Search papers, labs, and topics across Lattice.
The paper introduces REFORGE, a black-box red-teaming framework that uses adversarial image prompts to evaluate the robustness of Image Generation Model Unlearning (IGMU) techniques. REFORGE leverages a cross-attention-guided masking strategy to optimize perturbations in stroke-based images, focusing noise on concept-relevant regions. Experiments across various unlearning tasks and defenses show that REFORGE significantly improves attack success rates and semantic alignment compared to existing baselines, revealing vulnerabilities in current IGMU methods.
Current image generation unlearning methods are surprisingly brittle: adversarial image prompts, optimized with attention-guided masking, can effectively resurrect supposedly "forgotten" concepts.
Recent progress in image generation models (IGMs) enables high-fidelity content creation but also amplifies risks, including the reproduction of copyrighted content and the generation of offensive content. Image Generation Model Unlearning (IGMU) mitigates these risks by removing harmful concepts without full retraining. Despite growing attention, the robustness under adversarial inputs, particularly image-side threats in black-box settings, remains underexplored. To bridge this gap, we present REFORGE, a black-box red-teaming framework that evaluates IGMU robustness via adversarial image prompts. REFORGE initializes stroke-based images and optimizes perturbations with a cross-attention-guided masking strategy that allocates noise to concept-relevant regions, balancing attack efficacy and visual fidelity. Extensive experiments across representative unlearning tasks and defenses demonstrate that REFORGE significantly improves attack success rate while achieving stronger semantic alignment and higher efficiency than involved baselines. These results expose persistent vulnerabilities in current IGMU methods and highlight the need for robustness-aware unlearning against multi-modal adversarial attacks. Our code is at: https://github.com/Imfatnoily/REFORGE.
NVIDIA