{"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/dense-information-flow-for-neural-machine","title":"Dense Information Flow for Neural Machine Translation","arxiv_id":"1806.00722","date":"2018-06-03","proceeding":"NAACL 2018 6","authors":["Yanyao Shen","Xu Tan","Di He","Tao Qin","Tie-Yan Liu"],"abstract":"Recently, neural machine translation has achieved remarkable progress by\nintroducing well-designed deep neural networks into its encoder-decoder\nframework. From the optimization perspective, residual connections are adopted\nto improve learning performance for both encoder and decoder in most of these\ndeep architectures, and advanced attention connections are applied as well.\nInspired by the success of the DenseNet model in computer vision problems, in\nthis paper, we propose a densely connected NMT architecture (DenseNMT) that is\nable to train more efficiently for NMT. The proposed DenseNMT not only allows\ndense connection in creating new features for both encoder and decoder, but\nalso uses the dense attention structure to improve attention quality. Our\nexperiments on multiple datasets show that DenseNMT structure is more\ncompetitive and efficient.","url_abs":"http://arxiv.org/abs/1806.00722v2","url_pdf":"http://arxiv.org/pdf/1806.00722v2.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":"dense-information-flow-for-neural-machine","repo_url":"https://github.com/yanyao-shen/fairseq","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"pytorch","reach":{"status":"ok","spdx":"NOASSERTION"}}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"machine-translation","task_name":"Machine Translation"},{"task_slug":"nmt","task_name":"NMT"},{"task_slug":"translation","task_name":"Translation"}],"methods":[{"method_slug":"1x1-convolution","method_name":"1x1 Convolution"},{"method_slug":"average-pooling","method_name":"Average Pooling"},{"method_slug":"batch-normalization","method_name":"Batch Normalization"},{"method_slug":"concatenated-skip-connection","method_name":"Concatenated Skip Connection"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"dense-block","method_name":"Dense Block"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"global-average-pooling","method_name":"Global Average Pooling"},{"method_slug":"kaiming-initialization","method_name":"Kaiming Initialization"},{"method_slug":"max-pooling","method_name":"Max Pooling"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"softmax","method_name":"Softmax"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/machine-translation-on-wmt2014-english-german","task":"Machine Translation","dataset":"WMT2014 English-German","model":"DenseNMT","rank_in_archive_order":67,"of":91,"metrics":{"BLEU score":"25.52"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1806.00722","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}