TY - JOUR
T1 - Incremental growth in modular neural networks
AU - MacLeod, Christopher
AU - Maxwell, Grant
AU - Muthuraman, Sethuraman
PY - 2009/6
Y1 - 2009/6
N2 - This paper outlines an algorithm for incrementally growing Artificial Neural Networks. The algorithm allows the network to expand by adding new sub-networks or modules to an existing structure; the modules are trained using an Evolutionary Algorithm. Only the latest module added to the network is trained, the previous structure remains fixed. The algorithm allows information from different data domains to be integrated into the network and because the search space in each iteration is small, large and complex networks with a modular structure can emerge naturally. The paper describes an application of the algorithm to a legged robot and discusses its biological inspiration.
AB - This paper outlines an algorithm for incrementally growing Artificial Neural Networks. The algorithm allows the network to expand by adding new sub-networks or modules to an existing structure; the modules are trained using an Evolutionary Algorithm. Only the latest module added to the network is trained, the previous structure remains fixed. The algorithm allows information from different data domains to be integrated into the network and because the search space in each iteration is small, large and complex networks with a modular structure can emerge naturally. The paper describes an application of the algorithm to a legged robot and discusses its biological inspiration.
KW - Evolutionary algorithm
KW - Genetic algorithm
KW - Incremental evolution
KW - Incremental growth
KW - Modular artificial neural network
KW - Robotics
UR - http://www.scopus.com/inward/record.url?scp=67349164300&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=67349164300&partnerID=8YFLogxK
U2 - 10.1016/j.engappai.2008.11.002
DO - 10.1016/j.engappai.2008.11.002
M3 - Article
AN - SCOPUS:67349164300
SN - 0952-1976
VL - 22
SP - 660
EP - 666
JO - Engineering Applications of Artificial Intelligence
JF - Engineering Applications of Artificial Intelligence
IS - 4-5
ER -