Probabilistic Incremental Program Evolution:
Stochastic Search Through Program Space
R. Salustowicz, and J. Schmidhuber. Machine Learning: ECML-97, volume 1224 of Lecture Notes in Artificial Intelligence, page 213--220. Springer-Verlag, (1997)
Abstract
Probabilistic Incremental Program Evolution (PIPE) is
a novel technique for automatic program synthesis. We
combine probability vector coding of program
instructions Schmidhuber, 1997, Population-Based
Incremental Learning (PBIL) Baluja and Caruana, 1995
and tree-coding of programs used in variants of Genetic
Programming (GP) icga85:cramer ;
koza:book . PIPE uses a stochastic selection
method for successively generating better and better
programs according to an adaptive ``probabilistic
prototype tree''. No crossover operator is used. We
compare PIPE to Koza's GP variant on a function
regression problem and the 6-bit parity problem.
%0 Conference Paper
%1 Salustowicz:97ecml
%A Salustowicz, R. P.
%A Schmidhuber, J.
%B Machine Learning: ECML-97
%D 1997
%E van Someren, M.
%E Widmer, G.
%I Springer-Verlag
%K Incremental Learning, Population-Based Program Search Stochastic algorithms, genetic programming,
%P 213--220
%T Probabilistic Incremental Program Evolution:
Stochastic Search Through Program Space
%U ftp://ftp.idsia.ch/pub/rafal/ECML_PIPE.ps.gz
%V 1224
%X Probabilistic Incremental Program Evolution (PIPE) is
a novel technique for automatic program synthesis. We
combine probability vector coding of program
instructions Schmidhuber, 1997, Population-Based
Incremental Learning (PBIL) Baluja and Caruana, 1995
and tree-coding of programs used in variants of Genetic
Programming (GP) icga85:cramer ;
koza:book . PIPE uses a stochastic selection
method for successively generating better and better
programs according to an adaptive ``probabilistic
prototype tree''. No crossover operator is used. We
compare PIPE to Koza's GP variant on a function
regression problem and the 6-bit parity problem.
@inproceedings{Salustowicz:97ecml,
abstract = {Probabilistic Incremental Program Evolution (PIPE) is
a novel technique for automatic program synthesis. We
combine probability vector coding of program
instructions [Schmidhuber, 1997], Population-Based
Incremental Learning (PBIL) [Baluja and Caruana, 1995]
and tree-coding of programs used in variants of Genetic
Programming (GP) [ \cite{icga85:cramer} ;
\cite{koza:book} ]. PIPE uses a stochastic selection
method for successively generating better and better
programs according to an adaptive ``probabilistic
prototype tree''. No crossover operator is used. We
compare PIPE to Koza's GP variant on a function
regression problem and the 6-bit parity problem.},
added-at = {2008-06-19T17:46:40.000+0200},
author = {Salustowicz, R. P. and Schmidhuber, J.},
biburl = {https://www.bibsonomy.org/bibtex/261c547a179958a8d0b58833b4e1c0d8a/brazovayeye},
booktitle = {Machine Learning: ECML-97},
editor = {van Someren, M. and Widmer, G.},
interhash = {83a8cecbc3fa9f5de0bb5e5e65dd28a4},
intrahash = {61c547a179958a8d0b58833b4e1c0d8a},
keywords = {Incremental Learning, Population-Based Program Search Stochastic algorithms, genetic programming,},
notes = {ECML-97},
pages = {213--220},
publisher = {Springer-Verlag},
publisher_address = {Berlin},
series = {Lecture Notes in Artificial Intelligence},
size = {9 pages},
timestamp = {2008-06-19T17:50:58.000+0200},
title = {Probabilistic Incremental Program Evolution:
Stochastic Search Through Program Space},
url = {ftp://ftp.idsia.ch/pub/rafal/ECML_PIPE.ps.gz},
volume = 1224,
year = 1997
}