Evolving plasticity for autonomous learning under changing environmental conditions

Research output: Contribution to journalArticleAcademic

Abstract

A fundamental aspect of learning in biological neural networks (BNNs) is the plasticity property which allows them to modify their configurations during their lifetime. Hebbian learning is a biologically plausible mechanism for modeling the plasticity property based on the local activation of neurons. In this work, we employ genetic algorithms to evolve local learning rules, from Hebbian perspective, to produce autonomous learning under changing environmental conditions. Our evolved synaptic plasticity rules are capable of performing synaptic updates in distributed and self-organized fashion, based only on the binary activation states of neurons, and a reinforcement signal received from the environment. We demonstrate the learning and adaptation capabilities of the ANNs modified by the evolved plasticity rules on a foraging task in a continuous learning settings. Our results show that evolved plasticity rules are highly efficient at adapting the ANNs to task under changing environmental conditions.
LanguageEnglish
Article number1904.01709v1
Number of pages26
JournalarXiv
StatePublished - 2019

Fingerprint

Plasticity
Neurons
Chemical activation
Reinforcement
Genetic algorithms
Neural networks

Cite this

@article{58470a88309245a1a0bdb5b936f43563,
title = "Evolving plasticity for autonomous learning under changing environmental conditions",
abstract = "A fundamental aspect of learning in biological neural networks (BNNs) is the plasticity property which allows them to modify their configurations during their lifetime. Hebbian learning is a biologically plausible mechanism for modeling the plasticity property based on the local activation of neurons. In this work, we employ genetic algorithms to evolve local learning rules, from Hebbian perspective, to produce autonomous learning under changing environmental conditions. Our evolved synaptic plasticity rules are capable of performing synaptic updates in distributed and self-organized fashion, based only on the binary activation states of neurons, and a reinforcement signal received from the environment. We demonstrate the learning and adaptation capabilities of the ANNs modified by the evolved plasticity rules on a foraging task in a continuous learning settings. Our results show that evolved plasticity rules are highly efficient at adapting the ANNs to task under changing environmental conditions.",
author = "Anil Yaman and Decebal Mocanu and Giovanni Iacca and Matt Coler and George Fletcher and Mykola Pechenizkiy",
year = "2019",
language = "English",
journal = "arXiv",
publisher = "Cornell University Library",

}

Evolving plasticity for autonomous learning under changing environmental conditions. / Yaman, Anil; Mocanu, Decebal; Iacca, Giovanni; Coler, Matt; Fletcher, George; Pechenizkiy, Mykola.

In: arXiv, 2019.

Research output: Contribution to journalArticleAcademic

TY - JOUR

T1 - Evolving plasticity for autonomous learning under changing environmental conditions

AU - Yaman,Anil

AU - Mocanu,Decebal

AU - Iacca,Giovanni

AU - Coler,Matt

AU - Fletcher,George

AU - Pechenizkiy,Mykola

PY - 2019

Y1 - 2019

N2 - A fundamental aspect of learning in biological neural networks (BNNs) is the plasticity property which allows them to modify their configurations during their lifetime. Hebbian learning is a biologically plausible mechanism for modeling the plasticity property based on the local activation of neurons. In this work, we employ genetic algorithms to evolve local learning rules, from Hebbian perspective, to produce autonomous learning under changing environmental conditions. Our evolved synaptic plasticity rules are capable of performing synaptic updates in distributed and self-organized fashion, based only on the binary activation states of neurons, and a reinforcement signal received from the environment. We demonstrate the learning and adaptation capabilities of the ANNs modified by the evolved plasticity rules on a foraging task in a continuous learning settings. Our results show that evolved plasticity rules are highly efficient at adapting the ANNs to task under changing environmental conditions.

AB - A fundamental aspect of learning in biological neural networks (BNNs) is the plasticity property which allows them to modify their configurations during their lifetime. Hebbian learning is a biologically plausible mechanism for modeling the plasticity property based on the local activation of neurons. In this work, we employ genetic algorithms to evolve local learning rules, from Hebbian perspective, to produce autonomous learning under changing environmental conditions. Our evolved synaptic plasticity rules are capable of performing synaptic updates in distributed and self-organized fashion, based only on the binary activation states of neurons, and a reinforcement signal received from the environment. We demonstrate the learning and adaptation capabilities of the ANNs modified by the evolved plasticity rules on a foraging task in a continuous learning settings. Our results show that evolved plasticity rules are highly efficient at adapting the ANNs to task under changing environmental conditions.

UR - https://arxiv.org/abs/1904.01709

M3 - Article

JO - arXiv

T2 - arXiv

JF - arXiv

M1 - 1904.01709v1

ER -