TY - GEN
T1 - Efficient natural evolution strategies
AU - Sun, Yi
AU - Wierstra, Daan
AU - Schaul, Tom
AU - Schmidhuber, Juergen
N1 - Generated from Scopus record by KAUST IRTS on 2022-09-14
PY - 2009/12/31
Y1 - 2009/12/31
N2 - Efficient Natural Evolution Strategies (eNES) is a novel alternative to conventional evolutionary algorithms, using the natural gradient to adapt the mutation distribution. Unlike previous methods based on natural gradients, eNES uses a fast algorithm to calculate the inverse of the exact Fisher information matrix, thus increasing both robustness and performance of its evolution gradient estimation, even in higher dimensions. Additional novel aspects of eNES include optimal fitness baselines and importance mixing (a procedure for updating the population with very few fitness evaluations). The algorithm yields competitive results on both unimodal and multimodal benchmarks. Copyright 2009 ACM.
AB - Efficient Natural Evolution Strategies (eNES) is a novel alternative to conventional evolutionary algorithms, using the natural gradient to adapt the mutation distribution. Unlike previous methods based on natural gradients, eNES uses a fast algorithm to calculate the inverse of the exact Fisher information matrix, thus increasing both robustness and performance of its evolution gradient estimation, even in higher dimensions. Additional novel aspects of eNES include optimal fitness baselines and importance mixing (a procedure for updating the population with very few fitness evaluations). The algorithm yields competitive results on both unimodal and multimodal benchmarks. Copyright 2009 ACM.
UR - http://portal.acm.org/citation.cfm?doid=1569901.1569976
UR - http://www.scopus.com/inward/record.url?scp=72749090128&partnerID=8YFLogxK
U2 - 10.1145/1569901.1569976
DO - 10.1145/1569901.1569976
M3 - Conference contribution
SN - 9781605583259
SP - 539
EP - 545
BT - Proceedings of the 11th Annual Genetic and Evolutionary Computation Conference, GECCO-2009
ER -