Using genetic search to exploit the emergent behavior of neural networks
CNLS '89 Proceedings of the ninth annual international conference of the Center for Nonlinear Studies on Self-organizing, Collective, and Cooperative Phenomena in Natural and Artificial Computing Networks on Emergent computation
Proceedings of the third international conference on Genetic algorithms
Efficient reinforcement learning through symbiotic evolution
Machine Learning - Special issue on reinforcement learning
Towards designing artificial neural networks by evolution
Applied Mathematics and Computation - Special issue on articficial life and robotics
A Coevolutionary Approach to Learning Sequential Decision Rules
Proceedings of the 6th International Conference on Genetic Algorithms
Evolutionary Artificial Neural Network Design and Training for wood veneer classification
Engineering Applications of Artificial Intelligence
Hi-index | 0.00 |
We describe an efficient method of combining theglobal search of genetic algorithms (GAs) with thelocal search of gradient descent algorithms. Eachtechnique optimizes a mutually exclusive subset of thenetwork‘s weight parameters. The GA chromosome fixesthe feature detectors and their location, and agradient descent algorithm starting from randominitial values optimizes the remaining weights. Threealgorithms having different methods of encoding hiddenunit weights in the chromosome are applied tomultilayer perceptrons (MLPs) which classify noisydigital images. The fitness function measures the MLPclassification accuracy together with the confidenceof the networks.