CoG-Guided Weight Correction for Fault-Tolerant Deep Neural Networks

2026-07-17

AI Summary

This work addresses the vulnerability of deep neural networks to hardware or memory faults in safety-critical applications, which can corrupt weights and degrade reliability. The authors introduce, for the first time, the concept of Center of Gravity (CoG) into weight tensors and propose a fault-tolerant method that requires neither retraining nor architectural modifications. By analyzing the spatial distribution of weights in each layer, the approach employs a distance-aware rule to detect and correct faulty weights. Experimental results demonstrate that, at a bit error rate (BER) of 10⁻³, the method improves fault tolerance by 230× and 6.41× for StageNet and MTFNet, respectively, and by 49.55× and 20.79× for ResNet-18 and VGG-16, with negligible accuracy loss.

Abstract

Deep Neural Networks (DNNs) used in safety-critical applications are vulnerable to hardware and memory faults that corrupt network weights and degrade reliability. In this paper, we propose a Center of Gravity (CoG) guided weight correction method that restores faulty weights based on their spatial characteristics within each layer. The proposed approach detects and corrects weight faults using distance-aware correction rules, eliminating the need for retraining or architectural modification. The effectiveness of the proposed method in terms of the capability of tolerating hardware faults has been evaluated through performing fault injection at different Bit Error Rates (BERs). Experiments on safety-critical LSTM-based Networks, including StageNet for disease progression tracking and MTFNet for cardiac anomaly detection, demonstrate fault tolerance improvements of up to 230x and 6.41x, respectively, at a BER of 10^{-3}, with negligible accuracy loss. When extended to Convolutional Neural Networks (CNNs), the method achieves up to 49.55x and 20.79x improvements under comparable fault conditions on ResNet-18 and VGG-16, respectively. To the best of our knowledge, this is the first work to apply the CoG concept to neural network weight tensors for enhancing model reliability.

Problem

• fault-tolerant
• weight corruption
• hardware faults
• reliability
• safety-critical

Innovation

• Center of Gravity (CoG)
• weight correction
• fault-tolerant DNNs
• bit error rate (BER)
• retraining-free