Great article. Not being an machine learning specialist myself, despite having a current task requiring it, this is the first I’ve come across the concept of graph or geometric machine learning. It looks very exciting, extremely relevant to the overall general aims of the VRENAR project, which has at its heart a 3D virtual environment. I can feel a possible complementary article of my own, outlining some exciting applicability coming on already.

Thanks greatly for posting, and commendations to you for choosing to make your knowledge on this openly available. I think it is really important for powerful ML related info to be in the public domain.

The difference between conventional ML and geometric ML seems a little like the concepts of volume and mass. They look closely related, yet it took more than 1500 years for folk like Galileo and Newton to come up with the ideas of mass and gravity, after Archimedes came up with the idea of volume.

I am glad we didn’t need to wait that long for graph ML!

I will look forward to the promised future installments.

Perhaps you will add links to that initial article as an index to the others, as they appear?