IncreFA: Breaking the Static Wall of Generative Model Attribution
IncreFA: 破解生成模型归因的静态壁垒
机构 * School of Artificial Intelligence, Beijing University of Posts and Telecommunications(北京邮电大学人工智能学院) ; Tsinghua University(清华大学)
专题命中 图像生成评测 :diffusion(abstract);分类 cs.CV
AI总结 IncreFA将归因重新定义为结构化增量学习问题,通过层级约束和潜在记忆库机制,在28个生成模型上实现最先进的归因准确率和未见检测率。
Comments Accepted to CVPR 2026, with appendix