{"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/self-supervised-knowledge-distillation-using","title":"Self-supervised Knowledge Distillation Using Singular Value Decomposition","arxiv_id":"1807.06819","date":"2018-07-18","proceeding":"ECCV 2018 9","authors":["Seung Hyun Lee","Dae Ha Kim","Byung Cheol Song"],"abstract":"To solve deep neural network (DNN)'s huge training dataset and its high\ncomputation issue, so-called teacher-student (T-S) DNN which transfers the\nknowledge of T-DNN to S-DNN has been proposed. However, the existing T-S-DNN\nhas limited range of use, and the knowledge of T-DNN is insufficiently\ntransferred to S-DNN. To improve the quality of the transferred knowledge from\nT-DNN, we propose a new knowledge distillation using singular value\ndecomposition (SVD). In addition, we define a knowledge transfer as a\nself-supervised task and suggest a way to continuously receive information from\nT-DNN. Simulation results show that a S-DNN with a computational cost of 1/5 of\nthe T-DNN can be up to 1.1\\% better than the T-DNN in terms of classification\naccuracy. Also assuming the same computational cost, our S-DNN outperforms the\nS-DNN driven by the state-of-the-art distillation with a performance advantage\nof 1.79\\%. code is available on https://github.com/sseung0703/SSKD\\_SVD.","url_abs":"http://arxiv.org/abs/1807.06819v1","url_pdf":"http://arxiv.org/pdf/1807.06819v1.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":"self-supervised-knowledge-distillation-using","repo_url":"https://github.com/sseung0703/SSKD","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":{"status":"ok"}},{"paper_slug":"self-supervised-knowledge-distillation-using","repo_url":"https://github.com/sseung0703/SSKD_SVD","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}},{"paper_slug":"self-supervised-knowledge-distillation-using","repo_url":"https://github.com/wnma3mz/KD_Notes","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"ok"}}],"tasks":[{"task_slug":"knowledge-distillation","task_name":"Knowledge Distillation"},{"task_slug":"transfer-learning","task_name":"Transfer Learning"}],"methods":[{"method_slug":"knowledge-distillation","method_name":"Knowledge Distillation"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1807.06819","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}