{"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/igcv2-interleaved-structured-sparse","title":"IGCV$2$: Interleaved Structured Sparse Convolutional Neural Networks","arxiv_id":"1804.06202","date":"2018-04-17","proceeding":null,"authors":["Guotian Xie","Jingdong Wang","Ting Zhang","Jian-Huang Lai","Richang Hong","Guo-Jun Qi"],"abstract":"In this paper, we study the problem of designing efficient convolutional\nneural network architectures with the interest in eliminating the redundancy in\nconvolution kernels. In addition to structured sparse kernels, low-rank kernels\nand the product of low-rank kernels, the product of structured sparse kernels,\nwhich is a framework for interpreting the recently-developed interleaved group\nconvolutions (IGC) and its variants (e.g., Xception), has been attracting\nincreasing interests.\n  Motivated by the observation that the convolutions contained in a group\nconvolution in IGC can be further decomposed in the same manner, we present a\nmodularized building block, {IGCV$2$:} interleaved structured sparse\nconvolutions. It generalizes interleaved group convolutions, which is composed\nof two structured sparse kernels, to the product of more structured sparse\nkernels, further eliminating the redundancy. We present the complementary\ncondition and the balance condition to guide the design of structured sparse\nkernels, obtaining a balance among three aspects: model size, computation\ncomplexity and classification accuracy. Experimental results demonstrate the\nadvantage on the balance among these three aspects compared to interleaved\ngroup convolutions and Xception, and competitive performance compared to other\nstate-of-the-art architecture design methods.","url_abs":"http://arxiv.org/abs/1804.06202v1","url_pdf":"http://arxiv.org/pdf/1804.06202v1.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":"igcv2-interleaved-structured-sparse","repo_url":"https://github.com/homles11/IGCV3","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"igcv2-interleaved-structured-sparse","repo_url":"https://github.com/xxradon/IGCV3-pytorch","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[],"methods":[{"method_slug":"1x1-convolution","method_name":"1x1 Convolution"},{"method_slug":"average-pooling","method_name":"Average Pooling"},{"method_slug":"convolution","method_name":"Convolution"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"depthwise-convolution","method_name":"Depthwise Convolution"},{"method_slug":"depthwise-separable-convolution","method_name":"Depthwise Separable Convolution"},{"method_slug":"global-average-pooling","method_name":"Global Average Pooling"},{"method_slug":"max-pooling","method_name":"Max Pooling"},{"method_slug":"pointwise-convolution","method_name":"Pointwise Convolution"},{"method_slug":"relu","method_name":"ReLU"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"softmax","method_name":"Softmax"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1804.06202","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}