SAST applies sharpness-aware minimization to surrogate-gradient SNN training and reports large gains in hard-spike accuracy on N-MNIST and DVS Gesture, including under INT8/INT4 quantization.
Fast-classifying, high-accuracy spiking deep networks through weight and threshold balancing,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.NE 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Sharpness-Aware Surrogate Training for On-Sensor Spiking Neural Networks
SAST applies sharpness-aware minimization to surrogate-gradient SNN training and reports large gains in hard-spike accuracy on N-MNIST and DVS Gesture, including under INT8/INT4 quantization.