Feature Discovery with Deep Learning Algebra Networks

Created by W.Langdon from gp-bibliography.bib Revision:1.8051

@InProceedings{Korns:2021:GPTP,

• author = "Michael Korns",
• title = "Feature Discovery with Deep Learning Algebra Networks",
• booktitle = "Genetic Programming Theory and Practice XVIII",
• year = "2021",
• editor = "Wolfgang Banzhaf and Leonardo Trujillo and Stephan Winkler and Bill Worzel",
• series = "Genetic and Evolutionary Computation",
• pages = "109--127",
• address = "East Lansing, USA",
• month = "19-21 " # may,
• publisher = "Springer",
• keywords = "genetic algorithms, genetic programming, Symbolic regression, Symbolic classification, Deep learning algebra networks",
• isbn13 = "978-981-16-8112-7",
• DOI = "doi:10.1007/978-981-16-8113-4_6",
• abstract = "Deep learning neural networks have produced some notable well publicized successes in several fields. Genetic Programming has also produced well publicized notable successes. Inspired by the deep learning successes with neural nets, we experiment with deep learning algebra networks where the network remains unchanged but where the neurons are replaced with general algebraic expressions. The training algorithms replace back propagation, counter propagation, etc. with a combination of genetic programming to generate the algebraic expressions and multiple regression, logit regression, and discriminant analysis to train the deep learning algebra network. These enhanced algebra networks are trained on ten theoretical classification problems with good performance advances which show a clear statistical performance improvement as network architecture is expanded.",
• notes = "Part of \cite{Banzhaf:2021:GPTP} published after the workshop in 2022",

}

Genetic Programming entries for Michael Korns

Citations