Announcing StellarGraph Release 0.9.0

The StellarGraph team is happy to announce the latest release of StellarGraph: 0.9.0!

StellarGraph is an open-source Python library implementing a variety of state-of-the-art graph machine learning algorithms. The project is delivered as part of CSIRO’s Data61.

New algorithms:

  • Cluster-GCN
  • Relational-GCN (RGCN)
  • Link prediction for full-batch models

Other Major Improvements/Fixes:

  • Unsupervised GraphSAGE can now be used to create embeddings that are reproducible by fixing the appropriate random seeds.
  • A datasets subpackage provides easier access to sample datasets with inbuilt downloading
  • Various usability and performance improvements for the StellarGraph class.
  • Included as an experimental feature for now, is a way to construct a numpy/pandas based StellarGraph. This will become the new default in favour of the NetworkX-based model in the near future, and should enable some significant performance benefits!

See the full release notes here.