{"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/training-competitive-binary-neural-networks","title":"Training Competitive Binary Neural Networks from Scratch","arxiv_id":"1812.01965","date":"2018-12-05","proceeding":null,"authors":["Joseph Bethge","Marvin Bornstein","Adrian Loy","Haojin Yang","Christoph Meinel"],"abstract":"Convolutional neural networks have achieved astonishing results in different\napplication areas. Various methods that allow us to use these models on mobile\nand embedded devices have been proposed. Especially binary neural networks are\na promising approach for devices with low computational power. However,\ntraining accurate binary models from scratch remains a challenge. Previous work\noften uses prior knowledge from full-precision models and complex training\nstrategies. In our work, we focus on increasing the performance of binary\nneural networks without such prior knowledge and a much simpler training\nstrategy. In our experiments we show that we are able to achieve\nstate-of-the-art results on standard benchmark datasets. Further, to the best\nof our knowledge, we are the first to successfully adopt a network architecture\nwith dense connections for binary networks, which lets us improve the\nstate-of-the-art even further.","url_abs":"http://arxiv.org/abs/1812.01965v1","url_pdf":"http://arxiv.org/pdf/1812.01965v1.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":"training-competitive-binary-neural-networks","repo_url":"https://github.com/hpi-xnor/BMXNet-v2","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"mxnet","reach":{"status":"ok","spdx":"Apache-2.0"}}],"tasks":[],"methods":[{"method_slug":"dense-connections","method_name":"Dense Connections"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1812.01965","atlas_url":"https://app.syntology.ai/?focus=1812.01965","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}