{"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/can-we-gain-more-from-orthogonality","title":"Can We Gain More from Orthogonality Regularizations in Training Deep CNNs?","arxiv_id":"1810.09102","date":"2018-10-22","proceeding":"NeurIPS 2018","authors":["Nitin Bansal","Xiaohan Chen","Zhangyang Wang"],"abstract":"This paper seeks to answer the question: as the (near-) orthogonality of\nweights is found to be a favorable property for training deep convolutional\nneural networks, how can we enforce it in more effective and easy-to-use ways?\nWe develop novel orthogonality regularizations on training deep CNNs, utilizing\nvarious advanced analytical tools such as mutual coherence and restricted\nisometry property. These plug-and-play regularizations can be conveniently\nincorporated into training almost any CNN without extra hassle. We then\nbenchmark their effects on state-of-the-art models: ResNet, WideResNet, and\nResNeXt, on several most popular computer vision datasets: CIFAR-10, CIFAR-100,\nSVHN and ImageNet. We observe consistent performance gains after applying those\nproposed regularizations, in terms of both the final accuracies achieved, and\nfaster and more stable convergences. We have made our codes and pre-trained\nmodels publicly available:\nhttps://github.com/nbansal90/Can-we-Gain-More-from-Orthogonality.","url_abs":"http://arxiv.org/abs/1810.09102v1","url_pdf":"http://arxiv.org/pdf/1810.09102v1.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":"can-we-gain-more-from-orthogonality","repo_url":"https://github.com/nbansal90/Can-we-Gain-More-from-Orthogonality","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[],"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":"bottleneck-residual-block","method_name":"Bottleneck Residual Block"},{"method_slug":"convolution","method_name":"Convolution"},{"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":"residual-block","method_name":"Residual Block"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"wide-residual-block","method_name":"Wide Residual Block"},{"method_slug":"wideresnet","method_name":"WideResNet"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1810.09102","atlas_url":"https://app.syntology.ai/?focus=1810.09102","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}