[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"detail-sidebar-cat-0-en-105":3,"doc-seo-513534-105":59,"doc-detail-513534-en":130},{"code":4,"msg":5,"data":6},0,"success",[7,13,18,23,28,33,38,43,48,51,55],{"id":8,"doc_module":4,"doc_module_name":9,"category_name":10,"show_sort_weight":11,"slug":12},1,"Document","Story & Novel",90,"story-novel",{"id":14,"doc_module":4,"doc_module_name":9,"category_name":15,"show_sort_weight":16,"slug":17},2,"Literature",80,"literature",{"id":19,"doc_module":4,"doc_module_name":9,"category_name":20,"show_sort_weight":21,"slug":22},4,"Exam",70,"exam",{"id":24,"doc_module":4,"doc_module_name":9,"category_name":25,"show_sort_weight":26,"slug":27},5,"Comic",60,"comic",{"id":29,"doc_module":4,"doc_module_name":9,"category_name":30,"show_sort_weight":31,"slug":32},6,"Technology",50,"technology",{"id":34,"doc_module":4,"doc_module_name":9,"category_name":35,"show_sort_weight":36,"slug":37},7,"Healthcare",40,"healthcare",{"id":39,"doc_module":4,"doc_module_name":9,"category_name":40,"show_sort_weight":41,"slug":42},8,"Research & Report",30,"research-report",{"id":44,"doc_module":4,"doc_module_name":9,"category_name":45,"show_sort_weight":46,"slug":47},9,"Religion & Spirituality",20,"religion-spirituality",{"id":46,"doc_module":4,"doc_module_name":9,"category_name":49,"show_sort_weight":46,"slug":50},"World Cup","world-cup",{"id":52,"doc_module":4,"doc_module_name":9,"category_name":53,"show_sort_weight":52,"slug":54},10,"Lifestyle","lifestyle",{"id":56,"doc_module":4,"doc_module_name":9,"category_name":57,"show_sort_weight":24,"slug":58},19,"General","general",{"code":4,"msg":60,"data":61},"ok",{"site_id":62,"language":63,"slug":64,"title":65,"keywords":66,"description":67,"schema_data":68,"social_meta":123,"head_meta":125,"extra_data":127,"updated_unix":129},105,"en","training-end-to-end-dialogue-systems-with-the-ubuntu-dialogue-corpus-abstract","Training End-to-End Dialogue Systems with the Ubuntu Dialogue Corpus - Abstract","","Neural network-based end-to-end dialogue systems are analyzed using an updated version of the Ubuntu Dialogue Corpus, containing nearly 1 million multi-turn dialogues, 7 million utterances, and 100 million words. The study tests two training environments: retrieval over candidate next responses and generation by maximizing log-likelihood of utterances conditioned on dialogue context. A next utterance classification (NUC) recall task and vector-based topicality metrics evaluate performance. Qualitative error analysis identifies primary error causes and suggests future research directions.",{"@graph":69,"@context":122},[70,84,105],{"@type":71,"itemListElement":72},"BreadcrumbList",[73,77,79,82],{"item":74,"name":75,"@type":76,"position":8},"https://docshare.wps.com","Home","ListItem",{"item":78,"name":9,"@type":76,"position":14},"https://docshare.wps.com/document/",{"item":80,"name":40,"@type":76,"position":81},"https://docshare.wps.com/document/research-report/",3,{"item":83,"name":65,"@type":76,"position":19},"https://docshare.wps.com/document/training-end-to-end-dialogue-systems-with-the-ubuntu-dialogue-corpus-abstract/513534/",{"url":83,"name":65,"@type":85,"image":86,"author":91,"headline":65,"publisher":94,"fileFormat":97,"inLanguage":63,"description":67,"dateModified":98,"datePublished":99,"encodingFormat":97,"isAccessibleForFree":100,"interactionStatistic":101},"DigitalDocument",{"url":87,"@type":88,"width":89,"height":90},"https://docshare.wps.com/thumbnails/training-end-to-end-dialogue-systems-with-the-ubuntu-dialogue-corpus-abstract/513534.png","ImageObject",300,407,{"name":92,"@type":93},"aboduw","Person",{"url":74,"name":95,"@type":96},"DocShare","Organization","application/pdf","2026-11-10","2026-10-06",true,{"@type":102,"interactionType":103,"userInteractionCount":52},"InteractionCounter",{"@type":104},"ViewAction",{"@type":106,"mainEntity":107},"FAQPage",[108,114,118],{"name":109,"@type":110,"acceptedAnswer":111},"What dataset is used to train and evaluate the dialogue systems?","Question",{"text":112,"@type":113},"The paper uses an updated version of the Ubuntu Dialogue Corpus, which contains nearly one million multi-turn dyadic conversations and millions of utterances.","Answer",{"name":115,"@type":110,"acceptedAnswer":116},"How are the dialogue models trained in the paper?",{"text":117,"@type":113},"Two setups are evaluated: retrieval models select the correct next response from candidates, and generative models maximize the log-likelihood of generated utterances given the conversation context.",{"name":119,"@type":110,"acceptedAnswer":120},"How is model performance measured?",{"text":121,"@type":113},"Models are evaluated on a recall task called next utterance classification (NUC) and with vector-based metrics that capture the topicality of responses.","https://schema.org",{"og:url":83,"og:type":124,"og:title":65,"og:site_name":95,"og:description":67},"article",{"robots":126,"canonical":83},"index,follow",{"doc_id":128,"site_id":62},513534,1791337578,{"code":4,"msg":5,"data":131},{"doc_id":128,"user_id":132,"nickname":92,"user_avatar":133,"doc_module":4,"category_id":39,"category_name":40,"doc_title":65,"doc_description":67,"doc_content":134,"file_id":135,"file_url":136,"file_type":137,"file_size":138,"view_count":52,"is_deleted":4,"is_public":8,"is_downloadable":8,"audit_status":8,"page_count":139,"language":140,"language_code":63,"site_id":62,"html_lang":63,"table_of_contents":141,"faqs":142,"seo_title":143,"seo_description":67,"update_tm":144,"read_time":145,"preview_page_count":19,"preview_urls":146},3985747862266,"https://ap-avatar.wpscdn.com/davatar_085a072bc5b1113ac321206ff7593b45","Training End-to-End Dialogue Systems with the Ubuntu Dialogue Corpus  \nRyan Lowe  \nSchool of Computer Science, McGill University  \nNissan Pow  \nSchool of Computer Science, McGill University  \nIulian Vlad Serban  \nDIRO, Universite´ de Montre´al  \nLaurent Charlin  \nSchool of Computer Science, McGill University  \nChia-Wei Liu  \nSchool of Computer Science, McGill University  \nJoelle Pineau  \nSchool of Computer Science, McGill University  \nRYAN . LOWE @CS . MCGILL . CA  \nNISSAN . POW @MAIL . MCGILL . COM  \nJULIANSERBAN@GMAIL . COM  \nLCHARLIN @ GMAIL . COM  \nCHIA-WEI. LIU @MAIL . MCGILL . CA  \nJPINEAU @CS . MCGILL . CA  \nEditor: Amanda Stent  \nSubmitted 04/2016; Accepted 12/2016; Published online 01/2017  \nAbstract  \nIn this paper, we analyze neural network-based dialogue systems trained in an end-to-end manner using an updated version of the recent Ubuntu Dialogue Corpus, a dataset containing almost  \n1 million multi-turn dialogues, with a total of over 7 million utterances and 100 million words 1. This dataset is interesting because of its size, long context lengths, and technical nature; thus, it can be used to train large models directly from data with minimal feature engineering. We provide baselines in two different environments: one where models are trained to select the correct next response from a list of candidate responses, and one where models are trained to maximize the loglikelihood of a generated utterance conditioned on the context of the conversation. These are both evaluated on a recall task that we call next utterance classiﬁcation (NUC), and using vector-based metrics that capture the topicality of the responses. We observe that current end-to-end models are  \n1. This work is an extension of a paper appearing in SIGDIAL (Lowe et al., 2015) . This paper further includes results on generative dialogue models, more extensive evaluation of the retrieval models using vector-based generative metrics, and a qualitative examination of responses from the generative models and classiﬁcation errors made by the Dual Encoder model. Experiments are performed on a new version of the corpus, the Ubuntu Dialogue Corpus v2, which is publicly available: [https://github.com/rkadlec/ubuntu-ranking-dataset-creator](https://github.com/rkadlec/ubuntu-ranking-dataset-creator. The)[. The](https://github.com/rkadlec/ubuntu-ranking-dataset-creator. The) early dataset has been updated to add features and ﬁx bugs, which are detailed in Section 3 .  \n􀀍c2017 Ryan Lowe, Nissan Pow, Iulian Vlad Serban, Laurent Charlinn, Chia-Wei Liu and Joelle Pineau This is an open-access article distributed under the terms of a Creative Commons Attribution License ([http](http: ==creativecommons:org=licenses=by=3:0=)[:](http: ==creativecommons:org=licenses=by=3:0=)[ ==](http: ==creativecommons:org=licenses=by=3:0=)[creativecommons](http: ==creativecommons:org=licenses=by=3:0=)[:](http: ==creativecommons:org=licenses=by=3:0=)[org](http: ==creativecommons:org=licenses=by=3:0=)[=](http: ==creativecommons:org=licenses=by=3:0=)[licenses](http: ==creativecommons:org=licenses=by=3:0=)[=](http: ==creativecommons:org=licenses=by=3:0=)[by](http: ==creativecommons:org=licenses=by=3:0=)[=](http: ==creativecommons:org=licenses=by=3:0=)[3](http: ==creativecommons:org=licenses=by=3:0=)[:](http: ==creativecommons:org=licenses=by=3:0=)[0](http: ==creativecommons:org=licenses=by=3:0=)[=](http: ==creativecommons:org=licenses=by=3:0=)).  \nLOWE, POW, SERBAN, CHARLINN, LIU AND PINEAU  \nunable to completely solve these tasks; thus, we provide a qualitative error analysis to determine the primary causes of error for end-to-end models evaluated on NUC, and examine sample utterances from the generative models. As a result of this analysis, we suggest some promising directions for future research on the Ubuntu Dialogue Corpus, which can also be applied to end-to-end dialogue systems in general.  \n1. Introduction  \nDeriving statistical models that can naturally and coherently converse wit","cbCaid9JP5vd8Lxf","https://ap.wps.com/l/cbCaid9JP5vd8Lxf","pdf",781616,35,"English","# Abstract\n## Introduction","[{\"question\":\"What dataset is used to train and evaluate the dialogue systems?\",\"answer\":\"The paper uses an updated version of the Ubuntu Dialogue Corpus, which contains nearly one million multi-turn dyadic conversations and millions of utterances.\"},{\"question\":\"How are the dialogue models trained in the paper?\",\"answer\":\"Two setups are evaluated: retrieval models select the correct next response from candidates, and generative models maximize the log-likelihood of generated utterances given the conversation context.\"},{\"question\":\"How is model performance measured?\",\"answer\":\"Models are evaluated on a recall task called next utterance classification (NUC) and with vector-based metrics that capture the topicality of responses.\"}]","Training End-to-End Dialogue Systems with the Ubuntu Dialogue Corpus - Abstract | PDF",1791286753,88,[147,151,155,159],{"page_no":8,"url":148,"width":149,"height":150},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-17/95cabd9ba5d3ac788a1a7257229bab28.webp",983,1272,{"page_no":14,"url":152,"width":153,"height":154},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-17/dd55ba9d02a537451f9bf3c6fe5fa401.webp",982,1270,{"page_no":81,"url":156,"width":157,"height":158},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-17/8596a647ef1be6f1c43a2c58a08e0c75.webp",981,1269,{"page_no":19,"url":160,"width":161,"height":162},"https://res-academy.cache.wpscdn.com/upload/kdocshare/2026-11-17/7146c72ef4ca19995953e2c1eab32694.webp",980,1268]