SNN Backdoor Benchmark

Backdoor attacks and defenses on spiking neural networks with neuromorphic data. Every number is a cell of the same harness, with its seeds, site and device.

Attacks
Defenses

Rows are attacks and columns are defenses: the baselines first, then the oldest papers first. Each cell is the mean over seeds, with the standard deviation and the number of seeds. A grey cell is not applicable, invalid or failed: hover for the reason.