A Low-energy Machine-learning Classifier Based on Clocked Comparators for Direct Inference on Analog Sensors
This paper presents a system where clocked comparators consuming only CV2 energy directly derive classification decisions from analog sensor signals, thereby replacing instrumentation amplifiers, ADCs, and digital MACs, as typically required. A machine-learning algorithm for training the classifier is presented, which enables circuit non-idealities as well as severe energy/area scaling in analog circuits to be overcome. Further, a noise model of the system is presented and experimentally verified, providing a means to predict and optimize classification error probability in a given application. The noise model shows that superior noise efficiency is achieved by the comparator-based system compared to a system based on linear low-noise amplifiers. A prototype in 130nm CMOS performs image recognition of handwritten numerical digits, by taking raw analog pixels as the inputs. Due to pin limitations on the chip, the images with 28X28=784 pixels are resized and down-sampled to give 47 pixel features, yielding an accuracy of 90% for an ideal 10-way classification system (MATLAB simulated). The prototype comparator-based system achieves equivalent performance with a total energy of 543pJ per 10-way classification at a rate up to 1.3M images per second, representing 33 lower energy than an ADC/digital-MAC system.