Abstract:Brain-inspired neuromorphic computing has emerged as a promising solution to overcome the energy and speed limitations of conventional von Neumann architectures. In this context, in-memory computing utilizing memristors has gained attention as a key technology, harnessing their non-volatile characteristics to replicate synaptic behavior akin to the human brain. However, challenges arise from non-linearities, asymmetries, and device variations in memristive devices during synaptic weight updates, leading to inaccurate weight adjustments and diminished recognition accuracy. Moreover, the repetitive weight updates pose endurance challenges for these devices, adversely affecting latency and energy consumption. To address these issues, we propose a Siamese network learning approach to optimize the training of multi-level memristor neural networks. During neural inference, forward propagation takes place within the memristor neural network, enabling error and noise detection in the memristive devices and hardware circuits. Simultaneously, high-precision gradient computation occurs on the software side, initially updating the floating-point weights within the Siamese network with gradients. Subsequently, weight quantization is performed, and the memristor conductance values requiring updates are modified using a sparse update strategy. Additionally, we introduce gradient accumulation and weight quantization error compensation to further enhance network performance. The experimental results of MNIST data recognition, whether based on a MLP or a CNN model, demonstrate the rapid convergence of our network model. Moreover, our method successfully eliminates over 98% of weight updates for memristor conductance weights within a single epoch. This substantial reduction in weight updates leads to a significant decrease in energy consumption and time delay by more than 98% when compared to the basic closed-loop update method. Consequently, this approach effectively addresses the durability requirements of memristive devices.

Enhancing Security of Memristor Computing System Through Secure Weight Mapping

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

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

Bayesian Inference Based Robust Computing on Memristor Crossbar

Aging Aware Retraining for Memristor-based Neuromorphic Computing

Security Enhancement for RRAM Computing System through Obfuscating Crossbar Row Connections

Review of security techniques for memristor computing systems

TDPP: 2-D Permutation-Based Protection of Memristive Deep Neural Networks

Study of Weight Quantization Associations over a Weight Range for Application in Memristor Devices

Mitigating Slow-to-Write Errors in Memristor-Mapped Graph Neural Networks Induced by Adversarial Attacks

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

WESCO: Weight-encoded Reliability and Security Co-design for In-memory Computing Systems

QuantBayes: Weight Optimization for Memristive Neural Networks via Quantization-Aware Bayesian Inference

Power-balanced Memristive Cryptographic Implementation Against Side Channel Attacks

An Ultra-Efficient Memristor-Based DNN Framework with Structured Weight Pruning and Quantization Using ADMM

Novel Weight Mapping Method for Reliable NVM based Neural Network

Security enhancement of artificial neural network using physically transient form of heterogeneous memristors with tunable resistive switching behaviors

Memristor‐Based Analog Computation and Neural Network Classification with a Dot Product Engine

Enhancing in-situ updates of quantized memristor neural networks: a Siamese network learning approach

EFFICIENT WEIGHT MAPPING SCHEME WITHOUT VERIFICATION FOR RRAM BASED NEUROMORPHIC COMPUTING

Rescuing Memristor-Based Computing with Non-Linear Resistance Levels.