{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,6,21]],"date-time":"2025-06-21T04:29:58Z","timestamp":1750480198167},"reference-count":59,"publisher":"MIT Press - Journals","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Transactions of the Association for Computational Linguistics"],"published-print":{"date-parts":[[2020,12]]},"abstract":"<jats:p> We show that Bayes\u2019 rule provides an effective mechanism for creating document translation models that can be learned from only parallel sentences and monolingual documents a compelling benefit because parallel documents are not always available. In our formulation, the posterior probability of a candidate translation is the product of the unconditional (prior) probability of the candidate output document and the \u201creverse translation probability\u201d of translating the candidate output back into the source language. Our proposed model uses a powerful autoregressive language model as the prior on target language documents, but it assumes that each sentence is translated independently from the target to the source language. Crucially, at test time, when a source document is observed, the document language model prior induces dependencies between the translations of the source sentences in the posterior. The model\u2019s independence assumption not only enables efficient use of available data, but it additionally admits a practical left-to-right beam-search algorithm for carrying out inference. Experiments show that our model benefits from using cross-sentence context in the language model, and it outperforms existing document translation approaches. <\/jats:p>","DOI":"10.1162\/tacl_a_00319","type":"journal-article","created":{"date-parts":[[2020,7,8]],"date-time":"2020-07-08T14:53:11Z","timestamp":1594219991000},"page":"346-360","source":"Crossref","is-referenced-by-count":12,"title":["Better Document-Level Machine Translation with Bayes\u2019 Rule"],"prefix":"10.1162","volume":"8","author":[{"given":"Lei","family":"Yu","sequence":"first","affiliation":[{"name":"DeepMind."}]},{"given":"Laurent","family":"Sartran","sequence":"additional","affiliation":[{"name":"DeepMind."}]},{"given":"Wojciech","family":"Stokowiec","sequence":"additional","affiliation":[{"name":"DeepMind."}]},{"given":"Wang","family":"Ling","sequence":"additional","affiliation":[{"name":"DeepMind."}]},{"given":"Lingpeng","family":"Kong","sequence":"additional","affiliation":[{"name":"DeepMind."}]},{"given":"Phil","family":"Blunsom","sequence":"additional","affiliation":[{"name":"DeepMind"},{"name":"University of Oxford."}]},{"given":"Chris","family":"Dyer","sequence":"additional","affiliation":[{"name":"DeepMind."}]}],"member":"281","reference":[{"key":"bib1","volume-title":"Proceedings of ACL","author":"Artetxe Mikel","year":"2019"},{"key":"bib2","volume-title":"Proceedings of ICLR","author":"Artetxe Mikel","year":"2018"},{"key":"bib3","volume-title":"Proceedings of NAACL-HLT","author":"Bawden Rachel","year":"2018"},{"issue":"2","key":"bib4","first-page":"263","volume":"19","author":"Brown Peter F.","year":"1993","journal-title":"Computational Linguistics"},{"key":"bib5","volume-title":"Proceedings of ACL","author":"Cheng Yong","year":"2016"},{"key":"bib6","volume-title":"Proceedings of NAACL-HLT","author":"Chronopoulou Alexandra","year":"2019"},{"key":"bib7","volume-title":"Proceedings of ACL","author":"Dai Zihang","year":"2019"},{"key":"bib8","volume-title":"Proceedings of NAACL-HLT","author":"Devlin Jacob","year":"2019"},{"key":"bib9","volume":"1905","author":"Li Dong","year":"2019","journal-title":"CoRR"},{"key":"bib10","volume-title":"Proceedings of NAACL-HLT","author":"Edunov Sergey","year":"2019"},{"key":"bib11","volume-title":"Proceedings of EMNLP","author":"Edunov Sergey","year":"2018"},{"issue":"1","key":"bib12","first-page":"75","volume":"19","author":"Gale William A.","year":"1993","journal-title":"Computational Linguistics"},{"key":"bib13","volume":"1503","author":"G\u00fcl\u00e7ehre \u00c7aglar","year":"2015","journal-title":"CoRR"},{"key":"bib14","volume-title":"Proceedings of ACL","author":"Haffari Gholamreza","year":"2018"},{"key":"bib15","volume":"1704","author":"Jean S\u00e9bastien","year":"2017","journal-title":"CoRR"},{"key":"bib16","volume-title":"Proceedings of WMT","author":"Junczys-Dowmunt Marcin","year":"2019"},{"key":"bib17","volume-title":"Proceedings of EMNLP","author":"Kim Yoon","year":"2016"},{"key":"bib18","volume-title":"Proceedings of ICLR","author":"Kingma Diederik P.","year":"2015"},{"key":"bib19","volume-title":"Proceedings of ACL","author":"Koehn Philipp","year":"2007"},{"key":"bib20","volume":"1711","author":"Kuang Shaohui","year":"2017","journal-title":"CoRR"},{"key":"bib21","volume":"1901","author":"Lample Guillaume","year":"2019","journal-title":"CoRR"},{"key":"bib22","volume-title":"Proceedings of ICLR","author":"Lample Guillaume","year":"2018"},{"key":"bib23","volume-title":"Proceedings of EMNLP","author":"Lample Guillaume","year":"2018"},{"issue":"4","key":"bib24","volume":"14","author":"Mangu Lidia","year":"2000","journal-title":"Computer Speech & Language"},{"key":"bib25","volume-title":"Proceedings of NAACL-HLT","author":"Maruf Sameen","year":"2019"},{"key":"bib26","volume-title":"Proceedings of NeurIPS","author":"McCann Bryan","year":"2017"},{"key":"bib27","volume":"1803","author":"Merity Stephen","year":"2018","journal-title":"CoRR"},{"key":"bib28","volume-title":"Proceedings of ICLR","author":"Merity Stephen","year":"2018"},{"key":"bib29","volume-title":"Proceedings of WMT","author":"Ng Nathan","year":"2019"},{"key":"bib30","volume":"1909","author":"de Oliveira Luke","year":"2019","journal-title":"ArXiv"},{"key":"bib31","volume-title":"Proceedings of NAACL","author":"Peters Matthew E.","year":"2018"},{"key":"bib32","volume-title":"Proceedings of WMT","author":"Post Matt","year":"2018"},{"key":"bib33","unstructured":"Alec Radford, Karthik Narasimhan, Tim Salimans, and Ilya Sutskever. 2018. Improving language understanding by generative pre-training."},{"issue":"8","key":"bib34","volume":"1","author":"Radford Alec","year":"2019","journal-title":"OpenAI Blog"},{"key":"bib35","volume-title":"Proceedings of ACL","author":"Sennrich Rico","year":"2016"},{"key":"bib36","volume-title":"Proceedings of ACL","author":"Sennrich Rico","year":"2016"},{"key":"bib37","volume-title":"Proceedings of UAI","author":"Shachter Ross D.","year":"1998"},{"key":"bib38","volume-title":"Proceedings of ICML","author":"Shen Tianxiao","year":"2019"},{"key":"bib39","volume-title":"Proceedings of AMTA","author":"Snover Matthew","year":"2006"},{"key":"bib40","volume-title":"Proceedings of ICML","author":"Song Kaitao","year":"2019"},{"key":"bib41","volume-title":"Proceedings of WMT","author":"Sun Meng","year":"2019"},{"key":"bib42","volume-title":"Proceedings of DiscoMT@EMNLP","author":"Tiedemann J\u00f6rg","year":"2017"},{"key":"bib43","doi-asserted-by":"crossref","first-page":"407","DOI":"10.1162\/tacl_a_00029","volume":"6","author":"Zhaopeng Tu","year":"2018","journal-title":"Transactions of the Association for Computational Linguistics"},{"key":"bib44","volume-title":"Proceedings of NeurIPS","author":"Vaswani Ashish","year":"2017"},{"key":"bib45","volume-title":"Proceedings of EMNLP-IJCNLP","author":"Voita Elena","year":"2019"},{"key":"bib46","volume-title":"Proceedings of ACL","author":"Voita Elena","year":"2019"},{"key":"bib47","volume-title":"Proceedings of ACL","author":"Voita Elena","year":"2018"},{"key":"bib48","volume-title":"Proceedings of EMNLP","author":"Wang Longyue","year":"2017"},{"key":"bib49","volume-title":"Proceedings of EMNLP","author":"Werlen Lesly Miculicich","year":"2018"},{"key":"bib50","volume-title":"Proceedings of WMT","author":"Xia Yingce","year":"2019"},{"key":"bib51","volume-title":"Proceedings of AAAI","author":"Xiong Hao","year":"2019"},{"key":"bib52","volume":"1906","author":"Yang Zhilin","year":"2019","journal-title":"CoRR"},{"key":"bib53","volume-title":"Proceedings of EMNLP","author":"Yee Kyra","year":"2019"},{"key":"bib54","volume-title":"Proceedings of ICLR","author":"Lei Yu","year":"2017"},{"key":"bib55","volume-title":"Proceedings of EMNLP","author":"Lei Yu","year":"2016"},{"key":"bib56","author":"Zellers Rowan","year":"2019","journal-title":"arXiv preprint arXiv:1905.12616"},{"key":"bib57","volume":"1902","author":"Zhang Haoyu","year":"2019","journal-title":"CoRR"},{"key":"bib58","volume-title":"Proceedings of EMNLP","author":"Zhang Jiacheng","year":"2018"},{"key":"bib59","volume":"1908","author":"Ziegler Zachary M.","year":"2019","journal-title":"CoRR"}],"container-title":["Transactions of the Association for Computational Linguistics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mitpressjournals.org\/doi\/pdf\/10.1162\/tacl_a_00319","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2021,3,12]],"date-time":"2021-03-12T21:39:39Z","timestamp":1615585179000},"score":1,"resource":{"primary":{"URL":"https:\/\/direct.mit.edu\/tacl\/article\/96457"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2020,12]]},"references-count":59,"alternative-id":["10.1162\/tacl_a_00319"],"URL":"https:\/\/doi.org\/10.1162\/tacl_a_00319","relation":{},"ISSN":["2307-387X"],"issn-type":[{"value":"2307-387X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2020,12]]}}}