Classifying Analog and Digital Circuits with Machine Learning Techniques toward Mixed-Signal Design Automation

For modern system-on-chip (SoC) design, one of the most challenging and time-consuming tasks is the layout design of the mixed-signal integrated circuit (IC), which integrates both analog and digital circuits into a single chip. There is no industrial tool which can automatically identify analog and digital sub-circuits in a mixed-signal design to accelerate the layout design automation. In this paper, we first introduce a device sorting method to generate an unique sequence for circuit components. Then, we apply an unique matrix representation to encode circuit netlists. Finally, we employ machine learning algorithms to automatically classify/identify analog and digital sub-circuits. The experimental results show that the proposed method is promising based on the convolutional neural network (CNN) algorithm.