Research map / Graph learning / GNN architectures
Gaussian-Induced Convolution for Graphs
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 changed | Where | Measured | Effect |
|---|---|---|---|
| Use edge-induced Gaussian filtering instead of conventional graph convolution filtering (ablation) | GIC · MUTAG, PTC, NCI1, NCI109, ENZYMES, PROTEINS | accuracy | Helped |
| Add VI-GMM graph coarsening (ablation) | GIC · MUTAG, PTC, NCI1, NCI109, ENZYMES | accuracy: 93.33 -> 94.44; 68.23 -> 77.64; 79.17 -> 84.08; 77.81 -> 82.86; 52.00 -> 62.50 | Helped |
| Add VI-GMM graph coarsening (ablation) | GIC · PROTEINS | accuracy: 78.19 -> 77.65 | Hurt |
| Increase receptive-field size and Gaussian component count from 1 to 7 (ablation) | GIC · MUTAG | accuracy: 67.77 -> 94.44 | Helped |
| Increase receptive-field size and Gaussian component count from 3 to 7 (ablation) | GIC · ENZYMES | accuracy: 64.00 -> 62.50 | Hurt |
| Increase receptive-field size and Gaussian component count from 5 to 7 (ablation) | GIC · NCI1 | accuracy: 84.47 -> 84.08 | Hurt |
| Increase network depth from 2 to 8 layers (ablation) | GIC · MUTAG | accuracy: 86.66 -> 94.44 | Helped |
| Increase network depth from 2 to 8 layers (ablation) | GIC · ENZYMES | accuracy: 42.33 -> 62.50; peak 64.83 at N=6 | Mixed |
| Use GIC rather than GraphSAGE-LSTM aggregation | Node classification · PPI | micro-F1: 0.612 -> 0.661 | Helped |
| Use GIC rather than GraphSAGE-LSTM aggregation | Node classification · Reddit | micro-F1: 0.954 -> 0.952 | Hurt |
| Use GIC rather than the strongest listed graph-classification baseline | Graph classification · PTC, NCI1, ENZYMES, PROTEINS, COLLAB, IMDB-BINARY, IMDB-MULTI | accuracy | Helped |
| Use GIC rather than the strongest listed graph-classification baseline | Graph classification · MUTAG, NCI109, REDDIT-BINARY, REDDIT-MULTI-5K, REDDIT-MULTI-12K | accuracy | Hurt |
Closest papers by meaning
- Graph Convolution for Semi-Supervised Classification: Improved Linear Separability and Out-of-Distribution Generalization (ICML 2021 · similarity 0.90)
- Graph Convolutional Gaussian Processes (ICML 2019 · similarity 0.90)
- Effects of Graph Convolutions in Multi-layer Networks (ICLR 2023 · similarity 0.89)
- G$^2$CN: Graph Gaussian Convolution Networks with Concentrated Graph Filters (ICML 2022 · similarity 0.89)
- Learning Convolutional Neural Networks for Graphs (ICML 2016 · similarity 0.89)
- Neural Gaussian Similarity Modeling for Differential Graph Structure Learning (AAAI 2024 · similarity 0.87)
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.