Boheng Li

6

Papers

7

Total Citations

Papers (6)

Towards Resilient Safety-driven Unlearning for Diffusion Models against Downstream Fine-tuning

Transfer Learning of Real Image Features with Soft Contrastive Loss for Fake Image Detection

Invisible Backdoor Attack against Self-supervised Learning

TraceEvader: Making DeepFakes More Untraceable via Evading the Forgery Model Attribution

Nearest is Not Dearest: Towards Practical Defense against Quantization-conditioned Backdoor Attacks

Purifying Quantization-conditioned Backdoors via Layer-wise Activation Correction with Distribution Approximation