The paper claims the first gradient inversion attack on adapter-based PEFT in federated learning, reconstructing private images from adapter gradients via a poisoned pretrained model.
Tinytl: Reduce memory, not parameters for efficient on-device learning
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
eess.IV 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Gradient Inversion Attacks on Parameter-Efficient Fine-Tuning
The paper claims the first gradient inversion attack on adapter-based PEFT in federated learning, reconstructing private images from adapter gradients via a poisoned pretrained model.