Abstract:In recent years, memristive crossbar-based neuromorphic computing systems (NCS) have provided a promising solution to the acceleration of neural networks. However, stuck-at faults (SAFs) in the memristor devices significantly degrade the computing accuracy of NCS. Besides, the memristor suffers from the process variations, causing deviation of the actual programming resistance from its target resistance. In this paper, we propose a reliability-driven design framework for a memristive crossbar-based NCS in combination with general and chip-specific design optimizations. First, we design a general reliability-aware training scheme to enhance the robustness of NCS to SAFs and device variations; a dropconnect-inspired approach is developed to alleviate the impact of SAFs; a new weighted error function, including cross-entropy error (CEE), the $l_{2}$ -norm of weights, and the sum of squares of first-order derivatives of CEE with respect to weights, is proposed to obtain a smooth error curve, where the effects of variations are suppressed. Second, given the neural network model generated by the reliability-aware training scheme, we exploit chip-specific mapping and re- training to further improve computation accuracy loss incurred by SAFs. Experimental results show that the proposed method can boost the computation accuracy of NCS and improve the NCS robustness. Note to Practitioners—This work is motivated by the manufacturing reliability problem in a memristive crossbar-based NCS. To enhance the robustness of an NCS to SAFs and device variations, this paper presents a reliability-driven design framework with taking account of both general and chip-specific design optimizations. The experimental results have demonstrated that the proposed framework is superior to the prior arts, and can be easily integrated with existing industrial hardware-based fault tolerance solutions for higher accuracy at lower overhead. Memristive crossbar-based computing system gives hope for the anticipated efficient implementation of artificial neuromorphic networks. With the help of the reliability-driven designs, the computation accuracy is restored, and hence we can expect the wide use of memristive crossbar-based computing system in neuromorphic computing applications.

Reliable Memristor-based Neuromorphic Design Using Variation- and Defect-Aware Training

Aging Aware Retraining for Memristor-based Neuromorphic Computing

Reliability-Driven Neuromorphic Computing Systems Design.

Reliability-Driven Neural Network Training for Memristive Crossbar-Based Neuromorphic Computing Systems

Bayesian Inference Based Robust Computing on Memristor Crossbar

Reliability-Driven Memristive Crossbar Design in Neuromorphic Computing Systems

Robustness of Neuromorphic Computing with RRAM-based Crossbars and Optical Neural Networks.

BRoCoM: A Bayesian Framework for Robust Computing on Memristor Crossbar.

A Design Methodology for Fault-Tolerant Neuromorphic Computing Using Bayesian Neural Network

Resist: Robust Network Training for Memristive Crossbar-Based Neuromorphic Computing Systems.

Build Reliable And Efficient Neuromorphic Design With Memristor Technology

A Closed-Loop Design to Enhance Weight Stability of Memristor Based Neural Network Chips

Reduction and IR-drop Compensations Techniques for Reliable Neuromorphic Computing Systems

Robust Learning Approach for Neuro-Inspired Nanoscale Crossbar Architecture.

Nonideality-Aware Training for Accurate and Robust Low-Power Memristive Neural Networks

AD2VNCS: Adversarial Defense and Device Variation-tolerance in Memristive Crossbar-based Neuromorphic Computing Systems

ReMeCo: Reliable Memristor-Based In-Memory Neuromorphic Computation

Improving Dependability of Neuromorphic Computing With Non-Volatile Memory

A Variation Tolerant Scheme for Memristor Crossbar Based Neural Network Designs Via Two-Phase Weight Mapping and Memristor Programming

AD 2 VNCS: A Dversarial D Efense and D Evice V Ariation-Tolerance in Memristive Crossbar-based N Euromorphic C Omputing S Ystems

Reliability of Memristive Devices for High-Performance Neuromorphic Computing: (Invited Paper).