A modality-aware two-stage trust scoring and inspection scheme defends OTA-FL against bounded-scaling, Euclidean-constrained, Cosine-constrained, and Neurotoxin backdoor attacks while preserving main-task accuracy.
Attack of the tails: Yes, you really can backdoor federated learning.Advances in neural information processing systems, 33:16070–16084, 2020
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 and Mitigating Backdoor Attacks in OTA-FL Systems: A Two-Stage Robust Aggregation Scheme
A modality-aware two-stage trust scoring and inspection scheme defends OTA-FL against bounded-scaling, Euclidean-constrained, Cosine-constrained, and Neurotoxin backdoor attacks while preserving main-task accuracy.