{"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/scalable-convolutional-dictionary-learning","title":"Scalable Convolutional Dictionary Learning with Constrained Recurrent Sparse Auto-encoders","arxiv_id":"1807.04734","date":"2018-07-12","proceeding":null,"authors":["Bahareh Tolooshams","Sourav Dey","Demba Ba"],"abstract":"Given a convolutional dictionary underlying a set of observed signals, can a\ncarefully designed auto-encoder recover the dictionary in the presence of\nnoise? We introduce an auto-encoder architecture, termed constrained recurrent\nsparse auto-encoder (CRsAE), that answers this question in the affirmative.\nGiven an input signal and an approximate dictionary, the encoder finds a sparse\napproximation using FISTA. The decoder reconstructs the signal by applying the\ndictionary to the output of the encoder. The encoder and decoder in CRsAE\nparallel the sparse-coding and dictionary update steps in optimization-based\nalternating-minimization schemes for dictionary learning. As such, the\nparameters of the encoder and decoder are not independent, a constraint which\nwe enforce for the first time. We derive the back-propagation algorithm for\nCRsAE. CRsAE is a framework for blind source separation that, only knowing the\nnumber of sources (dictionary elements), and assuming sparsely-many can\noverlap, is able to separate them. We demonstrate its utility in the context of\nspike sorting, a source separation problem in computational neuroscience. We\ndemonstrate the ability of CRsAE to recover the underlying dictionary and\ncharacterize its sensitivity as a function of SNR.","url_abs":"http://arxiv.org/abs/1807.04734v1","url_pdf":"http://arxiv.org/pdf/1807.04734v1.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":"scalable-convolutional-dictionary-learning","repo_url":"https://github.com/ds2p/crsae","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"dictionary-learning","task_name":"Dictionary Learning"},{"task_slug":"spike-sorting","task_name":"Spike Sorting"},{"task_slug":"blind-source-separation","task_name":"blind source separation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}