Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

EraseAnything++: Enabling Concept Erasure in Rectified Flow Transformers Leveraging Multi-Object Optimization

About

Removing undesired concepts from large-scale text-to-image (T2I) and text-to-video (T2V) diffusion models while preserving overall generative quality remains a major challenge, particularly as modern models such as Stable Diffusion v3, Flux, and OpenSora employ flow-matching and transformer-based architectures and extend to long-horizon video generation. Existing concept erasure methods, designed for earlier T2I/T2V models, often fail to generalize to these paradigms. To address this issue, we propose EraseAnything++, a unified framework for concept erasure in both image and video diffusion models with flow-matching objectives. Central to our approach is formulating concept erasure as a constrained multi-objective optimization problem that explicitly balances concept removal with preservation of generative utility. To solve the resulting conflicting objectives, we introduce an efficient utility-preserving unlearning strategy based on implicit gradient surgery. Furthermore, by integrating LoRA-based parameter tuning with attention-level regularization, our method anchors erasure on key visual representations and propagates it consistently across spatial and temporal dimensions. In the video setting, we further enhance consistency through an anchor-and-propagate mechanism that initializes erasure on reference frames and enforces it throughout subsequent transformer layers, thereby mitigating temporal drift. Extensive experiments on both image and video benchmarks demonstrate that EraseAnything++ substantially outperforms prior methods in erasure effectiveness, generative fidelity, and temporal consistency, establishing a new state of the art for concept erasure in next-generation diffusion models.

Zhaoxin Fan, Nanxiang Jiang, Daiheng Gao, Shiji Zhou, Wenjun Wu• 2026

Related benchmarks

TaskDatasetResultRank
Text-to-Image GenerationMS-COCO
FID21.5
131
Nudity ErasureI2P
Total Count182
38
Image GenerationMS-COCO 10k (test)
FID21.67
24
Concept Erasure RobustnessUnlearn DiffAtk
Nudity Rate68.8
9
Concept Erasure RobustnessRing-a-Bell
Nudity Rate22.93
9
Concept Erasure RobustnessRing-A-Bell Union
Nudity Rate30.27
9
Artistic Style Erasure200-artist dataset
ACCe20.71
9
Concept ErasureConcept Erasure Abstraction
ACCe20.6
9
Concept ErasureConcept Erasure Relationship
ACCe18.5
9
Concept ErasureConcept Erasure Entity
ACCe12.4
9
Showing 10 of 15 rows

Other info

Follow for update