Publications

Is Forgetting Unified? Exploring Concept Unlearning in Unified Multimodal Models

Zhongkai Xue, Junkai Chen, Mingcong Lei, Rumo Zhang, Yixiao Liu, Zhimin Sun, Shu Wu, Qiang Liu, Lihua Zhang, Liang Wang

Unified multimodal models can generate images and understand visual content, but forgetting a concept in one pathway does not necessarily remove it from the other. We introduce UUBench to evaluate cross-pathway concept unlearning and propose UniForget to jointly suppress target concepts while preserving other model capabilities.