论文信息 - Spectral Ergodicity in Deep Learning Architectures via Surrogate Random Matrices

Spectral Ergodicity in Deep Learning Architectures via Surrogate Random Matrices

Mehmet Süzen, Cornelius Weber, and Joan J. Cerdà INM-6, Forschungszentrum Juelich and Department of Neurobiology, Ludwig-Maximilians-Universität München, Germany Department of Informatics, Faculty of Mathematics and Natural Sciences, University of Hamburg, Germany Department of Physics, University of the Balearic Islands, Crta. de Valldemossa, km 7.5, 07122 Palma (Illes Balears), Spain (Dated: July 12, 2017)

[1] M. Stephanov,et al. Random Matrices , 2005, hep-ph/0509286.

[2] Ole Winther,et al. Information Theory Workshop (ITW), 2014 IEEE , 2014 .

[3] A. Sayed,et al. Foundations and Trends ® in Machine Learning > Vol 7 > Issue 4-5 Ordering Info About Us Alerts Contact Help Log in Adaptation , Learning , and Optimization over Networks , 2011 .

[4] Dinh Phung,et al. Journal of Machine Learning Research: Preface , 2014 .

[5] F. Haake. Quantum signatures of chaos , 1991 .