Publication

Back to overview

Dynamics of Evolving Feed-Forward Neural Networks and Their Topological Invariants

Type of publication Peer-reviewed
Publikationsform Contribution to book (peer-reviewed)
Publication date 2016
Author Masulli Paolo, Villa Alessandro E.P.,
Project NeurEcA: Does working memory training affect decision making? A NeuroEconomic study of ADHD adults and healthy controls
Show all

Contribution to book (peer-reviewed)

Book Artificial Neural Networks and Machine Learning – ICANN 2016
Editor , Villa Alessandro E.P.
Publisher Springer Verlag, Switzerland
Page(s) 99 - 106
ISBN 978-3-319-44777-3
Title of proceedings Artificial Neural Networks and Machine Learning – ICANN 2016
DOI 10.1007/978-3-319-44778-0_12

Open Access

URL https://serval.unil.ch/resource/serval:BIB_8777B6A356F7.P001/REF
Type of Open Access Repository (Green Open Access)

Abstract

The evolution of a simulated feed-forward neural network with recurrent excitatory connections and inhibitory forward connections is studied within the framework of algebraic topology. The dynamics includes pruning and strengthening of the excitatory connections. The invariants that we define are based on the connectivity structure of the underlying graph and its directed clique complex. The computation of this complex and of its Euler characteristic are related with the dynamical evolution of the network. As the network evolves dynamically, its network topology changes because of the pruning and strengthening of the connections and algebraic topological invariants can be computed at different time steps providing a description of the process. We observe that the initial values of the topological invariant computed on the network before it evolves can predict the intensity of the activity.
-