Fault Injection Attacks in Spiking Neural Networks and Countermeasures

Keyphrases

• Spiking Neural Networks 100%
• Classification Accuracy 42%
• Accuracy Degradation 42%
• Learning Rate 28%
• Threshold Voltage 28%
• Energy Overhead 28%
• Decay Constant 28%
• Vulnerability 14%
• Area Overhead 14%
• Laser-induced 14%
• Design Decisions 14%
• Number of Neurons 14%
• Adversary 14%
• Voltage Change 14%
• High Energy Efficiency 14%
• Glitches 14%
• Power Lasers 14%
• Spiking Neurons 14%
• Design Parameters 14%
• Corrupt 14%
• Classification Task 14%
• Local Power 14%
• Neural Network Training 14%
• Neuronal Membrane 14%
• Threshold Potential 14%
• Spike Amplitude 14%
• Deep Neural Network 14%
• Robust Implementation 14%
• Voltage Fault 14%
• Driver Design 14%
• Hardware Level 14%
• Digit Recognition 14%
• Fault Injection 14%
• External Power Supply 14%
• Power-oriented 14%
• Current Driver 14%

Computer Science

• Neural Network 100%
• Fault Injection 100%
• Classification Accuracy 50%
• Threshold Voltage 33%
• Deep Neural Network 33%
• Learning Rate 33%
• Energy Efficiency 16%
• Design Parameter 16%
• Classification Task 16%
• Neural Network Training 16%
• Hardware Level 16%