Co-design of DNN Model Optimization for Binary ReRAM Array In-memory Processing

In this work, a novel design of ReRAM neuromorphic system is proposed to process deep neural network ( DNN) fully in array efficiently. A binary neural network model is constructed and optimized on MNIST dataset. The obtained model is simulated to be processed with the proposed ReRAM array. Co-design between hardware and network model optimization in software is analyzed to achieve feasible hardware design and generalizable model. Deployed with such co-design model, ReRAM array processes DNN with high robustness against fabrication fluctuation.