Genetic programming (GP) can learn complex concepts by
searching for the target concept through evolution of a
population of candidate hypothesis programs. However,
unlike some learning techniques, such as Artificial
Neural Networks (ANNs), GP does not have a principled
procedure for changing parts of a learned structure
based on that structure's performance on the training
data. GP is missing a clear, locally optimal update
procedure, the equivalent of gradient-descent
backpropagation for ANNs. This article introduces a new
algorithm, "internal reinforcement", for defining
and using performance feedback on program evolution.
This internal reinforcement principled mechanism is
developed within a new connectionist representation for
evolving parameterized programs, namely "neural
programming". We present the algorithms for the
generation of credit and blame assignment in the
process of learning programs using neural programming
and internal reinforcement. The article includes a
comprehensive overview of genetic programming and
empirical experiments that demonstrate the increased
learning rate obtained by using our principled program
evolution approach.
%0 Journal Article
%1 Teller:2000:AI
%A Teller, Astro
%A Veloso, Manuela
%D 2000
%J Artificial Intelligence
%K Bucket Evolutionary Internal Machine Neural Signal algorithms, brigade computation, genetic learning, programming, reinforcement, understanding,
%N 2
%P 165--198
%R doi:10.1016/S0004-3702(00)00023-0
%T Internal reinforcement in a connectionist genetic
programming approach
%U http://www.sciencedirect.com/science/article/B6TYF-40TY77M-1/1/c54fc0ab842b831a76c9e61e1c1c6b85
%V 120
%X Genetic programming (GP) can learn complex concepts by
searching for the target concept through evolution of a
population of candidate hypothesis programs. However,
unlike some learning techniques, such as Artificial
Neural Networks (ANNs), GP does not have a principled
procedure for changing parts of a learned structure
based on that structure's performance on the training
data. GP is missing a clear, locally optimal update
procedure, the equivalent of gradient-descent
backpropagation for ANNs. This article introduces a new
algorithm, "internal reinforcement", for defining
and using performance feedback on program evolution.
This internal reinforcement principled mechanism is
developed within a new connectionist representation for
evolving parameterized programs, namely "neural
programming". We present the algorithms for the
generation of credit and blame assignment in the
process of learning programs using neural programming
and internal reinforcement. The article includes a
comprehensive overview of genetic programming and
empirical experiments that demonstrate the increased
learning rate obtained by using our principled program
evolution approach.
@article{Teller:2000:AI,
abstract = {Genetic programming (GP) can learn complex concepts by
searching for the target concept through evolution of a
population of candidate hypothesis programs. However,
unlike some learning techniques, such as Artificial
Neural Networks (ANNs), GP does not have a principled
procedure for changing parts of a learned structure
based on that structure's performance on the training
data. GP is missing a clear, locally optimal update
procedure, the equivalent of gradient-descent
backpropagation for ANNs. This article introduces a new
algorithm, {"}internal reinforcement{"}, for defining
and using performance feedback on program evolution.
This internal reinforcement principled mechanism is
developed within a new connectionist representation for
evolving parameterized programs, namely {"}neural
programming{"}. We present the algorithms for the
generation of credit and blame assignment in the
process of learning programs using neural programming
and internal reinforcement. The article includes a
comprehensive overview of genetic programming and
empirical experiments that demonstrate the increased
learning rate obtained by using our principled program
evolution approach.},
added-at = {2008-06-19T17:46:40.000+0200},
author = {Teller, Astro and Veloso, Manuela},
biburl = {https://www.bibsonomy.org/bibtex/24971dbcd2080d490c2d30767cd9b700a/brazovayeye},
doi = {doi:10.1016/S0004-3702(00)00023-0},
interhash = {539270f24e78365e732873a11f1da68c},
intrahash = {4971dbcd2080d490c2d30767cd9b700a},
journal = {Artificial Intelligence},
keywords = {Bucket Evolutionary Internal Machine Neural Signal algorithms, brigade computation, genetic learning, programming, reinforcement, understanding,},
month = {July},
notes = {oai:CiteSeerPSU:558697
http://citeseer.ist.psu.edu/558697.html gives a
slightly different version},
number = 2,
pages = {165--198},
size = {34 pages},
timestamp = {2008-06-19T17:53:02.000+0200},
title = {Internal reinforcement in a connectionist genetic
programming approach},
url = {http://www.sciencedirect.com/science/article/B6TYF-40TY77M-1/1/c54fc0ab842b831a76c9e61e1c1c6b85},
volume = 120,
year = 2000
}