Zhou, B., & Roth, D. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2308.04756 · Cited by 1
“We use DPR and FiD-KD implemented by (Lin et al., 2021).”
MethodCitation-augmented text reuse detection
Muther, R. (2024) · Digital Repository Service (Northeastern University) · doi:10.17760/d20649817 · Cited by 1
A Cross-Lingual Statutory Article Retrieval Dataset for Taiwan Legal StudiesPreprint
Wang, Y., Su, F.-D., Yeh, T.-Y., & Fan, Y.-C. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2410.11450 · Cited by 1
Metadata
- Authors
- Lin, J., Ma, X., Lin, S.-C., Yang, J.-H., Pradeep, R., & Nogueira, R.
- Affiliations
- University of Waterloo (CA)
- Venue
- Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval
- Publisher
- ACM
- Published
- 2021-07-11
- Pages
- 2356-2362
- OpenAlex
- W3180230246
- Cited by
- 360
- References
- 41
- Citations by year
- 2026: 67 · 2025: 72 · 2024: 61 · 2023: 97 · 2022: 44 · 2021: 18
- FWCI
- 45.53
- Type
- conference paper
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- Locations
- International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR) (publishedVersion, cc-by)
- Funders
- Natural Sciences and Engineering Research Council of Canada
- Waterloo-Huawei Joint Innovation Laboratory
- Canada First Research Excellence Fund
- Topics
- Physical Sciences › Computer Science › Information Systems › Information Retrieval and Search Behavior
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Physical Sciences › Computer Science › Information Systems › Recommender Systems and Techniques
- Keywords
- reproducibility, sparse retrieval, dense retrieval, hybrid retrieval, BM25, transformer representations, nearest neighbor search, neural ranking, relevance assessment
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.