Current browse context:
stat.ML
Change to browse by:
References & Citations
Statistics > Machine Learning
Title: Spectra of the Conjugate Kernel and Neural Tangent Kernel for linear-width neural networks
(Submitted on 25 May 2020 (v1), last revised 10 Oct 2020 (this version, v3))
Abstract: We study the eigenvalue distributions of the Conjugate Kernel and Neural Tangent Kernel associated to multi-layer feedforward neural networks. In an asymptotic regime where network width is increasing linearly in sample size, under random initialization of the weights, and for input samples satisfying a notion of approximate pairwise orthogonality, we show that the eigenvalue distributions of the CK and NTK converge to deterministic limits. The limit for the CK is described by iterating the Marcenko-Pastur map across the hidden layers. The limit for the NTK is equivalent to that of a linear combination of the CK matrices across layers, and may be described by recursive fixed-point equations that extend this Marcenko-Pastur map. We demonstrate the agreement of these asymptotic predictions with the observed spectra for both synthetic and CIFAR-10 training data, and we perform a small simulation to investigate the evolutions of these spectra over training.
Submission history
From: Zhou Fan [view email][v1] Mon, 25 May 2020 01:11:49 GMT (634kb,D)
[v2] Tue, 2 Jun 2020 17:46:17 GMT (780kb,D)
[v3] Sat, 10 Oct 2020 16:47:46 GMT (1074kb,D)
Link back to: arXiv, form interface, contact.