Self-adaptive quasi-Gaussian circuits for analog on-chip-trainable multi-class classifiers

Abstract

Self-adaptive quasi-Gaussian circuits have been developed and introduced to an analog multi-class classifier in order to enhance its classification performance. By applying a floating threshold scheme to the quasi-Gaussian kernel, the kernel can extend its tail region adaptively according to the characteristics of input data. As a result, the… (More)
DOI: 10.1109/ISCAS.2012.6271919

Topics

7 Figures and Tables