Parallel Density‐Based Spatial Clustering with Dual‐Functional Memristive Crossbar Array

Abstract

Analog and digital switching performances in a Ta/HfO2/RuO2 (THR)memristor are studied to implement a density-based spatial clustering ofapplications with noise (DBSCAN) algorithm in a low-power,parallel-computing memristor crossbar structure. In the analog mode THRmemristor, more than 256 states can be stored through a fine-tuning processwith a denoising scheme. The analog mode crossbar array facilitatesEuclidean distance calculation between any points in the given graphicdataset. In the digital mode, the on/off ratio of more than three orders ofmagnitude between the binary states is achieved, providing functionality tocluster the data points with a reduced number of operations. The parallelcomputing capacity of the adopted crossbar decreases the time complexity ofthe original DBSCAN from O(n2) to O(n). Through array-level simulations, theeffectiveness of hardware functionality is validated using representativesynthetic datasets and single-cell RNA sequences datasets.