Universal Minimization on the Node Domain
An experimentation framework that assesses how well graph neural networks (GNN) can minimize various attributed graph functions on the node domain.
Developed by
License
Apache License 2.0 (Apache-2.0)
Main Characteristic
- Based on torch geometric.
- Modular architecture definition.
- Implementation of several diffusion-based architectures (GCN, GCNII, APPNP, S2GC, DeepSet on graphs).
- Several benchmarking tasks for the ability to approximate equivariant attributed graph functions.
- Uniform interface that treats multiple graphs as one disconnected graph.
Technical Categories
AI services
Last updated
27.05.2024 - 14:13
Detailed Description
Users of this framework first declare a predictive task on which to assess a GNN architecture, and obtain its training-test-validation data subtasks. If there are multiple graphs, they are packed together in one unconnected graph. Then, they declare an architecture and train it. Losses and predictive performances are retrieved from the predictive tasks.