{"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/locally-smoothed-neural-networks","title":"Locally Smoothed Neural Networks","arxiv_id":"1711.08132","date":"2017-11-22","proceeding":null,"authors":["Liang Pang","Yanyan Lan","Jun Xu","Jiafeng Guo","Xue-Qi Cheng"],"abstract":"Convolutional Neural Networks (CNN) and the locally connected layer are\nlimited in capturing the importance and relations of different local receptive\nfields, which are often crucial for tasks such as face verification, visual\nquestion answering, and word sequence prediction. To tackle the issue, we\npropose a novel locally smoothed neural network (LSNN) in this paper. The main\nidea is to represent the weight matrix of the locally connected layer as the\nproduct of the kernel and the smoother, where the kernel is shared over\ndifferent local receptive fields, and the smoother is for determining the\nimportance and relations of different local receptive fields. Specifically, a\nmulti-variate Gaussian function is utilized to generate the smoother, for\nmodeling the location relations among different local receptive fields.\nFurthermore, the content information can also be leveraged by setting the mean\nand precision of the Gaussian function according to the content. Experiments on\nsome variant of MNIST clearly show our advantages over CNN and locally\nconnected layer.","url_abs":"http://arxiv.org/abs/1711.08132v1","url_pdf":"http://arxiv.org/pdf/1711.08132v1.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":"locally-smoothed-neural-networks","repo_url":"https://gitlab.com/pl8787/CaffeAttention","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"face-verification","task_name":"Face Verification"},{"task_slug":"question-answering","task_name":"Question Answering"},{"task_slug":"visual-question-answering-1","task_name":"Visual Question Answering"},{"task_slug":"visual-question-answering","task_name":"Visual Question Answering (VQA)"}],"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}