{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/read-verify-machine-reading-comprehension","title":"Read + Verify: Machine Reading Comprehension with Unanswerable Questions","arxiv_id":"1808.05759","date":"2018-08-17","proceeding":null,"authors":["Minghao Hu","Furu Wei","Yuxing Peng","Zhen Huang","Nan Yang","Dongsheng Li"],"abstract":"Machine reading comprehension with unanswerable questions aims to abstain\nfrom answering when no answer can be inferred. In addition to extract answers,\nprevious works usually predict an additional \"no-answer\" probability to detect\nunanswerable cases. However, they fail to validate the answerability of the\nquestion by verifying the legitimacy of the predicted answer. To address this\nproblem, we propose a novel read-then-verify system, which not only utilizes a\nneural reader to extract candidate answers and produce no-answer probabilities,\nbut also leverages an answer verifier to decide whether the predicted answer is\nentailed by the input snippets. Moreover, we introduce two auxiliary losses to\nhelp the reader better handle answer extraction as well as no-answer detection,\nand investigate three different architectures for the answer verifier. Our\nexperiments on the SQuAD 2.0 dataset show that our system achieves a score of\n74.2 F1 on the test set, achieving state-of-the-art results at the time of\nsubmission (Aug. 28th, 2018).","url_abs":"http://arxiv.org/abs/1808.05759v5","url_pdf":"http://arxiv.org/pdf/1808.05759v5.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[],"tasks":[{"task_slug":"machine-reading-comprehension","task_name":"Machine Reading Comprehension"},{"task_slug":"question-answering","task_name":"Question Answering"},{"task_slug":"reading-comprehension","task_name":"Reading Comprehension"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/question-answering-on-squad20","task":"Question Answering","dataset":"SQuAD2.0","model":"Reinforced Mnemonic Reader + Answer Verifier (single model)","rank_in_archive_order":244,"of":286,"metrics":{"EM":"71.767","F1":"74.295"},"uses_additional_data":false},{"leaderboard":"/sota/question-answering-on-squad20-dev","task":"Question Answering","dataset":"SQuAD2.0 dev","model":"RMR + ELMo (Model-III)","rank_in_archive_order":11,"of":13,"metrics":{"EM":"72.3","F1":"74.8"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1808.05759","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}