{"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/improving-generalization-of-deep-networks-for","title":"Improving Generalization of Deep Networks for Inverse Reconstruction of Image Sequences","arxiv_id":"1903.02948","date":"2019-03-05","proceeding":null,"authors":["Sandesh Ghimire","Prashnna Kumar Gyawali","Jwala Dhamala","John L. Sapp","Milan Horacek","Linwei Wang"],"abstract":"Deep learning networks have shown state-of-the-art performance in many image\nreconstruction problems. However, it is not well understood what properties of\nrepresentation and learning may improve the generalization ability of the\nnetwork. In this paper, we propose that the generalization ability of an\nencoder-decoder network for inverse reconstruction can be improved in two\nmeans. First, drawing from analytical learning theory, we theoretically show\nthat a stochastic latent space will improve the ability of a network to\ngeneralize to test data outside the training distribution. Second, following\nthe information bottleneck principle, we show that a latent representation\nminimally informative of the input data will help a network generalize to\nunseen input variations that are irrelevant to the output reconstruction.\nTherefore, we present a sequence image reconstruction network optimized by a\nvariational approximation of the information bottleneck principle with\nstochastic latent space. In the application setting of reconstructing the\nsequence of cardiac transmembrane potential from bodysurface potential, we\nassess the two types of generalization abilities of the presented network\nagainst its deterministic counterpart. The results demonstrate that the\ngeneralization ability of an inverse reconstruction network can be improved by\nstochasticity as well as the information bottleneck.","url_abs":"http://arxiv.org/abs/1903.02948v1","url_pdf":"http://arxiv.org/pdf/1903.02948v1.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":"improving-generalization-of-deep-networks-for","repo_url":"https://github.com/sandeshgh/Improving-Generalization","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"},{"task_slug":"image-reconstruction","task_name":"Image Reconstruction"},{"task_slug":"learning-theory","task_name":"Learning Theory"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}