{"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/neural-belief-tracker-data-driven-dialogue","title":"Neural Belief Tracker: Data-Driven Dialogue State Tracking","arxiv_id":"1606.03777","date":"2016-06-12","proceeding":"ACL 2017 7","authors":["Nikola Mrkšić","Diarmuid Ó Séaghdha","Tsung-Hsien Wen","Blaise Thomson","Steve Young"],"abstract":"One of the core components of modern spoken dialogue systems is the belief\ntracker, which estimates the user's goal at every step of the dialogue.\nHowever, most current approaches have difficulty scaling to larger, more\ncomplex dialogue domains. This is due to their dependency on either: a) Spoken\nLanguage Understanding models that require large amounts of annotated training\ndata; or b) hand-crafted lexicons for capturing some of the linguistic\nvariation in users' language. We propose a novel Neural Belief Tracking (NBT)\nframework which overcomes these problems by building on recent advances in\nrepresentation learning. NBT models reason over pre-trained word vectors,\nlearning to compose them into distributed representations of user utterances\nand dialogue context. Our evaluation on two datasets shows that this approach\nsurpasses past limitations, matching the performance of state-of-the-art models\nwhich rely on hand-crafted semantic lexicons and outperforming them when such\nlexicons are not provided.","url_abs":"http://arxiv.org/abs/1606.03777v2","url_pdf":"http://arxiv.org/pdf/1606.03777v2.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":"dialogue-state-tracking","task_name":"Dialogue State Tracking"},{"task_slug":"representation-learning","task_name":"Representation Learning"},{"task_slug":"spoken-dialogue-systems","task_name":"Spoken Dialogue Systems"},{"task_slug":"spoken-language-understanding","task_name":"Spoken Language Understanding"}],"methods":[],"datasets_introduced":[{"slug":"wizard-of-oz","name":"Wizard-of-Oz","full_name":""}],"methods_introduced":[],"results":[{"leaderboard":"/sota/dialogue-state-tracking-on-second-dialogue","task":"Dialogue State Tracking","dataset":"Second dialogue state tracking challenge","model":"Neural belief tracker","rank_in_archive_order":5,"of":7,"metrics":{"Area":"90","Food":"84","Joint":"73.4","Price":"94","Request":"96.5"},"uses_additional_data":false},{"leaderboard":"/sota/dialogue-state-tracking-on-wizard-of-oz","task":"Dialogue State Tracking","dataset":"Wizard-of-Oz","model":"Neural belief tracker","rank_in_archive_order":9,"of":10,"metrics":{"Joint":"84.4","Request":"96.5"},"uses_additional_data":false}],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1606.03777","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}