Geometric Deep Learning: a Temperature Based Analysis of Graph Neural Networks

M. Lapenna; F. Faglioni; F. Zanchetta; R. Fioresi

Geometric Deep Learning: a Temperature Based Analysis of Graph Neural Networks

Machine Learning 2023-09-06 v1 Artificial Intelligence

Authors: M. Lapenna , F. Faglioni , F. Zanchetta , R. Fioresi

Abstract

We examine a Geometric Deep Learning model as a thermodynamic system treating the weights as non-quantum and non-relativistic particles. We employ the notion of temperature previously defined in [7] and study it in the various layers for GCN and GAT models. Potential future applications of our findings are discussed.

Keywords

deep neural networks graph neural network deep learning

Cite

@article{arxiv.2309.00699,
  title  = {Geometric Deep Learning: a Temperature Based Analysis of Graph Neural Networks},
  author = {M. Lapenna and F. Faglioni and F. Zanchetta and R. Fioresi},
  journal= {arXiv preprint arXiv:2309.00699},
  year   = {2023}
}

Comments

Published on Proceedings of GSI 2023

Related papers

View all related →