WISER: Deep Neural Network Weight-bit Inversion for State Error Reduction in MLC NAND Flash
Jaehun Jang1 and Jong Hwan Ko2
1Department of Semiconductor and Display Engineering Sungkyunkwan University, Suwon, Korea Memory Division, Samsung Electronics, Hwaseong, Korea
hun0317@g.skku.edu
2Department of Electronic and Electrical Engineering Sungkyunkwan University Suwon, Korea
jhko@skku.edu
ABSTRACT
When Flash memory is used to store the deep neural network (DNN) weights, inference accuracy can degrade due to the Flash memory state errors. To protect the weights from the state errors, the existing methods relied on ECC(Error Correction Code) or parity, which can incur power/storage overhead. In this study, we propose a weight bit inversion method that minimizes the accuracy loss due to the Flash memory state errors without using the ECC or parity. The method first applies WISE(Weightbit Inversion for State Elimination) that removes the most errorprone state from MLC NAND, thereby improving both the error robustness and the MSB page read speed. If the initial accuracy loss due to weight inversion of WISE is unacceptable, we apply WISER(Weight-bit Inversion for State Error Reduction) that reduces weight mapping to the error-prone state with minimum weight value changes. The simulation results show that after 16K program-erase cycles in NAND Flash, WISER reduces CIFAR- 100 accuracy loss by 2.92X for VGG-16 compared to the existing methods.
Keywords: NAND Flash Memory, Deep Neural Network, Image Classification, Reliability.
