Robust noisy speech recognition using deep neural support vector machines

Abstract

This paper aims to classify noisy sound samples in several daily indoor and outdoor acoustic scenes using an optimized deep neural networks (DNNs). The advantage of a traditional DNNs lies in using at the top layer a softmax activation function which is a logistic regression in order to learn the output label in a multi-class recognition problem. In this paper, we optimize the DNNs by replacing the softmax activation function by a linear support vector machine. In this paper, a novel deep neural networks (DN) using Support Vector Machines (SVM) instead of the multinomial logistic regression is proposed. We have verified the effectiveness of this new method using speech samples from Aurora speech database recorded in noisy conditions. The experimental results obtained with the method DN-SVM demonstrates a significant improvement of the performance with noisy sound samples classification.