Transparency of execution using epigenetic networks

Dethlefs, Nina.; Turner, Alexander.

doi:10.7551/ecal_a_068

Authors

Dr Nina Dethlefs N.Dethlefs@hull.ac.uk
Lecturer

Dr Alexander Turner Alexander.Turner@hull.ac.uk
Lecturer

Contributors

Carole Knibbe
Editor

Guillaume Beslon
Editor

David Parsons
Editor

Dusan Misevic
Editor

Jonathan Rouzaud-Cornabas
Editor

Nicolas Bredeche
Editor

Salima Hassas
Editor

Olivier Simonin
Editor

Hedi Soula
Editor

Abstract

This paper describes how the recurrent connectionist architecture epiNet, which is capable of dynamically modifying its topology, is able to provide a form of transparent execution. EpiNet, which is inspired by eukaryotic gene regulation in nature, is able to break its own architecture down into sets of smaller interacting networks. This allows for autonomous complex task decomposition, and by analysing these smaller interacting networks, it is possible to provide a real world understanding of why specific decisions have been made. We expect this work to be useful in fields where the risk of improper decision making is high, such as medical simulations, diagnostics and financial modelling. To test this hypothesis we apply epiNet to two data sets within UCI’s machine learning repository, each of which requires a specific set of behaviours to solve. We then perform analysis on the overall functionality of epiNet in order to deduce the underlying rules behind its functionality and in turn provide transparency of execution.

Journal Article Type	Article
Start Date	Sep 4, 2017
Publication Date	2017-09
Journal	The European Conference on Artificial Life 2017
Publisher	Massachusetts Institute of Technology Press (MIT Press)
Peer Reviewed	Not Peer Reviewed
Volume	14
Pages	404-411
Book Title	Proceedings of the ECAL 2017
ISBN	978-0-262-34633-7
DOI	https://doi.org/10.7551/ecal_a_068
Keywords	Transparency; Artificial intelligence; Topological Morphology
Publisher URL	http://cognet.mit.edu/proceed/10.7551/ecal_a_068