{"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/contextual-parameter-generation-for-universal","title":"Contextual Parameter Generation for Universal Neural Machine Translation","arxiv_id":"1808.08493","date":"2018-08-26","proceeding":"EMNLP 2018 10","authors":["Emmanouil Antonios Platanios","Mrinmaya Sachan","Graham Neubig","Tom Mitchell"],"abstract":"We propose a simple modification to existing neural machine translation (NMT)\nmodels that enables using a single universal model to translate between\nmultiple languages while allowing for language specific parameterization, and\nthat can also be used for domain adaptation. Our approach requires no changes\nto the model architecture of a standard NMT system, but instead introduces a\nnew component, the contextual parameter generator (CPG), that generates the\nparameters of the system (e.g., weights in a neural network). This parameter\ngenerator accepts source and target language embeddings as input, and generates\nthe parameters for the encoder and the decoder, respectively. The rest of the\nmodel remains unchanged and is shared across all languages. We show how this\nsimple modification enables the system to use monolingual data for training and\nalso perform zero-shot translation. We further show it is able to surpass\nstate-of-the-art performance for both the IWSLT-15 and IWSLT-17 datasets and\nthat the learned language embeddings are able to uncover interesting\nrelationships between languages.","url_abs":"http://arxiv.org/abs/1808.08493v1","url_pdf":"http://arxiv.org/pdf/1808.08493v1.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":"contextual-parameter-generation-for-universal","repo_url":"https://github.com/eaplatanios/symphony-mt","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":{"status":"ok","spdx":"Apache-2.0"}}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"domain-adaptation","task_name":"Domain Adaptation"},{"task_slug":"machine-translation","task_name":"Machine Translation"},{"task_slug":"nmt","task_name":"NMT"},{"task_slug":"translation","task_name":"Translation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1808.08493","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}