39fJ Analog Artificial Neural Network for Breast Cancer Classification in 65nm CMOS

Abstract

An analog artificial neural network (ANN) classifier using a common-source amplifier based nonlinear activation function is presented in this work. A shallow ANN is designed in 65nm CMOS to perform binary classification on breast cancer dataset and identify each patient data as either benign or malignant. Use of common-source amplifier structure simplifies the ANN and results in only 39fJ/classification at 0.8V power supply and core area of only 240μm2. The classifier is trained using Matlab and validated using Spectre simulations.