deep learning 1model inversion 1privacy leakage 1supply chain security 1training-time injection attack 1
From the 1 of 2 linked papers with an AI index.
2 papers
cs.CR2026
Don't Trust the AI Ecosystem: Analyzing Privacy Leakage in Compromised Open-Source Components
Jin-Seong Kim, Han-Ju Lee, Seok-Won Hong +4
The paper presents GradLock, a training-time injection attack that stealthily embeds private training data into model parameters, enabling near‑perfect reconstruction of the data f…
cs.CV2026
DPAC: Distribution-Preserving Adversarial Control for Diffusion Sampling
Han-Jin Lee, Han-Ju Lee, Jin-Seong Kim +1
Adversarially guided diffusion sampling often achieves the target class, but sample quality degrades as deviations between the adversarially controlled and nominal trajectories acc…