{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,23]],"date-time":"2026-04-23T14:55:48Z","timestamp":1776956148381,"version":"3.51.4"},"reference-count":45,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"1","license":[{"start":{"date-parts":[[2020,1,1]],"date-time":"2020-01-01T00:00:00Z","timestamp":1577836800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2020,1,1]],"date-time":"2020-01-01T00:00:00Z","timestamp":1577836800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2020,1,1]],"date-time":"2020-01-01T00:00:00Z","timestamp":1577836800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100000923","name":"Australian Research Council","doi-asserted-by":"publisher","award":["DP180103232"],"award-info":[{"award-number":["DP180103232"]}],"id":[{"id":"10.13039\/501100000923","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100000923","name":"Australian Research Council","doi-asserted-by":"publisher","award":["CE140100016"],"award-info":[{"award-number":["CE140100016"]}],"id":[{"id":"10.13039\/501100000923","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100000923","name":"Australian Research Council","doi-asserted-by":"publisher","award":["FL130100102"],"award-info":[{"award-number":["FL130100102"]}],"id":[{"id":"10.13039\/501100000923","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Pattern Anal. Mach. Intell."],"published-print":{"date-parts":[[2020,1,1]]},"DOI":"10.1109\/tpami.2018.2876413","type":"journal-article","created":{"date-parts":[[2018,10,16]],"date-time":"2018-10-16T18:54:45Z","timestamp":1539716085000},"page":"15-26","source":"Crossref","is-referenced-by-count":20,"title":["Approximate Fisher Information Matrix to Characterize the Training of Deep Neural Networks"],"prefix":"10.1109","volume":"42","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-9965-4511","authenticated-orcid":false,"given":"Zhibin","family":"Liao","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Tom","family":"Drummond","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Ian","family":"Reid","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5571-6220","authenticated-orcid":false,"given":"Gustavo","family":"Carneiro","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref39","article-title":"Mini-batch stochastic gradient descent with dynamic sample sizes","author":"metel","year":"2017"},{"key":"ref38","article-title":"Adaptive sampling strategies for stochastic optimization","author":"bollapragada","year":"2017"},{"key":"ref33","first-page":"1","article-title":"Don't decay the learning rate, increase the batch size","author":"smith","year":"2018","journal-title":"Proc Int Conf Learn Representations"},{"key":"ref32","first-page":"1","article-title":"A Bayesian perspective on generalization and stochastic gradient descent","author":"smith","year":"2017","journal-title":"Proc 2nd Workshop Bayesian Deep Learn"},{"key":"ref31","first-page":"1","article-title":"Qualitatively characterizing neural network optimization problems","author":"goodfellow","year":"2015","journal-title":"Proc Int Conf Learn Representations"},{"key":"ref30","first-page":"1","article-title":"Adam: A method for stochastic optimization","author":"kingma","year":"2015","journal-title":"Proc Int Conf Learn Representations"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.1007\/s10107-012-0572-5"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.1137\/110830629"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-015-0816-y"},{"key":"ref34","article-title":"Accurate, large minibatch SGD: Training ImageNet in 1 hour","author":"goyal","year":"2017"},{"key":"ref10","first-page":"1","article-title":"Entropy-SGD: Biasing gradient descent into wide valleys","author":"chaudhari","year":"2017","journal-title":"Proc Int Conf Learn Representations"},{"key":"ref40","first-page":"1504","article-title":"Big batch SGD: Automated inference using adaptive batch sizes","author":"de","year":"2017","journal-title":"Proc Int Conf Artif Intell Statist"},{"key":"ref11","article-title":"New insights and perspectives on the natural gradient method","author":"martens","year":"2014"},{"key":"ref12","first-page":"1","article-title":"Empirical analysis of the hessian of over-parametrized neural networks","author":"sagun","year":"2018","journal-title":"Proc Workshop Int Conf Learn Represent"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-01424-7_39"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1137\/16M1080173"},{"key":"ref15","article-title":"Reading digits in natural images with unsupervised feature learning","author":"netzer","year":"2011","journal-title":"Proc NIPS Workshop on Deep Learning and Unsupervised Feature Learning"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1109\/5.726791"},{"key":"ref17","author":"fletcher","year":"2013","journal-title":"Practical Methods of Optimization"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1137\/0916069"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1137\/S1052623494268522"},{"key":"ref28","first-page":"1","article-title":"Training neural networks with stochastic hessian-free optimization","author":"kiros","year":"2013","journal-title":"Proc Int Conf Learn Representations"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1214\/aoms\/1177729586"},{"key":"ref27","first-page":"735","article-title":"Deep learning via Hessian-free optimization","author":"martens","year":"2010","journal-title":"Proc Int Conf Mach Learn"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.243"},{"key":"ref6","article-title":"The loss surface of residual networks: Ensembles and the role of batch normalization","author":"littwin","year":"2016"},{"key":"ref29","article-title":"ADADELTA: An adaptive learning rate method","author":"zeiler","year":"2012"},{"key":"ref5","first-page":"1","article-title":"On large-batch training for deep learning: Generalization gap and sharp minima","author":"keskar","year":"2017","journal-title":"Proc Int Conf Learn Representations"},{"key":"ref8","first-page":"1246","article-title":"Gradient descent only converges to minimizers","author":"lee","year":"2016","journal-title":"Proc Conf Learn Theory"},{"key":"ref7","article-title":"No bad local minima: Data independent training error guarantees for multilayer neural networks","author":"soudry","year":"2016"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-46493-0_39"},{"key":"ref9","first-page":"7","article-title":"Learning multiple layers of features from tiny images","volume":"1","author":"krizhevsky","year":"2009"},{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1007\/3-540-44668-0_4"},{"key":"ref45","first-page":"1","article-title":"Torch7: A Matlab-like environment for machine learning","author":"collobert","year":"2011","journal-title":"NIPS Workshop BigLearn"},{"key":"ref22","first-page":"26","article-title":"Lecture 6.5-rmsprop: Divide the gradient by a running average of its recent magnitude","volume":"4","author":"tieleman","year":"2012","journal-title":"Neural Netw Mach Learning"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1162\/089976698300017746"},{"key":"ref42","first-page":"448","article-title":"Batch normalization: Accelerating deep network training by reducing internal covariate shift","author":"ioffe","year":"2015","journal-title":"Proc Int Conf Mach Learn"},{"key":"ref24","article-title":"Singularity of the hessian in deep learning","author":"sagun","year":"2016"},{"key":"ref41","first-page":"807","article-title":"Rectified linear units improve restricted Boltzmann machines","author":"nair","year":"2010","journal-title":"Proc Int Conf Mach Learn"},{"key":"ref23","first-page":"2121","article-title":"Adaptive subgradient methods for online learning and stochastic optimization","volume":"12","author":"duchi","year":"2011","journal-title":"J Mach Learn Res"},{"key":"ref44","article-title":"Training and investigating residual nets","year":"2017"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1994.6.1.147"},{"key":"ref43","author":"wilkinson","year":"1965","journal-title":"The Algebraic Eigenvalue Problem"},{"key":"ref25","first-page":"67","author":"wright","year":"1999","journal-title":"Numerical Optimization"}],"container-title":["IEEE Transactions on Pattern Analysis and Machine Intelligence"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/34\/8922815\/08493265.pdf?arnumber=8493265","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,4,27]],"date-time":"2022-04-27T13:59:28Z","timestamp":1651067968000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/8493265\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2020,1,1]]},"references-count":45,"journal-issue":{"issue":"1"},"URL":"https:\/\/doi.org\/10.1109\/tpami.2018.2876413","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":[[2020,1,1]]}}}