Abstract
A major goal of bio-inspired artificial intelligence is to design artificial neural networks with abilities that resemble those of animal nervous systems. It is commonly believed that two keys for evolving nature-like artificial neural networks are (1) the developmental process that links genes to nervous systems, which enables the evolution of large, regular neural networks, and (2) synaptic plasticity, which allows neural networks to change during their lifetime. So far, these two topics have been mainly studied separately. The present paper shows that they are actually deeply connected. Using a simple operant conditioning task and a classic evolutionary algorithm, we compare three ways to encode plastic neural networks: a direct encoding, a developmental encoding inspired by computational neuroscience models, and a developmental encoding inspired by morphogen gradients (similar to HyperNEAT). Our results suggest that using a developmental encoding could improve the learning abilities of evolved, plastic neural networks. Complementary experiments reveal that this result is likely the consequence of the bias of developmental encodings towards regular structures: (1) in our experimental setup, encodings that tend to produce more regular networks yield networks with better general learning abilities; (2) whatever the encoding is, networks that are the more regular are statistically those that have the best learning abilities.
Highlights
A major goal of bio-inspired artificial intelligence is to design artificial neural networks (ANNs) with abilities that resemble those of animal nervous systems [1,2,3,4]
The GLA score grows linearly with the size of the evolutionary training set, which is consistent with previous results [27], and the GLA scores obtained with small values of DED are statistically different from those obtained with larger values (e.g., 1 versus 7: p~4|10{4; 4 versus 7, p~2|10{3, 3 versus 6, p~2|10{3; unless otherwise specified, the statistical test in this paper is the Mann-Whitney Utest)
The HNN encoding leads to many networks with at least one symmetry axis, whereas the direct encoding leads to substantially fewer regular networks
Summary
A major goal of bio-inspired artificial intelligence is to design artificial neural networks (ANNs) with abilities that resemble those of animal nervous systems [1,2,3,4]. The genotype of animals does not encode each synapse individually, it instead describes rules of development that are executed multiple times to give birth to networks with regular patterns of connection. Influenced by this concept, many researchers proposed artificial developmental systems with diverse inspirations including chemical gradients [11,12], gene regulatory networks [13,14], cell divisions [15], computational neuro-science models [16] and L-systems [9]
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
