AI Paper Atlas

Research map / Label-efficient and robust learning / Semi-supervised learning

Semi-Supervised Learning for Molecular Graphs via Ensemble Consensus

Rasmus Hannibal Tirsgaard, Laurits Fredsgaard, Marisa Wodrich, Mikkel Jordahn, Mikkel N. Schmidt

ICML 2026 · Label-efficient and robust learning › Semi-supervised learning · approach: Holistic semi-supervised pipelines · 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
add ensemble consensus loss on unlabeled dataPaiNN individual member · QM9 U0MAE (meV): 24.8752 -> 19.9642Improved
add ensemble consensus loss on unlabeled dataPaiNN ensemble M=4 · QM9 epsilon HOMOMAE (meV): 76.4713 -> 73.0755Improved
use a consensus-trained individual member instead of a supervised ensemblePaiNN · QM9 U0MAE (meV): 20.9101 -> 19.9642Improved
use ensemble consensus instead of mean-teacherPaiNN ensemble · QM9 U0MAE (meV): 24.7288 -> 19.3816Improved
use ensemble consensus instead of PSEUDsigmaPaiNN · QM9 <R2>MAE: 0.9099 -> 0.5605Improved
increase ensemble size from 2 to 4 with consensus training (ablation)PaiNN ensemble · QM9 U0MAE: 20.43 -> 19.38Improved
add consensus loss to GCNGCN individual · ZINCMAE: .3163 -> .2406Improved
add consensus loss to GatedGCNGatedGCN individual · ogbg-molhivAUROC: .7312 -> .7361No clear effect
add consensus loss to GINGIN individual · Peptides-structMAE: .2966 -> .2944No clear effect
add mean-teacher instead of ensemble consensusGCN individual · ZINCMAE: .2406 (consensus) vs .2884 (mean teacher)Hurt
add consensus coupling in CIFAR-10 ResNet-18 ensembleResNet-18 ensemble · CIFAR-10 ensemble calibrationECE and BrierHurt
add consensus coupling to the ensembleindividual members on ogbg-molhiv · ogbg-molhivcalibrationImproved

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