{"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/on-the-properties-of-neural-machine","title":"On the Properties of Neural Machine Translation: Encoder-Decoder Approaches","arxiv_id":"1409.1259","date":"2014-09-03","proceeding":null,"authors":["Kyunghyun Cho","Bart van Merrienboer","Dzmitry Bahdanau","Yoshua Bengio"],"abstract":"Neural machine translation is a relatively new approach to statistical\nmachine translation based purely on neural networks. The neural machine\ntranslation models often consist of an encoder and a decoder. The encoder\nextracts a fixed-length representation from a variable-length input sentence,\nand the decoder generates a correct translation from this representation. In\nthis paper, we focus on analyzing the properties of the neural machine\ntranslation using two models; RNN Encoder--Decoder and a newly proposed gated\nrecursive convolutional neural network. We show that the neural machine\ntranslation performs relatively well on short sentences without unknown words,\nbut its performance degrades rapidly as the length of the sentence and the\nnumber of unknown words increase. Furthermore, we find that the proposed gated\nrecursive convolutional network learns a grammatical structure of a sentence\nautomatically.","url_abs":"http://arxiv.org/abs/1409.1259v2","url_pdf":"http://arxiv.org/pdf/1409.1259v2.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":"on-the-properties-of-neural-machine","repo_url":"https://github.com/AnkurDeria/HSI-Traditional-to-Deep-Models","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}},{"paper_slug":"on-the-properties-of-neural-machine","repo_url":"https://github.com/ictmcg/endef-sigir2022","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}},{"paper_slug":"on-the-properties-of-neural-machine","repo_url":"https://github.com/karpathy/makemore","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"machine-translation","task_name":"Machine Translation"},{"task_slug":"sentence","task_name":"Sentence"},{"task_slug":"translation","task_name":"Translation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1409.1259","atlas_url":"https://app.syntology.ai/?focus=1409.1259","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}