{"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/multilingual-nmt-with-a-language-independent","title":"Multilingual NMT with a language-independent attention bridge","arxiv_id":"1811.00498","date":"2018-11-01","proceeding":"WS 2019 8","authors":["Raúl Vázquez","Alessandro Raganato","Jörg Tiedemann","Mathias Creutz"],"abstract":"In this paper, we propose a multilingual encoder-decoder architecture capable\nof obtaining multilingual sentence representations by means of incorporating an\nintermediate {\\em attention bridge} that is shared across all languages. That\nis, we train the model with language-specific encoders and decoders that are\nconnected via self-attention with a shared layer that we call attention bridge.\nThis layer exploits the semantics from each language for performing translation\nand develops into a language-independent meaning representation that can\nefficiently be used for transfer learning. We present a new framework for the\nefficient development of multilingual NMT using this model and scheduled\ntraining. We have tested the approach in a systematic way with a multi-parallel\ndata set. We show that the model achieves substantial improvements over strong\nbilingual models and that it also works well for zero-shot translation, which\ndemonstrates its ability of abstraction and transfer learning.","url_abs":"http://arxiv.org/abs/1811.00498v1","url_pdf":"http://arxiv.org/pdf/1811.00498v1.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":"multilingual-nmt-with-a-language-independent","repo_url":"https://github.com/Helsinki-NLP/OpenNMT-py","is_official":0,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok","spdx":"MIT"}}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"nmt","task_name":"NMT"},{"task_slug":"sentence","task_name":"Sentence"},{"task_slug":"transfer-learning","task_name":"Transfer Learning"},{"task_slug":"translation","task_name":"Translation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1811.00498","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}