AI Paper Atlas

Research map / Label-efficient and robust learning / Knowledge distillation

Rectified Decoupled Dataset Distillation: A Closer Look for Fair and Comprehensive Evaluation

Xinhao Zhong, Shuoyang Sun, Xulin Gu, Chenyang Zhu, Bin Chen, Yaowei Wang

ICLR 2026 · Label-efficient and robust learning › Knowledge distillation · approach: Self and data-free distillation

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
re-evaluate all methods under a unified post-evaluation protocol (ablation)decoupled dataset distillation methods · ImageNet-1K IPC=10top-1 accuracy: SRe2L 21.3 -> 40.2; EDC 48.6 -> 46.9Improved
add multi-teacher hybrid soft labels (ablation)SRe2L, RDED, D4M, EDC · ImageNet-1K IPC=1top-1 accuracy: SRe2L 4.1 -> 12.2; RDED 7.6 -> 15.6Improved
add multi-teacher hybrid soft labels (ablation)EDC · ImageNet-1K IPC=50top-1 accuracy: 60.1 -> 57.1Hurt
use Gaussian noise instead of real/RDED initialization (ablation)EDC · ImageNet-1K IPC=10top-1 accuracy: 46.9 -> 36.5Hurt
initialize with random real images instead of noise (ablation)G-VBSM · ImageNet-1K IPC=10top-1 accuracy: 41.5 -> 46.4Improved
replace KL loss with MSE+GT loss (ablation)decoupled DD methods · ImageNet-1K IPC=1/10/50/100top-1 accuracy: EDC IPC=100 63.2 -> 64.1; SRe2L IPC=1 4.1 -> 3.6Mixed
random sampling with soft labels as distilled setrandom baseline vs distillation methods · ImageNet-1K IPC=10top-1 accuracy: random 45.8 vs SRe2L 40.2Improved
use soft labels instead of hard labels for random images (ablation)random sampling · ImageNet-1K IPC=10top-1 accuracy: 6.1 -> 45.8Improved
smaller learning-rate smoothing factor with larger eval models (ablation)evaluation protocol · ImageNet-1K IPC=10top-1 accuracyMixed
use diffusion-based generation (D4M) instead of optimization (SRe2L)evaluation on ViT-B-16 · ImageNet-1K IPC=50top-1 accuracy: 53.4 -> 63.7Improved

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