{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,8,23]],"date-time":"2025-08-23T05:22:54Z","timestamp":1755926574844},"reference-count":49,"publisher":"MIT Press","license":[{"start":{"date-parts":[[2021,5,28]],"date-time":"2021-05-28T00:00:00Z","timestamp":1622160000000},"content-version":"vor","delay-in-days":147,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["direct.mit.edu"],"crossmark-restriction":true},"short-container-title":[],"published-print":{"date-parts":[[2021,5,26]]},"abstract":"<jats:title>Abstract<\/jats:title><jats:p>Tracking dialogue states to better interpret user goals and feed downstream policy learning is a bottleneck in dialogue management. Common practice has been to treat it as a problem of classifying dialogue content into a set of pre-defined slot-value pairs, or generating values for different slots given the dialogue history. Both have limitations on considering dependencies that occur on dialogues, and are lacking of reasoning capabilities. This paper proposes to track dialogue states gradually with reasoning over dialogue turns with the help of the back-end data. Empirical results demonstrate that our method outperforms the state-of-the-art methods in terms of joint belief accuracy for MultiWOZ 2.1, a large-scale human--human dialogue dataset across multiple domains.<\/jats:p>","DOI":"10.1162\/tacl_a_00384","type":"journal-article","created":{"date-parts":[[2021,5,28]],"date-time":"2021-05-28T16:23:34Z","timestamp":1622219014000},"page":"557-569","update-policy":"http:\/\/dx.doi.org\/10.1162\/mitpressjournals.corrections.policy","source":"Crossref","is-referenced-by-count":13,"title":["Dialogue State Tracking with Incremental Reasoning"],"prefix":"10.1162","volume":"9","author":[{"given":"Lizi","family":"Liao","sequence":"first","affiliation":[{"name":"School of Computing, National University of Singapore, Singapore. liaolizi.llz@gmail.com"}]},{"given":"Le Hong","family":"Long","sequence":"additional","affiliation":[{"name":"School of Computing, National University of Singapore, Singapore. lehonglong@u.nus.edu"}]},{"given":"Yunshan","family":"Ma","sequence":"additional","affiliation":[{"name":"School of Computing, National University of Singapore, Singapore. yunshan.ma@gmail.com"}]},{"given":"Wenqiang","family":"Lei","sequence":"additional","affiliation":[{"name":"School of Computing, National University of Singapore, Singapore. wenqianglei@gmail.com"}]},{"given":"Tat-Seng","family":"Chua","sequence":"additional","affiliation":[{"name":"School of Computing, National University of Singapore, Singapore. chuats@comp.nus.edu.sg"}]}],"member":"281","published-online":{"date-parts":[[2021,5,27]]},"reference":[{"key":"2021060823392815600_bib1","unstructured":"Radford Alec , WuJeffrey, ChildRewon, LuanDavid, AmodeiDario, and SutskeverIlya. 2019. Language models are unsupervised multitask learners. Technical report, OpenAI."},{"key":"2021060823392815600_bib2","first-page":"1468","article-title":"BERT- DST: Scalable end-to-end dialogue state tracking with bidirectional encoder representations from transformer","author":"Chao","year":"2019"},{"key":"2021060823392815600_bib3","first-page":"7521","article-title":"Schema-guided multi-domain dialogue state tracking with graph attention neural networks","volume-title":"AAAI","author":"Chen","year":"2020"},{"key":"2021060823392815600_bib4","first-page":"1724","article-title":"Learning phrase representations using rnn encoder\u07dddecoder for statistical machine translation","volume-title":"EMNLP","author":"Cho","year":"2014"},{"key":"2021060823392815600_bib5","doi-asserted-by":"crossref","first-page":"160","DOI":"10.1145\/1390156.1390177","article-title":"A unified architecture for natural language processing: Deep neural networks with multitask learning","volume-title":"ICML","author":"Collobert","year":"2008"},{"key":"2021060823392815600_bib6","article-title":"Go for a walk and arrive at the answer: Reasoning over paths in knowledge bases using reinforcement learning","author":"Das","year":"2017","journal-title":"arXiv preprint arXiv:1711.05851"},{"key":"2021060823392815600_bib7","first-page":"4171","article-title":"BERT: Pre-training of deep bidirectional transformers for language understanding","volume-title":"NAACL","author":"Devlin","year":"2019"},{"key":"2021060823392815600_bib8","article-title":"MultiWOZ 2.1: Multi-domain dialogue state corrections and state tracking baselines","author":"Eric","year":"2019","journal-title":"CoRR"},{"key":"2021060823392815600_bib9","first-page":"264","article-title":"Dialog state tracking: A neural reading comprehension approach","volume-title":"SIGDIAL","author":"Gao","year":"2019"},{"key":"2021060823392815600_bib10","article-title":"HyST: A hybrid approach for flexible and accurate dialogue state tracking","author":"Goel","year":"2019","journal-title":"arXiv preprint arXiv:1907.00883"},{"key":"2021060823392815600_bib11","first-page":"173","article-title":"Neural collaborative filtering","volume-title":"WWW","author":"He","year":"2017"},{"key":"2021060823392815600_bib12","first-page":"35","article-title":"Trippy: A triple copy strategy for value independent neural dialog state tracking","volume-title":"SIGDIAL","author":"Heck","year":"2020"},{"key":"2021060823392815600_bib13","first-page":"263","article-title":"The second dialog state tracking challenge","volume-title":"SIGDIAL","author":"Henderson","year":"2014"},{"key":"2021060823392815600_bib14","first-page":"292","article-title":"Word-based dialog state tracking with recurrent neural networks","volume-title":"SIGDIAL","author":"Henderson","year":"2014"},{"key":"2021060823392815600_bib15","article-title":"A simple language model for task-oriented dialogue","author":"Hosseini-Asl","year":"2020","journal-title":"arXiv preprint arXiv:2005.00796"},{"key":"2021060823392815600_bib16","first-page":"567","article-title":"Efficient dialogue state tracking by selectively overwriting memory","volume-title":"ACL","author":"Kim","year":"2020"},{"key":"2021060823392815600_bib17","first-page":"414","article-title":"Recipe for building robust spoken dialog state trackers: Dialog state tracking challenge system description","volume-title":"SIGDIAL","author":"Lee","year":"2013"},{"key":"2021060823392815600_bib18","doi-asserted-by":"crossref","first-page":"801","DOI":"10.1145\/3240508.3240605","article-title":"Knowledge-aware multimodal dialogue systems","volume-title":"Proceedings of the 26th ACM international conference on Multimedia","author":"Liao","year":"2018"},{"key":"2021060823392815600_bib19","doi-asserted-by":"crossref","DOI":"10.1145\/3442381.3450134","article-title":"Multi-domain dialogue state tracking with recursive inference","volume-title":"The Web Conference","author":"Liao","year":"2021"},{"key":"2021060823392815600_bib20","first-page":"6294","article-title":"Learned in translation: Contextualized word vectors","volume-title":"NIPS","author":"McCann","year":"2017"},{"key":"2021060823392815600_bib21","first-page":"1777","article-title":"Neural belief tracker: Data- driven dialogue state tracking","volume-title":"ACL","author":"Mrk\u0161i\u0107","year":"2017"},{"key":"2021060823392815600_bib22","first-page":"467","article-title":"Loopy belief propagation for approximate inference: An empirical study","volume-title":"UAI","author":"Murphy","year":"1999"},{"key":"2021060823392815600_bib23","article-title":"Toward scalable neural dialogue state tracking model","author":"Nouri","year":"2018","journal-title":"arXiv preprint arXiv:1812.00899"},{"key":"2021060823392815600_bib24","first-page":"34","article-title":"Dialogue state tracking with explicit slot connection modeling","volume-title":"ACL","author":"Ouyang","year":"2020"},{"key":"2021060823392815600_bib25","unstructured":"Lawrence Page , SergeyBrin, RajeevMotwani, and TerryWinograd. 1999, The pagerank citation ranking: Bringing order to the Web. Stanford InfoLab."},{"key":"2021060823392815600_bib26","first-page":"305","article-title":"Dialog state tracking, a machine reading approach using memory network","volume-title":"EACL","author":"Perez","year":"2017"},{"key":"2021060823392815600_bib27","first-page":"432","article-title":"Large-scale multi-domain belief tracking with knowledge sharing","volume-title":"ACL","author":"Ramadan","year":"2018"},{"key":"2021060823392815600_bib28","first-page":"561","article-title":"Scalable multi-domain dialogue state tracking","volume-title":"ASRU Workshop","author":"Rastogi","year":"2017"},{"key":"2021060823392815600_bib29","first-page":"1876","article-title":"Scalable and accurate dialogue state tracking via hierarchical sequence generation","volume-title":"EMNLP","author":"Ren","year":"2019"},{"key":"2021060823392815600_bib30","first-page":"2780","article-title":"Towards universal dialogue state tracking","volume-title":"EMNLP","author":"Ren","year":"2018"},{"key":"2021060823392815600_bib31","first-page":"6322","article-title":"A contextual hierarchical attention network with adaptive objective for dialogue state tracking","volume-title":"ACL","author":"Shan","year":"2020"},{"key":"2021060823392815600_bib32","first-page":"330","article-title":"A generalized rule based tracker for dialogue state tracking","volume-title":"SLT Workshop","author":"Sun","year":"2014"},{"key":"2021060823392815600_bib33","first-page":"318","article-title":"The sjtu system for dialog state tracking challenge 2","volume-title":"SIGDIAL","author":"Sun","year":"2014"},{"issue":"4","key":"2021060823392815600_bib34","doi-asserted-by":"crossref","first-page":"562","DOI":"10.1016\/j.csl.2009.07.003","article-title":"Bayesian update of dialogue state: A POMDP framework for spoken dialogue systems","volume":"24","author":"Thomson","year":"2010","journal-title":"Computer Speech & Language"},{"key":"2021060823392815600_bib35","first-page":"5998","article-title":"Attention is all you need","volume-title":"NIPS","author":"Vaswani","year":"2017"},{"key":"2021060823392815600_bib36","first-page":"423","article-title":"A simple and generic belief tracking mechanism for the dialog state tracking challenge: On the believability of observed information","volume-title":"SIGDIAL","author":"Wang","year":"2013"},{"key":"2021060823392815600_bib37","first-page":"438","article-title":"A network-based end- to-end trainable task-oriented dialogue system","volume-title":"EACL","author":"Wen","year":"2017"},{"key":"2021060823392815600_bib38","first-page":"282","article-title":"Web-style ranking and slu combination for dialog state tracking","volume-title":"SIGDIAL","author":"Williams","year":"2014"},{"issue":"2","key":"2021060823392815600_bib39","doi-asserted-by":"crossref","first-page":"393","DOI":"10.1016\/j.csl.2006.06.008","article-title":"Partially observable markov decision processes for spoken dialog systems","volume":"21","author":"Williams","year":"2007","journal-title":"Computer Speech & Language"},{"key":"2021060823392815600_bib40","first-page":"808","article-title":"Transferable multi-domain state generator for task-oriented dialogue systems","volume-title":"ACL","author":"Chien-Sheng","year":"2019"},{"key":"2021060823392815600_bib41","article-title":"Google's neural machine translation system: Bridging the gap between human and machine translation","author":"Yonghui","year":"2016","journal-title":"arXiv preprint arXiv:1609.08144"},{"key":"2021060823392815600_bib42","first-page":"209","article-title":"Cost-sensitive active learning for dialogue state tracking","volume-title":"SIGDIAL","author":"Xie","year":"2018"},{"key":"2021060823392815600_bib43","first-page":"564","article-title":"Deeppath: A reinforcement learning method for knowledge graph reasoning","volume-title":"EMNLP","author":"Xiong","year":"2017"},{"key":"2021060823392815600_bib44","first-page":"1448","article-title":"An end-to-end approach for handling unknown slot values in dialogue state tracking","volume-title":"ACL","author":"Puyang","year":"2018"},{"key":"2021060823392815600_bib45","article-title":"Find or classify? Dual strategy for slot-value predictions on multi-domain dialog state tracking","author":"Zhang","year":"2019","journal-title":"arXiv preprint arXiv:1910.03544"},{"key":"2021060823392815600_bib46","doi-asserted-by":"crossref","first-page":"2401","DOI":"10.1145\/3308558.3313598","article-title":"Neural multimodal belief tracker with adaptive attention for dialogue systems","volume-title":"The World Wide Web Conference","author":"Zhang","year":"2019"},{"key":"2021060823392815600_bib47","first-page":"1458","article-title":"Global-locally self-attentive encoder for dialogue state tracking","volume-title":"ACL","author":"Zhong","year":"2018"},{"key":"2021060823392815600_bib48","article-title":"Multi-domain dialogue state tracking as dynamic knowledge graph enhanced question answering","author":"Li","year":"2019","journal-title":"arXiv preprint arXiv:1911.06192"},{"key":"2021060823392815600_bib49","first-page":"757","article-title":"Incremental lstm-based dialog state tracker","volume-title":"ASRU Workshop","author":"Zilka","year":"2015"}],"container-title":["Transactions of the Association for Computational Linguistics"],"original-title":[],"language":"en","link":[{"URL":"http:\/\/direct.mit.edu\/tacl\/article-pdf\/doi\/10.1162\/tacl_a_00384\/1923739\/tacl_a_00384.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"http:\/\/direct.mit.edu\/tacl\/article-pdf\/doi\/10.1162\/tacl_a_00384\/1923739\/tacl_a_00384.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,12,29]],"date-time":"2022-12-29T00:27:57Z","timestamp":1672273677000},"score":1,"resource":{"primary":{"URL":"https:\/\/direct.mit.edu\/tacl\/article\/doi\/10.1162\/tacl_a_00384\/101875\/Dialogue-State-Tracking-with-Incremental-Reasoning"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021]]},"references-count":49,"URL":"https:\/\/doi.org\/10.1162\/tacl_a_00384","relation":{},"ISSN":["2307-387X"],"issn-type":[{"value":"2307-387X","type":"electronic"}],"subject":[],"published-other":{"date-parts":[[2021]]},"published":{"date-parts":[[2021]]}}}