Picture for Hualong Ma

Hualong Ma

MEA-Defender: A Robust Watermark against Model Extraction Attack

Add code
Jan 26, 2024
Viaarxiv icon

DataElixir: Purifying Poisoned Dataset to Mitigate Backdoor Attacks via Diffusion Models

Add code
Dec 20, 2023
Viaarxiv icon

SSL-WM: A Black-Box Watermarking Approach for Encoders Pre-trained by Self-supervised Learning

Add code
Sep 08, 2022
Figure 1 for SSL-WM: A Black-Box Watermarking Approach for Encoders Pre-trained by Self-supervised Learning
Figure 2 for SSL-WM: A Black-Box Watermarking Approach for Encoders Pre-trained by Self-supervised Learning
Figure 3 for SSL-WM: A Black-Box Watermarking Approach for Encoders Pre-trained by Self-supervised Learning
Figure 4 for SSL-WM: A Black-Box Watermarking Approach for Encoders Pre-trained by Self-supervised Learning
Viaarxiv icon

DBIA: Data-free Backdoor Injection Attack against Transformer Networks

Add code
Nov 22, 2021
Figure 1 for DBIA: Data-free Backdoor Injection Attack against Transformer Networks
Figure 2 for DBIA: Data-free Backdoor Injection Attack against Transformer Networks
Figure 3 for DBIA: Data-free Backdoor Injection Attack against Transformer Networks
Figure 4 for DBIA: Data-free Backdoor Injection Attack against Transformer Networks
Viaarxiv icon