Deep linear discriminant analysis

Autoren Matthias Dorfer
Rainer Kelz
Gerhard Widmer
Editoren
Titel Deep linear discriminant analysis
Buchtitel Published as a conference paper at ICLR 2016 - 4th International Conference on Learning Representations, San Juan, Puerto Rico, May 2-4, 2016.
Typ in Konferenzband
Monat February
Jahr 2016
Seiten https://arxiv.org/abs/1511.04707
SCCH ID# 16095
Abstract

We introduce Deep Linear Discriminant Analysis (DeepLDA) which learns linearly separable latent representations in an end-to-end fashion. Classic LDA extracts features which preserve class separability and is used for dimensionality reduction for many classification problems. The central idea of this paper is to put LDA on top of a deep neural network. This can be seen as a non-linear extension of classic LDA. Instead of maximizing the likelihood of target labels for individual samples, we propose an objective function that pushes the network to produce feature distributions which: (a) have low variance within the same class and (b) high variance between different classes. Our objective is derived from the general LDA eigenvalue problem and still allows to train with stochastic gradient descent and back-propagation. For evaluation we test our approach on three different benchmark datasets (MNIST, CIFAR-10 and STL-10). DeepLDA produces competitive results on MNIST and CIFAR-10 and outperforms a network trained with categorical cross entropy (having the same architecture) on a supervised setting of STL-10.