qitianwu/GraphOOD-GNNSafe

The official implementation for ICLR23 paper "GNNSafe: Energy-based Out-of-Distribution Detection for Graph Neural Networks"

27
/ 100
Experimental

This project helps identify unusual or 'out-of-distribution' nodes within complex, interconnected datasets like social networks, citation graphs, or product recommendation systems. It takes in existing graph data, where some nodes are known to be typical and others are potentially unusual, and outputs predictions on whether new nodes are typical or unusual. Data scientists and machine learning engineers working with graph-structured data will find this useful for flagging anomalies.

No commits in the last 6 months.

Use this if you need to detect anomalous or novel data points in your graph-structured datasets, such as identifying fraudulent accounts in a social network or emerging research topics in a citation graph.

Not ideal if your data is not structured as a graph, or if you are solely focused on standard classification tasks where all test data is expected to come from the same distribution as your training data.

anomaly-detection graph-analytics data-quality network-security fraud-detection
No License Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 9 / 25
Maturity 8 / 25
Community 10 / 25

How are scores calculated?

Stars

84

Forks

6

Language

Python

License

Last pushed

Jul 27, 2023

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/qitianwu/GraphOOD-GNNSafe"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.