Classical graph modeling approaches such as Erd\H{o}s R\'{e}nyi (ER) randomgraphs or Barab\’asi-Albert (BA) graphs, here referred to as stylized models, aim to reproduce properties of real-world graphs in an interpretable way . Previous work by Stoehr et al. (2019)addresses these issues by learning the generation process from graph data . We focus on recovering thegenerative parameters of BA graphs by replacing their $beta$-VAE decoder with a sequential one . We first learn the generative BA parameters in a supervisedfashion using a Graph Neural Network (GNN) and a Random Forest Regressor, byminimizing the squared loss between the true generative parameters and thelatent variables . Next, we train a $beta-VAe model

Author(s) : Cristina Guzman, Daphna Keidar, Tristan Meynier, Andreas Opedal, Niklas Stoehr

Links : PDF - Abstract

Code :

https://github.com/doty-k/world_models


Coursera

Keywords : graphs - parameters - graph - ba - recovering -

Leave a Reply

Your email address will not be published. Required fields are marked *