Show simple item record

dc.contributor.advisor2016. The authors
dc.contributor.authorChurchill, AWen_US
dc.contributor.authorSigtia, Sen_US
dc.contributor.authorFernando, Cen_US
dc.date.accessioned2016-09-01T10:06:42Z
dc.date.submitted2016-06-20T10:20:54.853Z
dc.identifier.urihttp://qmro.qmul.ac.uk/xmlui/handle/123456789/14930
dc.description.abstractNeural networks and evolutionary computation have a rich intertwined history. They most commonly appear together when an evolutionary algorithm optimises the parameters and topology of a neural network for reinforcement learning problems, or when a neural network is applied as a surrogate fitness function to aid the evolutionary optimisation of expensive fitness functions. In this paper we take a different approach, asking the question of whether a neural network can be used to provide a mutation distribution for an evolutionary algorithm, and what advantages this approach may offer? Two modern neural network models are investigated, a Denoising Autoencoder modified to produce stochastic outputs and the Neural Autoregressive Distribution Estimator. Results show that the neural network approach to learning genotypes is able to solve many difficult discrete problems, such as MaxSat and HIFF, and regularly outperforms other evolutionary techniques.en_US
dc.language.isoenen_US
dc.subjectcs.NEen_US
dc.subjectcs.NEen_US
dc.titleLearning to Generate Genotypes with Neural Networksen_US
dc.typeArticle
pubs.author-urlhttp://arxiv.org/abs/1604.04153v1en_US
pubs.notesNot knownen_US


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record