A PAC-Bayesian Approach to Spectrally-Normalized Margin Bounds for Neural Networks
Behnam Neyshabur(Toyota Technological Institute at Chicago), Srinadh Bhojanapalli(Toyota Technological Institute at Chicago), Nathan Srebro(Toyota Technological Institute)
International Conference on Learning Representations
July 29, 2017
Cited by 256
Abstract
We present a generalization bound for feedforward neural networks in terms of the product of the spectral norm of the layers and the Frobenius norm of the weights. The generalization bound is derived using a PAC-Bayes analysis.
Related Papers
No related papers found
Powered by citation graph analysis