Welcome to GSNN’s Documentation!
The Graph Structured Neural Networks (GSNN) library provides flexible and scalable tools for learning on graph-structured data.
Overview
Tutorials
- General Premise
- Simulating structured data
- Performance comparison on simulated data
- Reinforcement learning for structure optimization
- Gradient checkpointing and compiling
- Uncertainty quantification with hypernetworks
- GSNN Interpretation methods
- DrugCell implementation example
- Inferring output edges
- Pathway latent factor regularization
- Function-node activity gating
- Per-channel function-node activity gating
- Inferring function-function edges (Tier 0)
- Online edge inference via auxiliary regression (Tier 0+)
- Function Edge Inferer
- Post-hoc edge inference via shared-embedding link prediction (node2vec)
API Reference