AI Paper Atlas

Research map / Learning theory and algorithms / Kernels and regression theory

GeoDM: Geometry-aware Distribution Matching for Dataset Distillation

Xuhui Li, Zhengquan Luo, Zihui Cui, Kai Zhao, zhiqiang xu

ICML 2026 · Learning theory and algorithms › Kernels and regression theory · approach: Statistical perspectives on generalization · 0 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 product space (Euclidean+hyperbolic+spherical) instead of Euclidean-only matching (ablation)NCFM-based distribution matching · CIFAR-10 IPC=10accuracy: 71.8 -> 73.5Improved
add learnable curvature and weights to product space (ablation)GeoDM · CIFAR-10 IPC=10accuracyImproved
add geometry-aware OT loss to the full product model (ablation)GeoDM · CIFAR-10 IPC=10accuracy: 73.5 -> 74.4Improved
replace NCFM objective with simple DM objective inside product space (ablation)GeoDM · CIFAR-10 IPC=10accuracy: up to 2.79% over single-space baselinesImproved
replace NCFM objective with DSDM inside product space (ablation)GeoDM · CIFAR-10 IPC=10accuracy: around 2%Improved
use product space with curvature and weights instead of Euclidean NCFM, at matched accuracy (ablation)NCFM · CIFAR-10 IPC=10 total costGFLOPs: 270.0 -> 181.7Improved
add OT loss to product space with curvature and weights (ablation)GeoDM · CIFAR-10 IPC=10 total costGFLOPs: 181.7 -> 242.4Hurt
transfer distilled data to ResNet-18 and AlexNetNCFM · CIFAR-10 IPC=10 cross-architectureaccuracy: ResNet-18 67.3±0.4 -> 69.1±0.5Improved
apply GeoDM to higher-resolution dataNCFM · Imagenette IPC=10accuracy: 77.3 -> 78.8Improved
apply GeoDM at high IPCNCFM · CIFAR-100 IPC=50accuracy: 54.7 -> 55.0No clear effect
apply GeoDM at low IPCNCFM · CIFAR-100 IPC=1accuracy: 34.4 -> 38.0Improved
replace 50% of CIFAR-10 with Gaussian noiseGeoDM learned curvature · CIFAR-10 noise testlearned curvature: spherical 1.41 -> 0.84; hyperbolic -0.79 -> -0.48Hurt

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