An Effective and Efficient Initialization Scheme for Training Multi-layer Feedforward Neural Networks
Abstract
Network initialization is the first and critical step for training neural networks. In this paper, we propose a novel network initialization scheme based on the celebrated Stein's identity. By viewing multi-layer feedforward neural networks as cascades of multi-index models, the projection weights to the first hidden layer are initialized using eigenvectors of the cross-moment matrix between the input's second-order score function and the response. The input data is then forward propagated to the next layer and such a procedure can be repeated until all the hidden layers are initialized. Finally, the weights for the output layer are initialized by generalized linear modeling. Such a proposed SteinGLM method is shown through extensive numerical results to be much faster and more accurate than other popular methods commonly used for training neural networks.
Cite
@article{arxiv.2005.08027,
title = {An Effective and Efficient Initialization Scheme for Training Multi-layer Feedforward Neural Networks},
author = {Zebin Yang and Hengtao Zhang and Agus Sudjianto and Aijun Zhang},
journal= {arXiv preprint arXiv:2005.08027},
year = {2020}
}