Evolution of Spiking Neural Networks Robust to Noise and Damage for Control of Simple Animats

Conference paper

DOI: 10.1007/978-3-319-45823-6_64

Part of the Lecture Notes in Computer Science book series (LNCS, volume 9921)
Cite this paper as:
Wróbel B. (2016) Evolution of Spiking Neural Networks Robust to Noise and Damage for Control of Simple Animats. In: Handl J., Hart E., Lewis P., López-Ibáñez M., Ochoa G., Paechter B. (eds) Parallel Problem Solving from Nature – PPSN XIV. PPSN 2016. Lecture Notes in Computer Science, vol 9921. Springer, Cham

Abstract

One of the central questions of biology is how complex biological systems can continue functioning in the presence of perturbations, damage, and mutational insults. This paper investigates evolution of spiking neural networks, consisting of adaptive exponential neurons. The networks are encoded in linear genomes in a manner inspired by genetic networks. The networks control a simple animat, with two sensors and two actuators, searching for targets in a simple environment. The results show that the presence of noise on the membrane voltage during evolution allows for evolution of efficient control and robustness to perturbations to the value of the neural parameters of neurons.

Keywords

Spiking neural networks Adaptive exponential integrate-and-fire model Genetic algorithm Robustness to noise Robustness to damage 

Copyright information

© Springer International Publishing AG 2016

Authors and Affiliations

  1. 1.Systems Modeling GroupIO PANSopotPoland
  2. 2.Evolutionary Systems GroupUniwersytet im. Adama MickiewiczaPoznańPoland

Personalised recommendations