{"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/interpretable-convolutional-neural-networks-1","title":"Interpretable Convolutional Neural Networks via Feedforward Design","arxiv_id":"1810.02786","date":"2018-10-05","proceeding":null,"authors":["C. -C. Jay Kuo","Min Zhang","Siyang Li","Jiali Duan","Yueru Chen"],"abstract":"The model parameters of convolutional neural networks (CNNs) are determined\nby backpropagation (BP). In this work, we propose an interpretable feedforward\n(FF) design without any BP as a reference. The FF design adopts a data-centric\napproach. It derives network parameters of the current layer based on data\nstatistics from the output of the previous layer in a one-pass manner. To\nconstruct convolutional layers, we develop a new signal transform, called the\nSaab (Subspace Approximation with Adjusted Bias) transform. It is a variant of\nthe principal component analysis (PCA) with an added bias vector to annihilate\nactivation's nonlinearity. Multiple Saab transforms in cascade yield multiple\nconvolutional layers. As to fully-connected (FC) layers, we construct them\nusing a cascade of multi-stage linear least squared regressors (LSRs). The\nclassification and robustness (against adversarial attacks) performances of BP-\nand FF-designed CNNs applied to the MNIST and the CIFAR-10 datasets are\ncompared. Finally, we comment on the relationship between BP and FF designs.","url_abs":"http://arxiv.org/abs/1810.02786v2","url_pdf":"http://arxiv.org/pdf/1810.02786v2.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":"interpretable-convolutional-neural-networks-1","repo_url":"https://github.com/yifan-fanyi/Pixelhop","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null},{"paper_slug":"interpretable-convolutional-neural-networks-1","repo_url":"https://github.com/yifan-fanyi/myMCL-Realization","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}