Graphs, recommendations, and time series · MLD-0489
Graph neural network
A neural model that uses graph structure to learn representations of nodes, edges, or whole graphs.
Original reference page · detailed expansion pending. Original definitions, examples, notes and citations are preserved below. New examples are illustrative, not benchmark results.
Original example
Predict properties of molecular graphs.
Original caveat
Different architectures have different structural expressiveness.
Context and aliases
Level: Intermediate. Stages: Architecture · Data · Training · Inference.
Tags: ML · DL · Graph ML.
Aliases: GNN.
Related concepts
Concepts linking here
Technology connections
- DGL — Framework for graph machine learning.
- Diffrax — Differential equation solving library for JAX.
- DoWhy — Causal inference framework.
- EconML — Machine-learning methods for heterogeneous treatment effects.
- GPyTorch — Gaussian-process modeling library in PyTorch.
- NetworkX — Python library for graph analysis.
- NumPyro — JAX-based probabilistic programming library.
- PyMC — Probabilistic programming library for Bayesian models.
- PyTorch Geometric — Graph neural-network library built on PyTorch.
Original references
- PyTorch Geometric documentation ↗ · Consulted
- scikit-learn User Guide ↗ · Consulted
- statsmodels: time series analysis ↗ · Reference link
- Original reference URL ↗