MIST detects Trojaned DNN updates by measuring spectral deviations in pre-activation representations against a benign fine-tuning reference, achieving high accuracy across datasets and attacks after a single update.
Abs: Scanning neural networks for back-doors by artificial brain stimulation,
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
-
Detecting Trojaned DNNs via Spectral Regression Analysis
MIST detects Trojaned DNN updates by measuring spectral deviations in pre-activation representations against a benign fine-tuning reference, achieving high accuracy across datasets and attacks after a single update.