{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,6]],"date-time":"2026-03-06T19:02:36Z","timestamp":1772823756534,"version":"3.50.1"},"reference-count":79,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","license":[{"start":{"date-parts":[[2022,1,1]],"date-time":"2022-01-01T00:00:00Z","timestamp":1640995200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/legalcode"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Pattern Anal. Mach. Intell."],"published-print":{"date-parts":[[2022]]},"DOI":"10.1109\/tpami.2022.3200344","type":"journal-article","created":{"date-parts":[[2022,8,22]],"date-time":"2022-08-22T19:58:19Z","timestamp":1661198299000},"page":"1-17","source":"Crossref","is-referenced-by-count":16,"title":["Quantifying the Knowledge in a DNN to Explain Knowledge Distillation for Classification"],"prefix":"10.1109","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-6108-2738","authenticated-orcid":false,"given":"Quanshi","family":"Zhang","sequence":"first","affiliation":[{"name":"Department of Computer Science and Engineering, the John Hopcroft Center, and the MoE Key Lab of Artificial Intelligence, AI Institute, Shanghai Jiao Tong University, China"}]},{"given":"Xu","family":"Cheng","sequence":"additional","affiliation":[{"name":"School of electronic information and electrical engineering, Shanghai Jiao Tong University, China"}]},{"given":"Yilan","family":"Chen","sequence":"additional","affiliation":[{"name":"School of electronic information and electrical engineering, Shanghai Jiao Tong University, China"}]},{"given":"Zhefan","family":"Rao","sequence":"additional","affiliation":[{"name":"School of electronic information and electrical engineering, Shanghai Jiao Tong University, China"}]}],"member":"263","reference":[{"key":"ref1","first-page":"6241","article-title":"Spectrally-normalized margin bounds for neural networks","author":"Bartlett","year":"2017","journal-title":"Adv. Neural Informat. Process. Syst."},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.354"},{"key":"ref3","first-page":"573","article-title":"Invertible residual networks","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Behrmann"},{"key":"ref4","first-page":"1","article-title":"Approximating CNNs with bag-of-local-features models works surprisingly well on imagenet","volume-title":"Proc. 7th Int. Conf. Learn. Representations","author":"Brendel"},{"key":"ref5","first-page":"1","article-title":"The intriguing role of module criticality in the generalization of deep networks","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Chatterji"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1109\/WACV.2018.00097"},{"key":"ref7","first-page":"8928","article-title":"This looks like that: Deep learning for interpretable image recognition","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Chen"},{"key":"ref8","first-page":"9916","article-title":"Residual flows for invertible generative modeling","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Chen"},{"key":"ref9","first-page":"2172","article-title":"InfoGAN: Interpretable representation learning by information maximizing generative adversarial nets","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Chen"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR42600.2020.01294"},{"key":"ref11","article-title":"Bert: Pre-training of deep bidirectional transformers for language understanding","author":"Devlin","year":"2018"},{"key":"ref12","first-page":"1019","article-title":"Sharp minima can generalize for deep nets","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Dinh"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.522"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-009-0275-4"},{"key":"ref15","article-title":"Transferring knowledge across learning processes","author":"Flennerhag","year":"2018"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00910"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2017.371"},{"key":"ref18","article-title":"Stiffness: A new perspective on generalization in neural networks","author":"Fort","year":"2019"},{"key":"ref19","article-title":"Born again neural networks","author":"Furlanello","year":"2018"},{"key":"ref20","first-page":"8789","article-title":"Loss surfaces, mode connectivity, and fast ensembling of dnns","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Garipov"},{"key":"ref21","first-page":"2299","article-title":"Estimating information flow in deep neural networks","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Goldfeld"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1007\/s10994-020-05929-w"},{"key":"ref23","first-page":"2454","article-title":"Towards a deep and unified understanding of deep neural models in nlp","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Guan"},{"key":"ref24","article-title":"beta-VAE: Learning basic visual concepts with a constrained variational framework","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Higgins"},{"key":"ref25","article-title":"Distilling the knowledge in a neural network","author":"Hinton","year":"2015"},{"key":"ref26","first-page":"1","article-title":"Matrix capsules with em routing","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Hinton"},{"key":"ref27","first-page":"9737","article-title":"A benchmark for interpretability methods in deep neural networks","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Hooker"},{"key":"ref28","article-title":"i-RevNet: Deep invertible networks","author":"Jacobsen","year":"2018"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2019.00505"},{"key":"ref31","article-title":"On large-batch training for deep learning: Generalization gap and sharp minima","author":"Keskar","year":"2016"},{"key":"ref32","article-title":"Interpretability beyond feature attribution: Quantitative testing with concept activation vectors (TCAV)","author":"Kim","year":"2017"},{"key":"ref33","article-title":"Learning how to explain neural networks: Patternnet and patternattribution","author":"Kindermans","year":"2017"},{"key":"ref34","first-page":"1097","article-title":"Imagenet classification with deep convolutional neural networks","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Krizhevsky"},{"issue":"7","key":"ref35","article-title":"Tiny imagenet visual recognition challenge","volume":"7","author":"Le","year":"2015","journal-title":"CS 231N"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR42600.2020.01465"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-10602-1_48"},{"key":"ref38","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.00271"},{"key":"ref39","first-page":"1","article-title":"Unifying distillation and privileged information","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Lopez-Paz"},{"key":"ref40","article-title":"Quantifying layerwise information discarding of neural networks","author":"Ma","year":"2019"},{"key":"ref41","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2015.7299155"},{"key":"ref42","article-title":"Why distillation helps: A statistical perspective","author":"Menon","year":"2020"},{"key":"ref43","first-page":"1","article-title":"A PAC-Bayesian approach to spectrally-normalized margin bounds for neural networks","volume-title":"Proc. Int. Conf. on Learn. Representations","author":"Neyshabur"},{"key":"ref44","first-page":"1376","article-title":"Norm-based capacity control in neural networks","volume-title":"Proc. Conf. Learn. Theory","author":"Neyshabur"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.1109\/SP.2016.41"},{"key":"ref46","first-page":"5142","article-title":"Towards understanding knowledge distillation","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Phuong"},{"key":"ref47","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.16"},{"key":"ref48","article-title":"PointNet++: Deep hierarchical feature learning on point sets in a metric space","author":"Qi","year":"2017"},{"key":"ref49","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/D16-1264"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/N16-3020"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v32i1.11491"},{"key":"ref52","article-title":"Fitnets: Hints for thin deep nets","author":"Romero","year":"2014"},{"key":"ref53","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-015-0816-y"},{"key":"ref54","first-page":"3856","article-title":"Dynamic routing between capsules","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Sabour"},{"key":"ref55","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2017.74"},{"key":"ref56","article-title":"Overfeat: Integrated recognition, localization and detection using convolutional networks","author":"Sermanet","year":"2013"},{"key":"ref57","doi-asserted-by":"publisher","DOI":"10.24963\/ijcai.2021\/409"},{"key":"ref58","article-title":"Opening the black box of deep neural networks via information","author":"Shwartz-Ziv","year":"2017"},{"key":"ref59","article-title":"Deep inside convolutional networks: Visualising image classification models and saliency maps","author":"Simonyan","year":"2017"},{"key":"ref60","first-page":"1","article-title":"Very deep convolutional networks for large-scale image recognition","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Simonyan"},{"key":"ref61","first-page":"5827","article-title":"A tail-index analysis of stochastic gradient noise in deep neural networks","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Simsekli"},{"key":"ref62","first-page":"1631","article-title":"Recursive deep models for semantic compositionality over a sentiment treebank","volume-title":"Proc. Conf. Empir. Methods Natural Lang. Process.","author":"Socher"},{"key":"ref63","article-title":"Understanding and improving knowledge distillation","author":"Tang","year":"2020"},{"key":"ref64","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00121"},{"key":"ref65","article-title":"The caltech-ucsd birds-200\u20132011 dataset","author":"Wah","year":"2011"},{"key":"ref66","doi-asserted-by":"publisher","DOI":"10.1145\/3326362"},{"key":"ref67","article-title":"Evaluating the robustness of neural networks: An extreme value theory approach","author":"Weng","year":"2018"},{"key":"ref68","doi-asserted-by":"publisher","DOI":"10.7551\/mitpress\/11909.003.0037"},{"key":"ref69","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.00985"},{"key":"ref70","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2015.7298801"},{"issue":"7","key":"ref71","doi-asserted-by":"crossref","first-page":"1639","DOI":"10.1109\/TPAMI.2017.2728788","article-title":"Learning and inferring \u201cdark matter","volume":"40","author":"Xie","year":"2018","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"ref72","first-page":"2524","article-title":"Information-theoretic analysis of generalization capability of learning algorithms","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Xu"},{"key":"ref73","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.754"},{"key":"ref74","article-title":"Understanding neural networks through deep visualization","author":"Yosinski","year":"2015"},{"key":"ref75","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR42600.2020.00396"},{"key":"ref76","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-10590-1_53"},{"key":"ref77","doi-asserted-by":"publisher","DOI":"10.1145\/3446776"},{"key":"ref78","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2020.2982882"},{"key":"ref79","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.319"}],"container-title":["IEEE Transactions on Pattern Analysis and Machine Intelligence"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/34\/4359286\/09864081.pdf?arnumber=9864081","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,3,2]],"date-time":"2024-03-02T05:26:37Z","timestamp":1709357197000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/9864081\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022]]},"references-count":79,"URL":"https:\/\/doi.org\/10.1109\/tpami.2022.3200344","relation":{},"ISSN":["0162-8828","2160-9292","1939-3539"],"issn-type":[{"value":"0162-8828","type":"print"},{"value":"2160-9292","type":"electronic"},{"value":"1939-3539","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022]]}}}