2 papers
cs.CV2026
High-Capacity Robust Medical Image Exfiltration via Neural Network Weight Replacement
Elie Thellier, Huiyu Li, Nicholas Ayache +1
Collaborative medical AI platforms allow researchers to train models on sensitive imaging data while restricting data export. However, trained models can serve as covert carriers o…
cs.LG2025
Mitigating Data Exfiltration Attacks through Layer-Wise Learning Rate Decay Fine-Tuning
Elie Thellier, Huiyu Li, Nicholas Ayache +1
Data lakes enable the training of powerful machine learning models on sensitive, high-value medical datasets, but also introduce serious privacy risks due to potential leakage of p…