AI Paper Atlas

Research map / Graph learning / GNN architectures

Learning Dynamic Stability Landscapes in Synchronization Networks

Christian Nauck, Junyou Zhu, Michael Lindner, Frank Hellmann

ICML 2026 · Graph learning › GNN architectures · approach: Physics-inspired GNNs · 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 CNN decoder instead of MLP decoder (ablation)DBGNN encoder · Heatmap SSIM, train 20 eval 100SSIM (%): 75.49 -> 81.03Improved
use CNN decoder instead of MLP decoder (ablation)TAG encoder · Heatmap SSIM, train 20 eval 20SSIM (%): 81.37 -> 84.31Improved
use DBGNN encoder instead of TAG (ablation)CNN decoder · Heatmap SSIM, train 20 eval 100SSIM (%): 74.36 -> 81.03Improved
use DBGNN encoder instead of TAG on real grids (ablation)CNN decoder · Germany zero-shotSSIM (%): 82.87 -> 53.10Hurt
predict landscapes then recover SNBS instead of direct SNBS regressionDBGNN · SNBS R2, train 20 eval 20R2 (%): 85.68 -> 83.99Mixed
predict landscapes then recover SNBS on OOD size shiftDBGNN · SNBS R2, train 20 eval 100R2 (%): 73.73 -> 70.66Hurt
train on 100-node graphs instead of 20-node graphs, evaluate on 100 (ablation)DBGNN-CNN · Heatmap SSIMSSIM (%): 81.03 -> 84.90Improved
MLP vs CNN decoder for contingency identification (ablation)TAG encoder · IoU, train 100 eval 100IoU: 0.41 -> 0.47Improved
MLP vs CNN decoder for contingency identification with TAG, train 20 eval 20 (ablation)TAG encoder · IoU, train 20 eval 20IoU: 0.58 -> 0.58No clear effect

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