*Memristor - An Emerging Device for Post-Moore's Computing and Applications*

#### **Figure 4.**

*(a) AIDX optimizes and applies separate voltage amplitude and duration ratios for each layer separately. (b) Basic test on memristance drift with all devices with and without AIDX. With all memristors initially set to 0.0052 S, half of the memristors receive pre-generated random sequences of positive and an identical sequence of negative pulses are applied to the other half of the crossbar.*

operates with a voltage range of 0.2 V to 0.2 V. Depending on the pulse amplitude optimization, AIDX's inference pulse can vary within this range. On average, there is 2% reduction in passive power consumption within the crossbar compared to the baseline system when using AIDX. A simple test of AIDX effectiveness is done by applying random positive and negative voltage pulses to memristors with and without AIDX. After 10000 inference operations, the baseline memristors deviated a max of 1.9% from its initial value while AIDX had a max deviation of only 0.17% (**Figure 4b**). When tested on the 10 benchmark tasks from the Proben1 dataset [63], the average classification accuracy degradation with AIDX is approximately 4%, 7%,

#### **Figure 5.**

*(a) Classification accuracy of AIDX and baseline neural network on ten tasks from the Proben1 dataset after 500, 2000, and 10000 inference operations. (b) Classification accuracy of AIDX and baseline on CIFAR10 dataset with various CNN architectures after 500, 2000, and 10000 inference operations. Figure reprinted by [56].*

*Mitigating State-Drift in Memristor Crossbar Arrays for Vector Matrix Multiplication DOI: http://dx.doi.org/10.5772/intechopen.100246*

#### **Figure 6.**

*(a) Reconstruction of sample images from MNIST dataset after 1, 500, 2000, and 10000 inference operations. (b) The average mean squared error in image reconstruction between AIDX and baseline autoencoder after set time steps. The percentage error improvement of AIDX over the baseline is also shown. Figure reprinted by [56].*

and 8% after 500, 2000, and 10000 inference operations **(Figure 5a)**. On average, AIDX reduced accuracy degradation by 42% as compared to the baseline test after 10000 inference operations. When testing AIDX on 10 CNN architectures using the CIFAR10 dataset [18], the classification accuracy decrease by an average of 4%, 7%, and 8% after 500, 2000, and 10000 inference operations **(Figure 5b)**. This accuracy degradation corresponds to a 22%, 35%, and 43% improvement over the baseline respectively. In addition, AIDX was also applied to image reconstruction by training a simple 3-layer autoencoder on the MNIST dataset [62]. The average mean squared error of the baseline auto-encoder was 0.033, 0.068, and 0.129 after 500, 2000, and 10000 inference operations respectively. With AIDX, the average mean squared error drops to 0.015, 0.021, and 0.028 after 500, 2000, and 10000 inference operations which is an improvement of 53.0%, 69.0%, and 78.6% over the baseline (**Figure 6**).
