AI Paper Atlas

Research map / Detection and segmentation / CNN features

CORDS - Continuous Representations of Discrete Structures

Tin Hadži Veljković, Erik J Bekkers, Michael Tiemann, Jan-Willem van de Meent

ICLR 2026 · Detection and segmentation › CNN features · approach: Efficient DETR query design

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
represent sets as density+feature fields (CORDS) instead of equivariant GNN generatorsCORDS non-equivariant Erwin backbone · QM9 unconditional generationmolecule stability: CORDS 82.3 vs GeoLDM 89.4, EDM 82.0Mixed
use CORDS fields instead of DETR/YOLO under OOD object countsCORDS · MultiMNIST detectionrelative AP drop: DETR 19.5 -> CORDS 16.4 (drop %)Improved
use CORDS instead of DETR in-distributionCORDS · MultiMNIST detection in-distributionAP: 81.2 -> 76.8Hurt
condition on continuous property directly instead of discretized binsCORDS · QM9 conditional generationatom-count distributionImproved
model non-categorical atom features (charges) in feature fieldsCORDS · GeomDrugsvalidity / atom stabilityImproved
use CORDS instead of GeoLDM on larger moleculesCORDS · GeomDrugsvalidity: 99.3 -> 94.6Hurt

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