{"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/sequence-to-sequence-models-can-directly","title":"Sequence-to-Sequence Models Can Directly Translate Foreign Speech","arxiv_id":"1703.08581","date":"2017-03-24","proceeding":null,"authors":["Ron J. Weiss","Jan Chorowski","Navdeep Jaitly","Yonghui Wu","Zhifeng Chen"],"abstract":"We present a recurrent encoder-decoder deep neural network architecture that\ndirectly translates speech in one language into text in another. The model does\nnot explicitly transcribe the speech into text in the source language, nor does\nit require supervision from the ground truth source language transcription\nduring training. We apply a slightly modified sequence-to-sequence with\nattention architecture that has previously been used for speech recognition and\nshow that it can be repurposed for this more complex task, illustrating the\npower of attention-based models. A single model trained end-to-end obtains\nstate-of-the-art performance on the Fisher Callhome Spanish-English speech\ntranslation task, outperforming a cascade of independently trained\nsequence-to-sequence speech recognition and machine translation models by 1.8\nBLEU points on the Fisher test set. In addition, we find that making use of the\ntraining data in both languages by multi-task training sequence-to-sequence\nspeech translation and recognition models with a shared encoder network can\nimprove performance by a further 1.4 BLEU points.","url_abs":"http://arxiv.org/abs/1703.08581v2","url_pdf":"http://arxiv.org/pdf/1703.08581v2.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":[{"paper_slug":"sequence-to-sequence-models-can-directly","repo_url":"https://github.com/colaprograms/speechify","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"machine-translation","task_name":"Machine Translation"},{"task_slug":"sequence-to-sequence-speech-recognition","task_name":"Sequence-To-Sequence Speech Recognition"},{"task_slug":"speech-recognition","task_name":"Speech Recognition"},{"task_slug":"translation","task_name":"Translation"},{"task_slug":"speech-recognition-1","task_name":"speech-recognition"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1703.08581","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}