The latest release marks an important next phase in Graphistry: AutoML for graph AI via Graph Neural Networks. Likewise, already popular with analysts, it adds the new multiselection panel. We’ll be sharing tutorials in the coming weeks on individual features, and some of our favorites already include:
- PyGraphistry[AI]: Automatic Graph AI with RGCNs: Graph neural networks are increasingly achieve best-of-class on different problems, and the new method g.embed() automates the RGCN model, which is great for heterogeneous data. Any Graphistry object, can now automatically be run through a GNN, such as for detecting anomalies in SSH logins.
- Multiselect panel: Select multiple events or entities, then filter/exclude them, and use the new selection expander tool to investigate paths related to them.
- Cross-origin embedding: The update significantly improves cross-origin usage around enabling authentication and authorization
- SSO: Added support to embedded usage
- PyGraphistry[AI]: GPU UMAP and neural search
Learn more in our public release notes.