{"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/key-value-memory-networks-for-directly","title":"Key-Value Memory Networks for Directly Reading Documents","arxiv_id":"1606.03126","date":"2016-06-09","proceeding":"EMNLP 2016 11","authors":["Alexander Miller","Adam Fisch","Jesse Dodge","Amir-Hossein Karimi","Antoine Bordes","Jason Weston"],"abstract":"Directly reading documents and being able to answer questions from them is an\nunsolved challenge. To avoid its inherent difficulty, question answering (QA)\nhas been directed towards using Knowledge Bases (KBs) instead, which has proven\neffective. Unfortunately KBs often suffer from being too restrictive, as the\nschema cannot support certain types of answers, and too sparse, e.g. Wikipedia\ncontains much more information than Freebase. In this work we introduce a new\nmethod, Key-Value Memory Networks, that makes reading documents more viable by\nutilizing different encodings in the addressing and output stages of the memory\nread operation. To compare using KBs, information extraction or Wikipedia\ndocuments directly in a single framework we construct an analysis tool,\nWikiMovies, a QA dataset that contains raw text alongside a preprocessed KB, in\nthe domain of movies. Our method reduces the gap between all three settings. It\nalso achieves state-of-the-art results on the existing WikiQA benchmark.","url_abs":"http://arxiv.org/abs/1606.03126v2","url_pdf":"http://arxiv.org/pdf/1606.03126v2.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":"key-value-memory-networks-for-directly","repo_url":"https://github.com/facebookresearch/ParlAI","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}},{"paper_slug":"key-value-memory-networks-for-directly","repo_url":"https://github.com/jojonki/key-value-memory-networks","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":{"status":"ok"}}],"tasks":[{"task_slug":"question-answering","task_name":"Question Answering"}],"methods":[],"datasets_introduced":[{"slug":"wikimovies","name":"WikiMovies","full_name":""}],"methods_introduced":[],"results":[{"leaderboard":"/sota/question-answering-on-wikiqa","task":"Question Answering","dataset":"WikiQA","model":"Key-Value Memory Network","rank_in_archive_order":12,"of":25,"metrics":{"MAP":"0.7069","MRR":"0.7265"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1606.03126","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}