RL-trained adversarial attacks cut a smart-inverter FDIA detector's accuracy from 71.6% to 3.6%, and sequential continual retraining forgets old attacks unless a rehearsal strategy is used.
A survey on the detection algo- rithms for false data injection attacks in smart grids,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Continual Adversarial Reinforcement Learning (CARL) of False Data Injection detection: forgetting and explainability
RL-trained adversarial attacks cut a smart-inverter FDIA detector's accuracy from 71.6% to 3.6%, and sequential continual retraining forgets old attacks unless a rehearsal strategy is used.