Lie Group Impression for Deep Learning

Mengduo Yang,Fanzhang Li

DOI: https://doi.org/10.1016/j.ipl.2018.03.006

IF: 0.851

2018-01-01

Information Processing Letters

Abstract:In this work, we exploit a novel algorithm for capturing the Lie group manifold structure of the visual impression. By developing the single-layer Lie group model, we show how the representation learning algorithm can be stacked to yield a deep architecture. In addition, we design a Lie group based gradient descent algorithm to solve the learning problem of network weights. We show that our proposed technique yields representations that significantly better suited for training deep network and is also computationally efficient.