Deep Neural Networks with Multistate Activation Functions. (10th September 2015)

• Record Type: Journal Article
• Authors: Cai, Chenghao; Xu, Yanyan; Ke, Dengfeng; Su, Kaile
• Other Names: Lin Cheng-Jian Academic Editor.
• Is Part Of: Computational intelligence and neuroscience. Volume 2015(2015)
• Publication Date: 2015-09-10
• Subjects: Neurosciences -- Data processing -- Periodicals; Computational intelligence -- Periodicals; Computational neuroscience -- Periodicals; 612.80285
• DOI: 10.1155/2015/721367
• Languages: English
• ISSNs: 1687-5265

Abstract

Abstract : We propose multistate activation functions (MSAFs) for deep neural networks (DNNs). These MSAFs are new kinds of activation functions which are capable of representing more than two states, including the N -order MSAFs and the symmetrical MSAF. DNNs with these MSAFs can be trained via conventional Stochastic Gradient Descent (SGD) as well as mean-normalised SGD. We also discuss how these MSAFs perform when used to resolve classification problems. Experimental results on the TIMIT corpus reveal that, on speech recognition tasks, DNNs with MSAFs perform better than the conventional DNNs, getting a relative improvement of 5.60% on phoneme error rates. Further experiments also reveal that mean-normalised SGD facilitates the training processes of DNNs with MSAFs, especially when being with large training sets. The models can also be directly trained without pretraining when the training set is sufficiently large, which results in a considerable relative improvement of 5.82% on word error rates.