AI Paper Atlas

Research map / Image generation / Text-to-image diffusion

Mass Concept Erasure in Diffusion Models with Concept Hierarchy

Jiahang Tu, Ye Li, Yiming Wu, Hanbin Zhao, Chao Zhang, Hui Qian

AAAI 2026 · Image generation › Text-to-image diffusion · approach: Concept erasure in diffusion models · 3 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add concept hierarchy with group-wise erasure (ablation)SuPLoRA erasure · Mass erasure benchmarkparameters (MB): 28.5MB -> 7.11MBImproved
add concept hierarchy with group-wise erasure (ablation)SuPLoRA erasure · Mass erasure benchmarkerasure accuracy: 6.040% -> 5.830%No clear effect
add SuPLoRA (ablation)hierarchical group-wise erasure · Mass erasure benchmarkdomain-specific accuracy: 67.81% -> 72.94%Improved
add SuPLoRA (ablation)hierarchical group-wise erasure · Mass erasure benchmarksupertype CLIP score: 25.19 -> 26.21Improved
add diffusion regularization loss (ablation)SuPLoRA erasure · Mass erasure benchmarkMS-COCO FID: 18.97 -> 17.94Improved
add diffusion regularization loss (ablation)SuPLoRA erasure · Mass erasure benchmarkerasure accuracy: 4.360% -> 5.830%Hurt
train both A and B instead of freezing B (ablation)SuPLoRA · Mass erasure benchmarksupertype CLIP score: 26.21 -> 25.22Hurt
freeze random down-projection B (ablation)standard LoRA · Mass erasure benchmarksupertype CLIP score: 25.19 -> 26.08Improved
initialize frozen B with orthogonal complement of supertype subspace (ablation)LoRA with frozen B · Mass erasure benchmarksupertype CLIP score: 26.08 -> 26.21Improved
use SuPLoRA instead of MACEmass erasure · Mass erasure benchmarkretained celebrity accuracy: 78.50% -> 83.38%Improved
use SuPLoRA instead of MACEmass erasure · Mass erasure benchmarkcelebrity erasure accuracy: 6.250% -> 7.500%Hurt
use SuPLoRA instead of MACEmass erasure · Mass erasure benchmarkstorage (MB): 198 -> 154Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy