{"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/learning-ordered-representations-with-nested","title":"Learning Ordered Representations with Nested Dropout","arxiv_id":"1402.0915","date":"2014-02-05","proceeding":null,"authors":["Oren Rippel","Michael A. Gelbart","Ryan P. Adams"],"abstract":"In this paper, we study ordered representations of data in which different\ndimensions have different degrees of importance. To learn these representations\nwe introduce nested dropout, a procedure for stochastically removing coherent\nnested sets of hidden units in a neural network. We first present a sequence of\ntheoretical results in the simple case of a semi-linear autoencoder. We\nrigorously show that the application of nested dropout enforces identifiability\nof the units, which leads to an exact equivalence with PCA. We then extend the\nalgorithm to deep models and demonstrate the relevance of ordered\nrepresentations to a number of applications. Specifically, we use the ordered\nproperty of the learned codes to construct hash-based data structures that\npermit very fast retrieval, achieving retrieval in time logarithmic in the\ndatabase size and independent of the dimensionality of the representation. This\nallows codes that are hundreds of times longer than currently feasible for\nretrieval. We therefore avoid the diminished quality associated with short\ncodes, while still performing retrieval that is competitive in speed with\nexisting methods. We also show that ordered representations are a promising way\nto learn adaptive compression for efficient online data reconstruction.","url_abs":"http://arxiv.org/abs/1402.0915v1","url_pdf":"http://arxiv.org/pdf/1402.0915v1.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":"learning-ordered-representations-with-nested","repo_url":"https://github.com/PhilippeNguyen/nested_dropout","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}}],"tasks":[{"task_slug":"retrieval","task_name":"Retrieval"}],"methods":[{"method_slug":"pca","method_name":"PCA"},{"method_slug":"speed","method_name":"SPEED"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":"https://syntology.ai/paper/1402.0915","atlas_url":"https://app.syntology.ai/?focus=1402.0915","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}