{"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/dynamic-capacity-networks","title":"Dynamic Capacity Networks","arxiv_id":"1511.07838","date":"2015-11-24","proceeding":null,"authors":["Amjad Almahairi","Nicolas Ballas","Tim Cooijmans","Yin Zheng","Hugo Larochelle","Aaron Courville"],"abstract":"We introduce the Dynamic Capacity Network (DCN), a neural network that can\nadaptively assign its capacity across different portions of the input data.\nThis is achieved by combining modules of two types: low-capacity sub-networks\nand high-capacity sub-networks. The low-capacity sub-networks are applied\nacross most of the input, but also provide a guide to select a few portions of\nthe input on which to apply the high-capacity sub-networks. The selection is\nmade using a novel gradient-based attention mechanism, that efficiently\nidentifies input regions for which the DCN's output is most sensitive and to\nwhich we should devote more capacity. We focus our empirical evaluation on the\nCluttered MNIST and SVHN image datasets. Our findings indicate that DCNs are\nable to drastically reduce the number of computations, compared to traditional\nconvolutional neural networks, while maintaining similar or even better\nperformance.","url_abs":"http://arxiv.org/abs/1511.07838v7","url_pdf":"http://arxiv.org/pdf/1511.07838v7.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":"dynamic-capacity-networks","repo_url":"https://github.com/beopst/dcn.tf","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"dynamic-capacity-networks","repo_url":"https://github.com/philqc/Dynamic-Capacity-Network","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok"}}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1511.07838","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}