AI Paper Atlas

Research map / Graph learning / Graph self-supervised learning

Uncertainty-Constrained Trustworthiness for Graph Learning

Chunhui Zhang, Pengqi Li, Lizhong Ding, Ye Yuan, Peng Yang, Changsheng Li, Guoren Wang

ICML 2026 · Graph learning › Graph self-supervised learning · approach: Heterophily and imbalance on graphs

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
train GNN with DICT objectiveGCN · NBADelta Acc (%): 26.11 -> 1.73Improved
train GNN with DICT objectiveGCN · GermanDelta SP (%): 7.04 -> 1.08Improved
train GNN with DICT objectiveGCN · GermanDelta EO (%): 7.47 -> 0.63Improved
train GNN with DICT objectiveGraphSAGE · BailClean Acc (%): 89.13 -> 94.69Improved
use DICT instead of KCES under NettackGCN on Cora · Cora (Nettack)Attack Acc (%): 78.25 -> 79.32Improved
use DICT instead of KCES under MetattackGCN on Cora · Cora (Metattack)Attack Acc (%): 76.20 -> 78.30Improved
increase Wasserstein radius r beyond 1.0 (ablation)DICT · Germanrobustness and fairnessNo clear effect
set kappa_A to zero (no penalty on structure perturbations) (ablation)DICT · GermanrobustnessHurt
increase kappa_s excessively (ablation)DICT · GermanfairnessHurt
apply DICT to four GNN backbones (GCN, GraphSAGE, GIN, GAT)DICT · German, Bail, NBADelta Acc, Delta SP, Delta EOImproved

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