{"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/adjusting-for-dropout-variance-in-batch","title":"Adjusting for Dropout Variance in Batch Normalization and Weight Initialization","arxiv_id":"1607.02488","date":"2016-07-08","proceeding":null,"authors":["Dan Hendrycks","Kevin Gimpel"],"abstract":"We show how to adjust for the variance introduced by dropout with corrections\nto weight initialization and Batch Normalization, yielding higher accuracy.\nThough dropout can preserve the expected input to a neuron between train and\ntest, the variance of the input differs. We thus propose a new weight\ninitialization by correcting for the influence of dropout rates and an\narbitrary nonlinearity's influence on variance through simple corrective\nscalars. Since Batch Normalization trained with dropout estimates the variance\nof a layer's incoming distribution with some inputs dropped, the variance also\ndiffers between train and test. After training a network with Batch\nNormalization and dropout, we simply update Batch Normalization's variance\nmoving averages with dropout off and obtain state of the art on CIFAR-10 and\nCIFAR-100 without data augmentation.","url_abs":"http://arxiv.org/abs/1607.02488v2","url_pdf":"http://arxiv.org/pdf/1607.02488v2.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":"adjusting-for-dropout-variance-in-batch","repo_url":"https://github.com/hendrycks/init","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"data-augmentation","task_name":"Data Augmentation"}],"methods":[{"method_slug":"batch-normalization","method_name":"Batch Normalization"},{"method_slug":"dropout","method_name":"Dropout"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1607.02488","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}