Using pretrained models as the starting point for a surrogate improves model extraction accuracy and fidelity against federated learning victim models, especially at small query budgets.
On safeguarding privacy and security in the framework of federated learning,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Evaluating Query Efficiency and Accuracy of Transfer Learning-based Model Extraction Attack in Federated Learning
Using pretrained models as the starting point for a surrogate improves model extraction accuracy and fidelity against federated learning victim models, especially at small query budgets.