{"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/a-thorough-examination-of-the-cnndaily-mail","title":"A Thorough Examination of the CNN/Daily Mail Reading Comprehension Task","arxiv_id":"1606.02858","date":"2016-06-09","proceeding":"ACL 2016 8","authors":["Danqi Chen","Jason Bolton","Christopher D. Manning"],"abstract":"Enabling a computer to understand a document so that it can answer\ncomprehension questions is a central, yet unsolved goal of NLP. A key factor\nimpeding its solution by machine learned systems is the limited availability of\nhuman-annotated data. Hermann et al. (2015) seek to solve this problem by\ncreating over a million training examples by pairing CNN and Daily Mail news\narticles with their summarized bullet points, and show that a neural network\ncan then be trained to give good performance on this task. In this paper, we\nconduct a thorough examination of this new reading comprehension task. Our\nprimary aim is to understand what depth of language understanding is required\nto do well on this task. We approach this from one side by doing a careful\nhand-analysis of a small subset of the problems and from the other by showing\nthat simple, carefully designed systems can obtain accuracies of 73.6% and\n76.6% on these two datasets, exceeding current state-of-the-art results by\n7-10% and approaching what we believe is the ceiling for performance on this\ntask.","url_abs":"http://arxiv.org/abs/1606.02858v2","url_pdf":"http://arxiv.org/pdf/1606.02858v2.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":[{"paper_slug":"a-thorough-examination-of-the-cnndaily-mail","repo_url":"https://github.com/danqi/rc-cnn-dailymail","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"a-thorough-examination-of-the-cnndaily-mail","repo_url":"https://github.com/clarenceguan/rc-cnn-daily-pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"a-thorough-examination-of-the-cnndaily-mail","repo_url":"https://github.com/clarenceguan/rc-cnn-dailymail-pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[{"task_slug":"articles","task_name":"Articles"},{"task_slug":"reading-comprehension","task_name":"Reading Comprehension"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/question-answering-on-cnn-daily-mail","task":"Question Answering","dataset":"CNN / Daily Mail","model":"Attentive + relabling + ensemble","rank_in_archive_order":3,"of":16,"metrics":{"CNN":"77.6","Daily Mail":"79.2"},"uses_additional_data":false},{"leaderboard":"/sota/question-answering-on-cnn-daily-mail","task":"Question Answering","dataset":"CNN / Daily Mail","model":"AttentiveReader + bilinear attention","rank_in_archive_order":11,"of":16,"metrics":{"CNN":"72.4","Daily Mail":"75.8"},"uses_additional_data":false},{"leaderboard":"/sota/question-answering-on-cnn-daily-mail","task":"Question Answering","dataset":"CNN / Daily Mail","model":"Classifier","rank_in_archive_order":14,"of":16,"metrics":{"CNN":"67.9","Daily Mail":"68.3"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1606.02858","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}