AI Paper Atlas

Research map / Graph learning / GNN architectures

RADE: Random Add-Drop Edge as a Regularizer

Danial Saber, Amirali Salehi-Abari

ICML 2026 · Graph learning › GNN architectures · approach: GNN generalization theory · 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
use RADE-OF instead of plain GCNGCN · ogbn-arxivaccuracy: 70.51 -> 71.09Improved
use RADE-OF instead of DropEdgeGCN · Coraaccuracy: 80.28 -> 81.08Improved
use RADE-OFS instead of GCNGCN · peptides-funcAP: 55.14 -> 56.97Improved
use RADE-OFS instead of GCNGCN · MUTAGaccuracy: 84.00 -> 86.24Improved
RADE-OFS vs RADE-OF on node classificationRADE · node classification (8 datasets)accuracyNo clear effect
add GradNorm adaptive rates (ablation)RADE-OF · Flickrmicro-F1: 51.42 -> 52.25Improved
add expectation-preserving correction (with fixed rates) (ablation)RADE-OF · Flickrmicro-F1: 49.35 -> 51.42Improved
add expectation-preserving correction (with fixed rates) (ablation)RADE-OF · Coraaccuracy: 78.94 -> 80.90Improved
use nonlinear GCN instead of linearized SGC backbone (ablation)RADE-OF · Flickrmicro-F1: 50.65 -> 52.25Improved
combine add and drop instead of drop-only (ablation)RADE-OF · ogbn-arxivaccuracy: 70.68 -> 71.09Improved
combine add and drop instead of add-only (ablation)RADE-OF · Coraaccuracy: 80.68 -> 81.08Improved
use adaptive GradNorm rates instead of tuned fixed rates (ablation)RADE-OFS · Computeraccuracy: 88.31 -> 89.94Improved

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