{"url":"/task/machine-reading-comprehension","name":"Machine Reading Comprehension","slug":"machine-reading-comprehension","description_markdown":"**Machine Reading Comprehension** is one of the key problems in Natural Language Understanding, where the task is to read and comprehend a given text passage, and then answer questions based on it.\r\n\r\n<span class=\"description-source\">Source: [Making Neural Machine Reading Comprehension Faster ](https://arxiv.org/abs/1904.00796)</span>","categories":[{"name":"Computer Vision","url":"/area/computer-vision"},{"name":"Natural Language Processing","url":"/area/natural-language-processing"}],"source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","slug_source":"archive_url"},"counts":{"papers_tagged":555,"papers_with_code":206,"benchmarks":4,"benchmark_tables_in_archive":4,"benchmark_tables_shown":4,"benchmark_tables_withheld_as_spam":0,"benchmark_definition":"a leaderboard table with at least one row; benchmark_tables_shown also counts the zero-row tables; benchmark_tables_in_archive adds the tables withheld as spam","datasets":43,"subtasks":0,"parent_tasks":2},"benchmarks":[{"leaderboard":"/sota/machine-reading-comprehension-on-dream","slug":"machine-reading-comprehension-on-dream","dataset":"DREAM","dataset_url":"/dataset/dream","rows_in_archive":3,"metrics":["Accuracy"],"first_row_in_archive_order":{"model":"ASA + RoBERTa","paper_title":"Adversarial Self-Attention for Language Understanding","paper_url":"/paper/adversarial-self-attention-for-language","paper_date":"2022-06-25","arxiv_id":"2206.12608","code_links":[{"title":"gingasan/adversarialsa","url":"https://github.com/gingasan/adversarialsa"}],"syntology":null}},{"leaderboard":"/sota/machine-reading-comprehension-on-reclor","slug":"machine-reading-comprehension-on-reclor","dataset":"ReClor","dataset_url":"/dataset/reclor","rows_in_archive":3,"metrics":["Accuracy","Accuracy (easy)","Accuracy (hard)"],"first_row_in_archive_order":{"model":"XLNet-large","paper_title":"ReClor: A Reading Comprehension Dataset Requiring Logical Reasoning","paper_url":"/paper/reclor-a-reading-comprehension-dataset-1","paper_date":"2020-02-11","arxiv_id":"2002.04326","code_links":[{"title":"yuweihao/reclor","url":"https://github.com/yuweihao/reclor"}],"syntology":{"n":2,"n_ran":0,"n_unverified":2,"n_pointer_only":2}}},{"leaderboard":"/sota/machine-reading-comprehension-on-uquad","slug":"machine-reading-comprehension-on-uquad","dataset":"UQuAD","dataset_url":"/dataset/uquad","rows_in_archive":2,"metrics":["Exact Match","F1"],"first_row_in_archive_order":{"model":"BERT","paper_title":"UQuAD1.0: Development of an Urdu Question Answering Training Data for Machine Reading Comprehension","paper_url":"/paper/uquad1-0-development-of-an-urdu-question","paper_date":"2021-11-02","arxiv_id":"2111.01543","code_links":[],"syntology":null}},{"leaderboard":"/sota/machine-reading-comprehension-on-biomrc","slug":"machine-reading-comprehension-on-biomrc","dataset":"BIOMRC","dataset_url":"/dataset/biomrc","rows_in_archive":1,"metrics":["Acc"],"first_row_in_archive_order":{"model":"MLP-based-weighting (on BIOMRC Lite)","paper_title":"Contextual embedding and model weighting by fusing domain knowledge on Biomedical Question Answering","paper_url":"/paper/contextual-embedding-and-model-weighting-by","paper_date":"2022-06-26","arxiv_id":"2206.12866","code_links":[{"title":"leoleoasd/mlp-based-weighting","url":"https://github.com/leoleoasd/mlp-based-weighting"}],"syntology":null}}],"datasets":[{"url":"/dataset/logiqa","name":"LogiQA","full_name":"","num_papers_in_archive":127},{"url":"/dataset/mrqa-2019","name":"MRQA","full_name":"","num_papers_in_archive":116},{"url":"/dataset/mctest","name":"MCTest","full_name":"","num_papers_in_archive":114},{"url":"/dataset/cbt","name":"CBT","full_name":"Children’s Book Test","num_papers_in_archive":92},{"url":"/dataset/reclor","name":"ReClor","full_name":"","num_papers_in_archive":88},{"url":"/dataset/cmrc","name":"CMRC","full_name":"Chinese Machine Reading Comprehension","num_papers_in_archive":69},{"url":"/dataset/belebele","name":"Belebele","full_name":"","num_papers_in_archive":68},{"url":"/dataset/dream","name":"DREAM","full_name":"","num_papers_in_archive":68},{"url":"/dataset/mutual","name":"MuTual","full_name":"MuTual","num_papers_in_archive":55},{"url":"/dataset/c3","name":"C3","full_name":"","num_papers_in_archive":54},{"url":"/dataset/drcd","name":"DRCD","full_name":"Delta Reading Comprehension Dataset","num_papers_in_archive":53},{"url":"/dataset/quoref","name":"Quoref","full_name":"Quoref","num_papers_in_archive":50},{"url":"/dataset/cmrc-2018","name":"CMRC 2018","full_name":"Chinese Machine Reading Comprehension 2018","num_papers_in_archive":49},{"url":"/dataset/duorc","name":"DuoRC","full_name":"DuoRC","num_papers_in_archive":42},{"url":"/dataset/chid","name":"ChID","full_name":"Chinese IDiom dataset","num_papers_in_archive":38},{"url":"/dataset/visualmrc","name":"VisualMRC","full_name":"VisualMRC: Machine Reading Comprehension on Document Images","num_papers_in_archive":36},{"url":"/dataset/holl-e","name":"Holl-E","full_name":"","num_papers_in_archive":30},{"url":"/dataset/molweni","name":"Molweni","full_name":"","num_papers_in_archive":29},{"url":"/dataset/adversarialqa","name":"AdversarialQA","full_name":"","num_papers_in_archive":26},{"url":"/dataset/klue","name":"KLUE","full_name":"Korean Language Understanding Evaluation","num_papers_in_archive":21},{"url":"/dataset/tweetqa","name":"TweetQA","full_name":"","num_papers_in_archive":19},{"url":"/dataset/uit-viquad","name":"UIT-ViQuAD","full_name":"","num_papers_in_archive":15},{"url":"/dataset/biomrc","name":"BIOMRC","full_name":"","num_papers_in_archive":13},{"url":"/dataset/who-did-what","name":"Who-did-What","full_name":"Who did What","num_papers_in_archive":13},{"url":"/dataset/cmrc-2017","name":"CMRC 2017","full_name":"Chinese Machine Reading Comprehension 2017","num_papers_in_archive":11},{"url":"/dataset/reco","name":"ReCO","full_name":"","num_papers_in_archive":9},{"url":"/dataset/cmrc-2019","name":"CMRC 2019","full_name":"Chinese Machine Reading Comprehension 2019","num_papers_in_archive":8},{"url":"/dataset/bipar","name":"BiPaR","full_name":"BiPaR","num_papers_in_archive":6},{"url":"/dataset/muserc","name":"MuSeRC","full_name":"Russian Multi-Sentence Reading Comprehension","num_papers_in_archive":6},{"url":"/dataset/onestopqa","name":"OneStopQA","full_name":"","num_papers_in_archive":5},{"url":"/dataset/rucos","name":"RuCoS","full_name":"Russian Reading Comprehension with Commonsense Reasoning","num_papers_in_archive":5},{"url":"/dataset/uit-vinewsqa","name":"UIT-ViNewsQA","full_name":"","num_papers_in_archive":5},{"url":"/dataset/vimmrc","name":"ViMMRC","full_name":"Vietnamese Multiple-choice Machine Reading Comprehension Corpus","num_papers_in_archive":4},{"url":"/dataset/uit-viwikiqa","name":"UIT-ViWikiQA","full_name":"","num_papers_in_archive":3},{"url":"/dataset/expmrc","name":"ExpMRC","full_name":"","num_papers_in_archive":2},{"url":"/dataset/mc-afp","name":"MC-AFP","full_name":"","num_papers_in_archive":2},{"url":"/dataset/merger-agreement-understanding-dataset-maud","name":"Merger Agreement Understanding Dataset (MAUD)","full_name":"","num_papers_in_archive":2},{"url":"/dataset/uquad","name":"UQuAD","full_name":"Urdu Question Answering Dataset","num_papers_in_archive":2},{"url":"/dataset/wikireading-recycled","name":"WikiReading Recycled","full_name":"","num_papers_in_archive":2},{"url":"/dataset/idk-mrc","name":"IDK-MRC","full_name":"","num_papers_in_archive":1},{"url":"/dataset/nerel-bio","name":"NEREL-BIO","full_name":"","num_papers_in_archive":1},{"url":"/dataset/persianqa","name":"PersianQA","full_name":"Persian Question Answering Dataset","num_papers_in_archive":1},{"url":"/dataset/qbsum","name":"QBSUM","full_name":"","num_papers_in_archive":1}],"subtasks":[],"parent_tasks":[{"url":"/task/reading-comprehension","name":"Reading Comprehension"},{"url":"/task/visual-question-answering","name":"Visual Question Answering (VQA)"}],"papers":{"order":"repositories listed in the archive (desc), then date (desc); the archive holds no stars","population":"papers tagged with this task that list at least one repository in the archive","shown":30,"of":206,"tagged_in_all":555,"items":[{"url":"/paper/ms-marco-a-human-generated-machine-reading","title":"MS MARCO: A Human Generated MAchine Reading COmprehension Dataset","date":"2016-11-28","arxiv_id":"1611.09268","repositories_listed":14,"syntology":{"n":33,"n_ran":3,"n_unverified":30,"n_pointer_only":0}},{"url":"/paper/a-unified-mrc-framework-for-named-entity","title":"A Unified MRC Framework for Named Entity Recognition","date":"2019-10-25","arxiv_id":"1910.11476","repositories_listed":8,"syntology":{"n":5,"n_ran":2,"n_unverified":3,"n_pointer_only":2}},{"url":"/paper/sdnet-contextualized-attention-based-deep","title":"SDNet: Contextualized Attention-based Deep Network for Conversational Question Answering","date":"2018-12-10","arxiv_id":"1812.03593","repositories_listed":6,"syntology":null},{"url":"/paper/stochastic-answer-networks-for-machine","title":"Stochastic Answer Networks for Machine Reading Comprehension","date":"2017-12-10","arxiv_id":"1712.03556","repositories_listed":6,"syntology":null},{"url":"/paper/stochastic-answer-networks-for-squad-20","title":"Stochastic Answer Networks for SQuAD 2.0","date":"2018-09-24","arxiv_id":"1809.09194","repositories_listed":5,"syntology":null},{"url":"/paper/multi-task-learning-for-machine-reading","title":"Multi-task Learning with Sample Re-weighting for Machine Reading Comprehension","date":"2018-09-18","arxiv_id":"1809.06963","repositories_listed":5,"syntology":null},{"url":"/paper/klue-korean-language-understanding-evaluation","title":"KLUE: Korean Language Understanding Evaluation","date":"2021-05-20","arxiv_id":"2105.09680","repositories_listed":4,"syntology":null},{"url":"/paper/dice-loss-for-data-imbalanced-nlp-tasks","title":"Dice Loss for Data-imbalanced NLP Tasks","date":"2019-11-07","arxiv_id":"1911.02855","repositories_listed":4,"syntology":{"n":3,"n_ran":0,"n_unverified":3,"n_pointer_only":0}},{"url":"/paper/chinesebert-chinese-pretraining-enhanced-by","title":"ChineseBERT: Chinese Pretraining Enhanced by Glyph and Pinyin Information","date":"2021-06-30","arxiv_id":"2106.16038","repositories_listed":3,"syntology":null},{"url":"/paper/dureaderrobust-a-chinese-dataset-towards","title":"DuReader_robust: A Chinese Dataset Towards Evaluating Robustness and Generalization of Machine Reading Comprehension in Real-World Applications","date":"2020-04-23","arxiv_id":"2004.11142","repositories_listed":3,"syntology":null},{"url":"/paper/clue-a-chinese-language-understanding","title":"CLUE: A Chinese Language Understanding Evaluation Benchmark","date":"2020-04-13","arxiv_id":"2004.05986","repositories_listed":3,"syntology":null},{"url":"/paper/dual-multi-head-co-attention-for-multi-choice","title":"DUMA: Reading Comprehension with Transposition Thinking","date":"2020-01-26","arxiv_id":"2001.09415","repositories_listed":3,"syntology":null},{"url":"/paper/dureader-a-chinese-machine-reading","title":"DuReader: a Chinese Machine Reading Comprehension Dataset from Real-world Applications","date":"2017-11-14","arxiv_id":"1711.05073","repositories_listed":3,"syntology":null},{"url":"/paper/reinforced-mnemonic-reader-for-machine","title":"Reinforced Mnemonic Reader for Machine Reading Comprehension","date":"2017-05-08","arxiv_id":"1705.02798","repositories_listed":3,"syntology":null},{"url":"/paper/the-belebele-benchmark-a-parallel-reading","title":"The Belebele Benchmark: a Parallel Reading Comprehension Dataset in 122 Language Variants","date":"2023-08-31","arxiv_id":"2308.16884","repositories_listed":2,"syntology":{"n":6,"n_ran":0,"n_unverified":6,"n_pointer_only":6}},{"url":"/paper/rethinking-label-smoothing-on-multi-hop","title":"Rethinking Label Smoothing on Multi-hop Question Answering","date":"2022-12-19","arxiv_id":"2212.09512","repositories_listed":2,"syntology":null},{"url":"/paper/genius-sketch-based-language-model-pre","title":"GENIUS: Sketch-based Language Model Pre-training via Extreme and Selective Masking for Text Generation and Augmentation","date":"2022-11-18","arxiv_id":"2211.10330","repositories_listed":2,"syntology":null},{"url":"/paper/fact-driven-logical-reasoning","title":"Fact-driven Logical Reasoning for Machine Reading Comprehension","date":"2021-05-21","arxiv_id":"2105.10334","repositories_listed":2,"syntology":null},{"url":"/paper/dependency-parsing-as-mrc-based-span-span","title":"Dependency Parsing as MRC-based Span-Span Prediction","date":"2021-05-17","arxiv_id":"2105.07654","repositories_listed":2,"syntology":null},{"url":"/paper/bidirectional-machine-reading-comprehension","title":"Bidirectional Machine Reading Comprehension for Aspect Sentiment Triplet Extraction","date":"2021-03-13","arxiv_id":"2103.07665","repositories_listed":2,"syntology":null},{"url":"/paper/deer-a-data-efficient-language-model-for","title":"ECONET: Effective Continual Pretraining of Language Models for Event Temporal Reasoning","date":"2020-12-30","arxiv_id":"2012.15283","repositories_listed":2,"syntology":null},{"url":"/paper/logiqa-a-challenge-dataset-for-machine","title":"LogiQA: A Challenge Dataset for Machine Reading Comprehension with Logical Reasoning","date":"2020-07-16","arxiv_id":"2007.08124","repositories_listed":2,"syntology":null},{"url":"/paper/teaching-machine-comprehension-with","title":"Teaching Machine Comprehension with Compositional Explanations","date":"2020-05-02","arxiv_id":"2005.00806","repositories_listed":2,"syntology":{"n":10,"n_ran":10,"n_unverified":0,"n_pointer_only":10}},{"url":"/paper/retrospective-reader-for-machine-reading","title":"Retrospective Reader for Machine Reading Comprehension","date":"2020-01-27","arxiv_id":"2001.09694","repositories_listed":2,"syntology":null},{"url":"/paper/asking-questions-the-human-way-scalable","title":"Asking Questions the Human Way: Scalable Question-Answer Generation from Text Corpus","date":"2020-01-27","arxiv_id":"2002.00748","repositories_listed":2,"syntology":{"n":8,"n_ran":1,"n_unverified":7,"n_pointer_only":2}},{"url":"/paper/numnet-machine-reading-comprehension-with","title":"NumNet: Machine Reading Comprehension with Numerical Reasoning","date":"2019-10-15","arxiv_id":"1910.06701","repositories_listed":2,"syntology":null},{"url":"/paper/mmm-multi-stage-multi-task-learning-for-multi","title":"MMM: Multi-stage Multi-task Learning for Multi-choice Reading Comprehension","date":"2019-10-01","arxiv_id":"1910.00458","repositories_listed":2,"syntology":null},{"url":"/paper/pre-training-with-whole-word-masking-for","title":"Pre-Training with Whole Word Masking for Chinese BERT","date":"2019-06-19","arxiv_id":"1906.08101","repositories_listed":2,"syntology":null},{"url":"/paper/text-understanding-with-the-attention-sum","title":"Text Understanding with the Attention Sum Reader Network","date":"2016-03-04","arxiv_id":"1603.01547","repositories_listed":2,"syntology":{"n":10,"n_ran":0,"n_unverified":10,"n_pointer_only":0}},{"url":"/paper/mrceval-a-comprehensive-challenging-and","title":"MRCEval: A Comprehensive, Challenging and Accessible Machine Reading Comprehension Benchmark","date":"2025-03-10","arxiv_id":"2503.07144","repositories_listed":1,"syntology":null}],"syntology_records":7,"syntology_note":"a paper without a record is not a recorded non-run: it may lack an arXiv id or simply be absent from the graph layer"},"description_links":{"kept":0,"unwrapped_to_text":0,"bare_urls_linked":0,"relative_images_dropped":0,"rule":"internal links are kept only when the target slug exists in the catalog"},"syntology":{"read_at":"2026-09-24T18:15:14+00:00","claim":"Per-sample execution status on synthesized fixtures ('ran N of M samples'); not a correctness claim and not a ranking signal.","status_vocabulary":{"ran_honours":"ran, honoured the contract we drafted","ran_violates":"ran, violated the contract we drafted","ran_draft_wrong":"ran; our contract draft was wrong, not the code","ran_fixture":"ran; our fixture could not drive it","ran":"ran on a synthesized input","unverified":"unverified (harvested, no recorded run)"}},"not_shown":{"libraries":"the archive has no per-task library table","trend_sparklines":"the Trend column of the benchmarks table was a rendered image; it is not in the archive","social_and_latest_sorts":"stars and social signals are not in the archive"}}