AI Paper Atlas

Research map / Image generation / Fast diffusion sampling

Moderating the Generalization of Score-based Generative Model

Wan Jiang, He Wang, Xin Zhang, Dan Guo, Zhaoxin Fan, Yunfeng Diao, Richang Hong

ICCV 2025 · Image generation › Fast diffusion sampling · approach: Distributional control of diffusion · 1 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
Apply MSGM score adjustment instead of re-training only on retained dataVE SDE · Toy Gaussian mixture, NSFG data D_fNLL on D_f: 11.59 -> 39.01Improved
Use Orthogonal-MSGM instead of Unseen by Re-trainingVE SDE · MNIST classes 3 and 7Unlearning ratio (%): 4.1 -> 1.2Improved
Use Orthogonal-MSGM instead of Unseen by Re-trainingVE SDE · MNIST classes 3 and 7NLL on D_f: 3.01 -> 13.23Improved
Use Obtuse-MSGM instead of Unseen by Re-trainingVP SDE · CelebA bangsUnlearning ratio (%): 6.7 -> 0.7Improved
Use Obtuse-MSGM instead of Orthogonal-MSGM for feature unlearningVP SDE · CelebA bangsUnlearning ratio (%): 3.5 -> 0.7Improved
Use Orthogonal-MSGM instead of Obtuse-MSGM for class unlearningVE SDE · MNIST classes 3 and 7Unlearning ratio (%): 5.1 -> 1.2Improved
Use MSGM instead of EraseDiffVP SDE · CIFAR-10 automobile and dogUnlearning ratio (%): 17.9 -> 11.9Improved
Fine-tune a pre-trained DDPM with Obtuse-MSGM instead of EraseDiffDDPM · CIFAR-10 automobile and dog (fine-tuned)Unlearning ratio (%): 7.4 -> 5.2Improved
Use MSGM instead of Unseen in SD v1.4 unlearningSD v1.4 · Imagenette tenchFID of D_g: 3.09 -> 3.08Improved
Use pseudo-label cyprinoid with EraseDiff rather than MSGMEraseDiff · Imagenette tenchClassification rate: 0.00 -> 0.47Hurt
Use Obtuse-MSGM instead of Standard on inpaintingVP SDE · CelebA inpainting D_fClassification accuracy (%): 53.0 -> 0.5Improved
Use Obtuse-MSGM instead of Standard on inpaintingVP SDE · CIFAR-10 inpainting D_fClassification accuracy (%): 75.0 -> 49.6Improved

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