Compositional pattern producing networks: A novel
abstraction of development
K. Stanley. Genetic Programming and Evolvable Machines, 8 (2):
131--162(Juni 2007)Special issue on developmental systems.
DOI: doi:10.1007/s10710-007-9028-8
Zusammenfassung
Natural DNA can encode complexity on an enormous
scale. Researchers are attempting to achieve the same
representational efficiency in computers by
implementing developmental encodings, i.e. encodings
that map the genotype to the phenotype through a
process of growth from a small starting point to a
mature form. A major challenge in in this effort is to
find the right level of abstraction of biological
development to capture its essential properties without
introducing unnecessary inefficiencies. In this paper,
a novel abstraction of natural development, called
Compositional Pattern Producing Networks (CPPNs), is
proposed. Unlike currently accepted abstractions such
as iterative rewrite systems and cellular growth
simulations, CPPNs map to the phenotype without local
interaction, that is, each individual component of the
phenotype is determined independently of every other
component. Results produced with CPPNs through
interactive evolution of two dimensional images show
that such an encoding can nevertheless produce
structural motifs often attributed to more conventional
developmental abstractions, suggesting that local
interaction may not be essential to the desirable
properties of natural encoding in the way that is
usually assumed.
%0 Journal Article
%1 Stanley:2007:GPEM
%A Stanley, Kenneth O.
%D 2007
%J Genetic Programming and Evolvable Machines
%K CPPN,Evolutionary systems,
%N 2
%P 131--162
%R doi:10.1007/s10710-007-9028-8
%T Compositional pattern producing networks: A novel
abstraction of development
%V 8
%X Natural DNA can encode complexity on an enormous
scale. Researchers are attempting to achieve the same
representational efficiency in computers by
implementing developmental encodings, i.e. encodings
that map the genotype to the phenotype through a
process of growth from a small starting point to a
mature form. A major challenge in in this effort is to
find the right level of abstraction of biological
development to capture its essential properties without
introducing unnecessary inefficiencies. In this paper,
a novel abstraction of natural development, called
Compositional Pattern Producing Networks (CPPNs), is
proposed. Unlike currently accepted abstractions such
as iterative rewrite systems and cellular growth
simulations, CPPNs map to the phenotype without local
interaction, that is, each individual component of the
phenotype is determined independently of every other
component. Results produced with CPPNs through
interactive evolution of two dimensional images show
that such an encoding can nevertheless produce
structural motifs often attributed to more conventional
developmental abstractions, suggesting that local
interaction may not be essential to the desirable
properties of natural encoding in the way that is
usually assumed.
@article{Stanley:2007:GPEM,
abstract = {Natural DNA can encode complexity on an enormous
scale. Researchers are attempting to achieve the same
representational efficiency in computers by
implementing developmental encodings, i.e. encodings
that map the genotype to the phenotype through a
process of growth from a small starting point to a
mature form. A major challenge in in this effort is to
find the right level of abstraction of biological
development to capture its essential properties without
introducing unnecessary inefficiencies. In this paper,
a novel abstraction of natural development, called
Compositional Pattern Producing Networks (CPPNs), is
proposed. Unlike currently accepted abstractions such
as iterative rewrite systems and cellular growth
simulations, CPPNs map to the phenotype without local
interaction, that is, each individual component of the
phenotype is determined independently of every other
component. Results produced with CPPNs through
interactive evolution of two dimensional images show
that such an encoding can nevertheless produce
structural motifs often attributed to more conventional
developmental abstractions, suggesting that local
interaction may not be essential to the desirable
properties of natural encoding in the way that is
usually assumed.},
added-at = {2013-09-10T07:11:01.000+0200},
author = {Stanley, Kenneth O.},
biburl = {https://www.bibsonomy.org/bibtex/2d13939ab4a0186add6bb55856534babc/andrewmatthews},
doi = {doi:10.1007/s10710-007-9028-8},
interhash = {5d8d4da4a8c4834e3cebc9eb55cedb73},
intrahash = {d13939ab4a0186add6bb55856534babc},
issn = {1389-2576},
journal = {Genetic Programming and Evolvable Machines},
keywords = {CPPN,Evolutionary systems,},
month = {June},
note = {Special issue on developmental systems},
number = 2,
pages = {131--162},
size = {32 pages},
timestamp = {2013-09-10T07:11:01.000+0200},
title = {Compositional pattern producing networks: {A} novel
abstraction of development},
volume = 8,
year = 2007
}