AdvScan detects adversarial examples in black-box TinyML on ARM Cortex-M devices via one-sample t-test on runtime power signatures against a benign baseline, reporting 99.984% detection with 40 false negatives and zero false positives over 318400 inputs.
Gran: An efficient gradient-norm based detector for adversarial and misclassified examples,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
AdvScan: Black-Box Adversarial Example Detection at Runtime through Power Analysis
AdvScan detects adversarial examples in black-box TinyML on ARM Cortex-M devices via one-sample t-test on runtime power signatures against a benign baseline, reporting 99.984% detection with 40 false negatives and zero false positives over 318400 inputs.