计算机应用研究2024，Vol.41Issue(10)：3068-3072,5.DOI:10.19734/j.issn.1001-3695.2024.01.0046
基于分区再训练的RRAM阵列多缺陷容忍算法
Partition retraining based multi-fault tolerance algorithm for RRAM crossbar
摘要
Abstract
To address the issue of calculation errors in neural network matrix-vector multiplication caused by manufacturing processes of RRAM cells,this paper modeled the characteristics of multiple faults in RRAM crossbar arrays and proposed a multi-fault tolerant algorithm.Firstly,it modeled the impacts of common transition fault and stuck at fault in RRAM crossbar arrays on the accuracy of neural network computations.Secondly,it partitioned the neural network and conducted partitioned training based on an improved knowledge distillation method.Lastly,it further optimized the algorithm by selecting an appro-priate loss function and incorporating normalization layers.Experimental results on the MNIST and Cifar-10 datasets demon-strate that the proposed method can achieve a recovery rate of over 98％across multiple neural networks,indicating its effec-tiveness in mitigating the impact of multiple faults in RRAM crossbar arrays on the accuracy of neural network computations.关键词
RRAM阵列/缺陷容忍/神经网络/知识蒸馏Key words
RRAM crossbar/fault tolerance/neural network/knowledge distillation分类
信息技术与安全科学引用本文复制引用
王梦可,杨朝晖,查晓婧,夏银水..基于分区再训练的RRAM阵列多缺陷容忍算法[J].计算机应用研究,2024,41(10):3068-3072,5.基金项目
国家自然科学基金资助项目(62131010,U22A2013) （62131010,U22A2013）
国家自然科学基金青年项目(62304115) （62304115）
浙江省自然科学基金创新群体资助项目(LDT23F04021F04) （LDT23F04021F04）
浙江省科研计划一般项目(Y202248965) （Y202248965）
