{"id":1699,"date":"2022-12-13T03:04:08","date_gmt":"2022-12-13T02:04:08","guid":{"rendered":"https:\/\/www-staging.systransoft.com\/papers\/maxime-sit-quis-dolore-rerum-ea\/"},"modified":"2023-11-16T18:22:31","modified_gmt":"2023-11-16T17:22:31","slug":"non-autoregressive-machine-translation-with-translation-memories","status":"publish","type":"paper","link":"https:\/\/www-staging.systransoft.com\/fr\/ressources\/papers-and-publications\/non-autoregressive-machine-translation-with-translation-memories\/","title":{"rendered":"Traduction automatique non autor\u00e9gressive avec m\u00e9moires de traduction"},"content":{"rendered":"Non-autoregressive machine translation (NAT) has recently made great progress. However, most works to date have focused on standard translation tasks, even though some edit-based NAT models, such as the Levenshtein Transformer (LevT), seem well suited to translate with a Translation Memory (TM). This is the scenario considered here. We first analyze the vanilla LevT model and explain why it does not do well in this setting. We then propose a new variant, TM-LevT, and show how to effectively train this model. By modifying the data presentation and introducing an extra deletion operation, we obtain performance that are on par with an autoregressive approach, while reducing the decoding load. We also show that incorporating TMs during training dispenses to use knowledge distillation, a well-known trick used to mitigate the multimodality issue.","protected":false},"excerpt":{"rendered":"<p>La traduction automatique non autor\u00e9gressive (NAT) a r\u00e9cemment fait de grands progr\u00e8s. Cependant, la plupart des travaux r\u00e9alis\u00e9s \u00e0 ce jour ont port\u00e9 sur des t\u00e2ches de traduction standard, m\u00eame si certains mod\u00e8les NAT bas\u00e9s sur l'\u00e9dition, tels que le Levenshtein Transformer (LevT), semblent bien adapt\u00e9s \u00e0 la traduction avec une m\u00e9moire de traduction (TM). C'est le sc\u00e9nario envisag\u00e9 ici. Nous analysons d'abord le mod\u00e8le LevT vanille... <a href=\"https:\/\/www-staging.systransoft.com\/fr\/ressources\/papers-and-publications\/non-autoregressive-machine-translation-with-translation-memories\/\">Suite<\/a><\/p>","protected":false},"featured_media":0,"template":"","class_list":["post-1699","paper","type-paper","status-publish","hentry"],"acf":[],"_links":{"self":[{"href":"https:\/\/www-staging.systransoft.com\/fr\/wp-json\/wp\/v2\/paper\/1699","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www-staging.systransoft.com\/fr\/wp-json\/wp\/v2\/paper"}],"about":[{"href":"https:\/\/www-staging.systransoft.com\/fr\/wp-json\/wp\/v2\/types\/paper"}],"wp:attachment":[{"href":"https:\/\/www-staging.systransoft.com\/fr\/wp-json\/wp\/v2\/media?parent=1699"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}