Fault Injection Attack

Fault injection attacks target the parameters of neural networks, aiming to manipulate their behavior for malicious purposes, such as misclassification or backdoor insertion. Current research focuses on developing both novel attacks, particularly bit-flip attacks tailored to specific architectures like graph neural networks and spiking neural networks, and robust defense mechanisms, including encoding-based protection and contrastive learning for detection and recovery. These efforts are crucial for securing the growing deployment of neural networks in safety-critical applications and ensuring the reliability of machine learning systems.