{"url":"/sota/ad-hoc-information-retrieval-on-trec-robust04","task":{"name":"Ad-Hoc Information Retrieval","url":"/task/ad-hoc-information-retrieval","note":null},"dataset":{"name":"TREC Robust04","url":"/dataset/robust04"},"category":"Natural Language Processing","categories":["Natural Language Processing"],"category_note":null,"description":"Ad-hoc information retrieval refers to the task of returning information resources related to a user query formulated in natural language.","description_from":"task","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","rank":"the archive's row order at snapshot; not re-ranked","rows_end_at":"2025-07-28","rows_withheld_as_spam":0,"metric_values":"the archive's strings, untouched"},"metrics":["nDCG@20","MAP","P@20"],"metric_direction":{"note":"inferred from the metric name only (the archive records no direction); null = not inferred, chart draws points only","by_metric":{"nDCG@20":"higher","MAP":"higher","P@20":null}},"counts":{"rows":21,"rows_with_code":21,"rows_with_paper_page":21,"rows_dated":21,"rows_using_additional_data":0},"rows":[{"rank_in_archive_order":1,"model":"monoT5-3B (zero-shot)","metrics":{"MAP":"0.3876","P@20":"0.5165","nDCG@20":"0.6091"},"uses_additional_data":false,"paper_date":"2020-03-14","paper":"/paper/document-ranking-with-a-pretrained-sequence","paper_url":"https://arxiv.org/abs/2003.06713v1","paper_title":"Document Ranking with a Pretrained Sequence-to-Sequence Model","code":"https://github.com/castorini/pygaggle","n_code_links":2,"syntology":null},{"rank_in_archive_order":2,"model":"PARADE(ELECTRA)","metrics":{"P@20":"0.4604","nDCG@20":"0.5399"},"uses_additional_data":false,"paper_date":"2020-08-20","paper":"/paper/parade-passage-representation-aggregation-for","paper_url":"https://arxiv.org/abs/2008.09093v2","paper_title":"PARADE: Passage Representation Aggregation for Document Reranking","code":"https://github.com/canjiali/PARADE","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":3,"n_samples":3,"n_pointer_only_licence":0}},{"rank_in_archive_order":3,"model":"CEDR-KNRM","metrics":{"P@20":"0.4667","nDCG@20":"0.5381"},"uses_additional_data":false,"paper_date":"2019-04-15","paper":"/paper/190407094","paper_url":"https://arxiv.org/abs/1904.07094v3","paper_title":"CEDR: Contextualized Embeddings for Document Ranking","code":"https://github.com/Georgetown-IR-Lab/contextualized-reps-for-ranking","n_code_links":7,"syntology":{"n_ran":2,"n_unverified":5,"n_samples":7,"n_pointer_only_licence":0}},{"rank_in_archive_order":4,"model":"PARADE(BERT)","metrics":{"P@20":"0.4486","nDCG@20":"0.5252"},"uses_additional_data":false,"paper_date":"2020-08-20","paper":"/paper/parade-passage-representation-aggregation-for","paper_url":"https://arxiv.org/abs/2008.09093v2","paper_title":"PARADE: Passage Representation Aggregation for Document Reranking","code":"https://github.com/canjiali/PARADE","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":3,"n_samples":3,"n_pointer_only_licence":0}},{"rank_in_archive_order":5,"model":"BERT-MaxP","metrics":{"nDCG@20":"0.469"},"uses_additional_data":false,"paper_date":"2019-05-22","paper":"/paper/deeper-text-understanding-for-ir-with","paper_url":"https://arxiv.org/abs/1905.09217v1","paper_title":"Deeper Text Understanding for IR with Contextual Neural Language Modeling","code":"https://github.com/AdeDZY/SIGIR19-BERT-IR","n_code_links":1,"syntology":{"n_ran":2,"n_unverified":5,"n_samples":7,"n_pointer_only_licence":0}},{"rank_in_archive_order":6,"model":"BERT-SumP","metrics":{"nDCG@20":"0.467"},"uses_additional_data":false,"paper_date":"2019-05-22","paper":"/paper/deeper-text-understanding-for-ir-with","paper_url":"https://arxiv.org/abs/1905.09217v1","paper_title":"Deeper Text Understanding for IR with Contextual Neural Language Modeling","code":"https://github.com/AdeDZY/SIGIR19-BERT-IR","n_code_links":1,"syntology":{"n_ran":2,"n_unverified":5,"n_samples":7,"n_pointer_only_licence":0}},{"rank_in_archive_order":7,"model":"POSIT-DRMM-MV","metrics":{"MAP":"0.271","P@20":"0.389","nDCG@20":"0.464"},"uses_additional_data":false,"paper_date":"2018-09-05","paper":"/paper/deep-relevance-ranking-using-enhanced","paper_url":"http://arxiv.org/abs/1809.01682v2","paper_title":"Deep Relevance Ranking Using Enhanced Document-Query Interactions","code":"https://github.com/nlpaueb/deep-relevance-ranking","n_code_links":1,"syntology":null},{"rank_in_archive_order":8,"model":"Vanilla BERT","metrics":{"P@20":"0.4042","nDCG@20":"0.4541"},"uses_additional_data":false,"paper_date":"2019-04-15","paper":"/paper/190407094","paper_url":"https://arxiv.org/abs/1904.07094v3","paper_title":"CEDR: Contextualized Embeddings for Document Ranking","code":"https://github.com/Georgetown-IR-Lab/contextualized-reps-for-ranking","n_code_links":7,"syntology":{"n_ran":2,"n_unverified":5,"n_samples":7,"n_pointer_only_licence":0}},{"rank_in_archive_order":9,"model":"NPRF-DRMM","metrics":{"MAP":"0.2904","P@20":"0.4064","nDCG@20":"0.4502"},"uses_additional_data":false,"paper_date":"2018-10-30","paper":"/paper/nprf-a-neural-pseudo-relevance-feedback","paper_url":"http://arxiv.org/abs/1810.12936v1","paper_title":"NPRF: A Neural Pseudo Relevance Feedback Framework for Ad-hoc Information Retrieval","code":"https://github.com/ucasir/NPRF","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":6,"n_samples":6,"n_pointer_only_licence":0}},{"rank_in_archive_order":10,"model":"PACRR","metrics":{"MAP":"0.258","P@20":"0.374","nDCG@20":"0.445"},"uses_additional_data":false,"paper_date":"2018-09-05","paper":"/paper/deep-relevance-ranking-using-enhanced","paper_url":"http://arxiv.org/abs/1809.01682v2","paper_title":"Deep Relevance Ranking Using Enhanced Document-Query Interactions","code":"https://github.com/nlpaueb/deep-relevance-ranking","n_code_links":1,"syntology":null},{"rank_in_archive_order":11,"model":"BERT-FirstP","metrics":{"nDCG@20":"0.444"},"uses_additional_data":false,"paper_date":"2019-05-22","paper":"/paper/deeper-text-understanding-for-ir-with","paper_url":"https://arxiv.org/abs/1905.09217v1","paper_title":"Deeper Text Understanding for IR with Contextual Neural Language Modeling","code":"https://github.com/AdeDZY/SIGIR19-BERT-IR","n_code_links":1,"syntology":{"n_ran":2,"n_unverified":5,"n_samples":7,"n_pointer_only_licence":0}},{"rank_in_archive_order":12,"model":"SNRM-PRF","metrics":{"MAP":"0.2971","P@20":"0.3948","nDCG@20":"0.4391"},"uses_additional_data":false,"paper_date":"2018-10-22","paper":"/paper/from-neural-re-ranking-to-neural-ranking","paper_url":"https://dl.acm.org/citation.cfm?id=3271800","paper_title":"From Neural Re-Ranking to Neural Ranking: Learning a Sparse Representation for Inverted Indexing","code":"https://github.com/hamed-zamani/snrm","n_code_links":1,"syntology":null},{"rank_in_archive_order":13,"model":"NPRF-KNRM","metrics":{"MAP":"0.2846","P@20":"0.3926","nDCG@20":"0.4327"},"uses_additional_data":false,"paper_date":"2018-10-30","paper":"/paper/nprf-a-neural-pseudo-relevance-feedback","paper_url":"http://arxiv.org/abs/1810.12936v1","paper_title":"NPRF: A Neural Pseudo Relevance Feedback Framework for Ad-hoc Information Retrieval","code":"https://github.com/ucasir/NPRF","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":6,"n_samples":6,"n_pointer_only_licence":0}},{"rank_in_archive_order":14,"model":"SNRM","metrics":{"MAP":"0.2856","P@20":"0.3766","nDCG@20":"0.4310"},"uses_additional_data":false,"paper_date":"2018-10-22","paper":"/paper/from-neural-re-ranking-to-neural-ranking","paper_url":"https://dl.acm.org/citation.cfm?id=3271800","paper_title":"From Neural Re-Ranking to Neural Ranking: Learning a Sparse Representation for Inverted Indexing","code":"https://github.com/hamed-zamani/snrm","n_code_links":1,"syntology":null},{"rank_in_archive_order":15,"model":"DRMM","metrics":{"MAP":"0.279","P@20":"0.382","nDCG@20":"0.431"},"uses_additional_data":false,"paper_date":"2017-11-23","paper":"/paper/a-deep-relevance-matching-model-for-ad-hoc","paper_url":"http://arxiv.org/abs/1711.08611v1","paper_title":"A Deep Relevance Matching Model for Ad-hoc Retrieval","code":"https://github.com/sebastian-hofstaetter/neural-ranking-drmm","n_code_links":3,"syntology":null},{"rank_in_archive_order":16,"model":"KNRM","metrics":{"MAP":"0.2464","P@20":"0.3510","nDCG@20":"0.3989"},"uses_additional_data":false,"paper_date":"2018-10-30","paper":"/paper/nprf-a-neural-pseudo-relevance-feedback","paper_url":"http://arxiv.org/abs/1810.12936v1","paper_title":"NPRF: A Neural Pseudo Relevance Feedback Framework for Ad-hoc Information Retrieval","code":"https://github.com/ucasir/NPRF","n_code_links":1,"syntology":{"n_ran":0,"n_unverified":6,"n_samples":6,"n_pointer_only_licence":0}},{"rank_in_archive_order":17,"model":"BERT FT(Microblog)","metrics":{"MAP":"0.3278","P@20":"0.4287"},"uses_additional_data":false,"paper_date":"2019-03-26","paper":"/paper/simple-applications-of-bert-for-ad-hoc","paper_url":"http://arxiv.org/abs/1903.10972v1","paper_title":"Simple Applications of BERT for Ad Hoc Document Retrieval","code":"https://github.com/castorini/birch","n_code_links":2,"syntology":null},{"rank_in_archive_order":18,"model":"Anserini BM25+RM3","metrics":{"MAP":"0.302","P@20":"0.4012"},"uses_additional_data":false,"paper_date":"2018-12-01","paper":"/paper/the-neural-hype-and-comparisons-against-weak","paper_url":"https://dl.acm.org/citation.cfm?id=3308781","paper_title":"The Neural Hype and Comparisons Against Weak Baselines","code":"https://github.com/castorini/Anserini","n_code_links":1,"syntology":null},{"rank_in_archive_order":19,"model":"FNRM-RankProb_Embed","metrics":{"MAP":"0.2837"},"uses_additional_data":false,"paper_date":"2017-04-28","paper":"/paper/neural-ranking-models-with-weak-supervision","paper_url":"http://arxiv.org/abs/1704.08803v2","paper_title":"Neural Ranking Models with Weak Supervision","code":"https://github.com/mikvrax/TrecingLab","n_code_links":1,"syntology":null},{"rank_in_archive_order":20,"model":"FNRM-Rank_Embed","metrics":{"MAP":"0.2811"},"uses_additional_data":false,"paper_date":"2017-04-28","paper":"/paper/neural-ranking-models-with-weak-supervision","paper_url":"http://arxiv.org/abs/1704.08803v2","paper_title":"Neural Ranking Models with Weak Supervision","code":"https://github.com/mikvrax/TrecingLab","n_code_links":1,"syntology":null},{"rank_in_archive_order":21,"model":"QL","metrics":{"MAP":"0.2499"},"uses_additional_data":false,"paper_date":"2018-10-22","paper":"/paper/from-neural-re-ranking-to-neural-ranking","paper_url":"https://dl.acm.org/citation.cfm?id=3271800","paper_title":"From Neural Re-Ranking to Neural Ranking: Learning a Sparse Representation for Inverted Indexing","code":"https://github.com/hamed-zamani/snrm","n_code_links":1,"syntology":null}],"since_archive":{"claim":"Results that newer papers report for their own method, placed here by Syntology. A model pointed at the cell in the paper's own table; the number was read from that cell and checked against this leaderboard's metric, dataset, split and scale; an independent check that saw this leaderboard's other rows and every other leaderboard on the same dataset accepted it. Not reviewed by the paper's authors or by the archive's editors, and not ranked against the archive rows.","extraction_file_present":true,"measurement":{"test_papers":883,"papers_with_output":881,"judged_true":108,"judged":110,"wilson95_lower":0.9361,"measured_on":"2026-09-24","frozen_commit":"0e3de0df94"},"measurement_note":"blind adjudication of accepted entries on a held-out split of archive papers, rules frozen before the test","coverage":{"sentence":"Syntology has checked 6,885 of the 9,623 papers on this site that are newer than the archive; results from the others appear after they are checked.","complete":false,"papers_newer_than_archive":9623,"papers_checked":6885,"papers_extracted_not_yet_verified":0,"boards_without_verdict":2,"papers_not_yet_extracted":2737},"order":"newest first by month (arXiv date, else the arXiv-id month), then arXiv id descending","columns":[],"entries":[]},"syntology":{"read_at":"2026-09-25T09:33:49+00:00","claim":"Per row: N of M harvested code samples from that row's paper executed on a synthesized fixture; the other M-N are unverified. Not a reproduction of the row's number; not a correctness claim. n_pointer_only_licence counts samples the site points at rather than redistributes (a licence axis, independent of ran/unverified).","rows_with_graph_line":10,"rows_with_any_sample_ran":5,"distinct_papers_with_graph_line":4,"distinct_papers_with_any_sample_ran":2,"samples_over_distinct_papers":{"n_ran":4,"n_unverified":19,"n_samples":23,"n_pointer_only_licence":0,"note":"each paper (arXiv id) counted once, however many rows it is behind; this is the page-level figure"},"samples_row_weighted":{"n_ran":10,"n_unverified":49,"n_samples":59,"n_pointer_only_licence":0,"note":"row-weighted: a paper behind several rows is counted once per row; inflated relative to samples_over_distinct_papers by design, kept for readers summing the per-row syntology blocks"}}}