AI Paper Atlas

Research map / Point clouds / Point cloud classification

From Small Scales to Large Scales: Distance-to-Measure Density based Geometric Analysis of Complex Data

Katharina Proksch, Christoph Alexander Weikamp, Thomas Staudt, Benoit Lelandais, Christophe Zimmer

JMLR 2024 · 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
use mass parameter m=1/4980 instead of m=1/1245 (c=2 vs c=0, 5% labeled, k=1) (ablation)DTM-density kNN classification · Simulated chromatin loops c=2 vs c=0misclassification rate: 0.029 -> 0.012Improved
use mass parameter m=1/4980 instead of m=1/9960 (c=2 vs c=0, 5% labeled, k=1) (ablation)DTM-density kNN classification · Simulated chromatin loops c=2 vs c=0misclassification rate: 0.029 -> 0.012Improved
use m=1/4980 instead of m=1/1245 (c=2 vs c=0, 10% labeled, k=5) (ablation)DTM-density kNN classification · Simulated chromatin loops c=2 vs c=0misclassification rate: 0.025 -> 0.012Improved
increase labeled fraction from 5% to 10% (m=1/4980, k=1) (ablation)DTM-density kNN classification · Simulated chromatin loops c=2 vs c=0misclassification rate: 0.012 -> 0.001Improved
increase neighbors k from 1 to 5 (m=1/9960, 5% labeled) (ablation)DTM-density kNN classification · Simulated chromatin loops c=2 vs c=0misclassification rate: 0.029 -> 0.069Hurt
increase neighbors k from 1 to 5 (m=1/1245, 5% labeled) (ablation)DTM-density kNN classification · Simulated chromatin loops c=2 vs c=0misclassification rate: 0.029 -> 0.100Hurt
use a small m instead of m=1 to separate loop densities (ablation)DTM-density kNN classification · Simulated chromatin loops c=0 vs c=25misclassification rateImproved
use m=0.2 instead of m=0.4 or larger (n=10,000) (ablation)DTM-density kNN classification · Wasserstein-geodesic shapes Y1..Y7misclassification rateHurt
increase sample size n from 500 to 2500 (m=1) (ablation)DTM-density kNN classification · Wasserstein-geodesic shapes Y1..Y7misclassification rateImproved
use m=1/9960 or 1/4980 instead of m=1/1245 for clustering (ablation)DTM-density kNN classification · Simulated chromatin loops c=2 vs c=0misclassification rateImproved

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