My team often use Python to perform GNN/Graph Embedding (i.e. https://paperswithcode.com/paper/accurately-modeling-biased-random-walks-on, https://paperswithcode.com/paper/hdmi-high-order-deep-multiplex-infomax, etc.) As these are fast becoming generic algorithms used in many applications, I look forward to native implementation in Mathematica. Is this the case, does anyone know if/when Mathematica will embrace the Geometric Deep Learning community? Kind thanks!