Research map / Graph learning / GNN architectures
The Infinite Contextual Graph Markov Model
Daniele Castellana, Federico Errica, Davide Bacciu, Alessio Micheli
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 changed | Where | Measured | Effect |
|---|---|---|---|
| Replace CGMM's grid-searched layer size with HDP nonparametric states (learned alpha0, gamma) | ICGMM · NCI1 | accuracy: 77.6 vs 76.2 for CGMM | No clear effect |
| Use Bayesian nonparametric HDP layers instead of CGMM | ICGMM · D&D | accuracy: +1.4 over CGMM (76.3 vs 74.9) | Improved |
| Use ICGMM in place of CGMM on social datasets | ICGMM · IMDB-B, REDDIT-B, COLLAB | accuracy: ICGMM better on average than other methods on IMDB-B, REDDIT-B, COLLAB | Improved |
| Replace exact Gibbs sampling with batch-of-one-graph approximate sampling (ICGMMf) (ablation) | ICGMM · chemical datasets (NCI1, D&D, PROTEINS) | accuracy: NCI1 76.4 vs 77.6; D&D 75.1 vs 76.3; PROTEINS 73.2 vs 73.1 | No clear effect |
| Fix alpha0 and gamma (ICGMMαγ) instead of learning them with hyper-priors (ablation) | ICGMM · REDDIT-B | accuracy: 91.3 vs 91.6 for ICGMMf | Hurt |
| Increase depth from 1 to 20 layers (ablation) | ICGMM · NCI1 validation accuracy | accuracy: validation accuracy rises with depth then plateaus near 20 layers | Improved |
| Learn number of states adaptively (HDP) instead of fixed C=20 in CGMM | ICGMM · NCI1 graph embedding size | embedding size: embedding size approximately 7% of CGMM's | Improved |
| Learn number of states adaptively (HDP) instead of fixed C=20 in CGMM | ICGMM · NCI1 unused states | unused states: near-zero unused states vs CGMM's linearly growing unused states | Improved |
| Use Bayesian nonparametric treatment and fixed hyper-priors, reducing configurations tried | ICGMM · social and chemical tasks | number of configurations: 6x reduction in configurations tried | Improved |
Closest papers by meaning
- Contextual Graph Markov Model: A Deep and Generative Approach to Graph Processing (ICML 2018 · similarity 0.92)
- Probabilistic Learning on Graphs via Contextual Architectures (JMLR 2020 · similarity 0.88)
- Graph Neural Networks with a Distribution of Parametrized Graphs (ICML 2024 · similarity 0.86)
- State Space Models on Temporal Graphs: A First-Principles Study (NeurIPS 2024 · similarity 0.86)
- Quality Measures for Dynamic Graph Generative Models (ICLR 2025 · similarity 0.85)
- A Non-Asymptotic Convergent Analysis for Scored-Based Graph Generative Model via a System of Stochastic Differential Equations (ICML 2025 · similarity 0.85)
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.
