There's some really interesting research going on in the deep learning world where they are trying to find the best way to represent graphs in vector space.

For example hyperbolic embeddings for trees: https://dawn.cs.stanford.edu/2018/03/19/hyperbolics/

I wonder if this book will be useful for that purpose? (the term "vector labeling" came up a few hundred times but I am not sure if that's the same thing)