my most recent publication is also my 1st sole-authored paper. it brings transparency to graph Laplacian methods by converting the eigenvector subspace to a topologically-informed probabilistic/fuzzy/soft embedding of the graph nodes. the source code is not available because i am working on commercial applications of this technique. if anyone has a graph/network analysis problem they want to try it on, please let me know! maybe we can develop a product for it :) #ml

https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0204096