Show HN: KarateClub a Python library for unsupervised machine learning on graphs

I am by no means an expert in this space, but I'm pretty sure the most straightforward answers will be some form of "clustering". Find cliques among personal connections, business networks among companies, etc

I am the OP.

A few examples:

1. DeepWalk: You have the social network of users and for some of them their age. You want to predict the missing age values so that You are able to design a good targeted marketing campaign. The solution that DeepWalk gives: Learn a node level embedding, learn to predict the age based on the embedding, make a prediction for the missing users.

2. Graph2Vec: You have the molecules of hundred thousands of materials. You know that some of them is a carcinogen, while others are not. How can you predict whether a molecule is a carcinogen? The solution that Graph2Vec gives: Learn a graph level embedding, learn to predict the carcinogen / not carcinogen target and predict whether a molecule with unknown status is a carcinogen or not.

Is dead. Dead does not mean fake.