AI Paper Atlas

Research map / Graph learning / GNN architectures

Manifold GCN: Diffusion-based Convolutional Neural Network for Manifold-valued Graphs

Martin Hanik, Gabriele Steidl, Christoph von Tycowicz, Alzheimer’s Disease Neuroimaging Initiative

IJCV 2026 · Graph learning › GNN architectures · approach: Physics-inspired GNNs · 9 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 one-hot embedding instead of degree embeddingManifold GCN · Synthetic graphs, 90 graphsmean F1: 55.2 -> 67.8Improved
use one-hot embedding instead of degree embeddingManifold GCN · Synthetic graphs, 2880 graphsmean F1: 73.3 -> 85.2Improved
replace hyperbolic H100 with SPD(15) embedding spaceManifold GCN · Synthetic graphs, 6000 graphsmean accuracy: 86.1% -> 85.0%No clear effect
add manifold GCN (one-hot) versus H2H-GCNH2H-GCN · Synthetic graphs, 6000 graphsmean accuracy: 80.9% -> 86.1%Improved
use manifold GCN instead of HGNN with little dataHGNN · Synthetic graphs, 90 graphsmean F1: 41.8 -> 67.8Improved
add tMLPs (instead of manifold FCL) (ablation)Manifold GCN · ADNI normalsaccuracy: 74.5 -> 75.3Improved
add diffusion layers (ablation)Manifold GCN · ADNI normalsaccuracy: 70.5 -> 75.3Improved
add node-wise MLPs (ablation)Manifold GCN · ADNI normalsaccuracy: 70.5 -> 75.3Improved
add equivariance to tMLP (ablation)Manifold GCN · ADNI normalsaccuracy: 70.8 -> 75.3Improved
use sphere-aware S2 normals instead of Euclidean GCN on normalsNormals features · ADNIaccuracy: 70.9 -> 75.3Improved
use differential coordinates instead of normals as featuresManifold GCN · ADNIaccuracy: 75.3 -> 76.9Improved
use manifold GCN instead of DiffusionNet on same differential coordinatesDifferential Coords. · ADNIaccuracy: 67.2 -> 76.9Improved

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