{"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/modulating-and-attending-the-source-image","title":"Modulating and attending the source image during encoding improves Multimodal Translation","arxiv_id":"1712.03449","date":"2017-12-09","proceeding":null,"authors":["Jean-Benoit Delbrouck","Stéphane Dupont"],"abstract":"We propose a new and fully end-to-end approach for multimodal translation\nwhere the source text encoder modulates the entire visual input processing\nusing conditional batch normalization, in order to compute the most informative\nimage features for our task. Additionally, we propose a new attention mechanism\nderived from this original idea, where the attention model for the visual input\nis conditioned on the source text encoder representations. In the paper, we\ndetail our models as well as the image analysis pipeline. Finally, we report\nexperimental results. They are, as far as we know, the new state of the art on\nthree different test sets.","url_abs":"http://arxiv.org/abs/1712.03449v1","url_pdf":"http://arxiv.org/pdf/1712.03449v1.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":"modulating-and-attending-the-source-image","repo_url":"https://github.com/jbdel/mmt_cbn","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":null}],"tasks":[{"task_slug":"translation","task_name":"Translation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1712.03449","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}