AI Paper Atlas

Research map / Point clouds / Point cloud classification

Dimensionality Reduction with Point-distributions Similarity Invariant

Hang Zhang, Kai Ming Ting

ICML 2026 · Point clouds › Point cloud classification · approach: Equivariant point cloud networks

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
preserve point-distribution similarity instead of point-point similarity (Ψ-DR vs PP-DR) (ablation)Ψ-DR · BanknoteCalinski-Harabasz: 108 -> 37146Improved
preserve point-distribution similarity instead of point-point similarity (ablation)Ψ-DR · SkinDavies-Bouldin: 3.47 -> 0.13Improved
preserve point-distribution similarity instead of point-point similarity (ablation)Ψ-DR · MNISTrandom forest accuracy: 0.841 -> 0.97Improved
preserve point-distribution similarity instead of point-point similarity (ablation)Ψ-DR · Gisettek-means NMI: 0.063 -> 0.655Improved
preserve point-distribution similarity instead of point-point similarity (ablation)Ψ-DR · USPSk-means NMI: 0.802 -> 0.799No clear effect
preserve point-distribution similarity instead of point-point similarity (ablation)Ψ-DR · Rcv1k-means NMI: 0.348 -> 0.343No clear effect
preserve point-distribution similarity instead of point-point similarity (ablation)Ψ-DR · LettersCalinski-Harabasz: 509 -> 418.739Hurt
set number of distributions k near the true number of classes (ablation)Ψ-DR · COIL20 and MNISTCH, DB, ACC, NMIImproved
use Ψ-DR instead of t-SNEMNIST embeddings · MNISTrandom forest accuracy: 0.938 -> 0.970Improved
use Ψ-DR instead of t-SNEPendigits embeddings · Pendigitsrandom forest accuracy: 0.996 -> 0.906Hurt
use Ψ-DR instead of t-SNEMNIST embeddings · MNISTcompute time (CPU s): 1030 -> 28.55Improved
use Ψ-DR instead of PaCMAPMNIST embeddings · MNISTCalinski-Harabasz: 81764 -> 52307Hurt

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