Abstract
We propose a method to evolve adaptive behavior of learning in an artificial neural network (ANN). The adaptive behavior of learning emerges from the coordination of learning rules. Each learning rule is defined as a function of local information of a corresponding neuron only and modifies the connective strength between the neuron and its neighbors. It is exposed to selective pressure based on the fitness value, which represents the importance in producing the correct output signals of the ANN. The learning rules with lower fitness values are replaced by new ones generated by genetic programming techniques. Experimental results demonstrate that the proposed method produces adaptive behavior of learning in single-layered and two-layered ANNs. This means that efficient learning rules evolve. Further, the learning rules in the two-layered ANN coordinate with each other and macroscopic adaptive behavior of learning emerges.
Similar content being viewed by others
References
Koza JR, Rice JP (1991) Generation of both the weights and architecture for a neural network. IJCNN-91-Seattle, vol II, pp 397–404
Kitano H (1994) Neurogenetic learning: an integrated method of designing and training neural networks using genetic algorithms. Physica D 75:225–238
Bornholdt S, Graudenz D (1992) General asymmetric neural networks and structure design by genetic algorithms. Neural Networks 5:327–334
Maniezzo V (1994) Genetic evolution of the topology and weight distribution of neural networks. IEEE Trans on Neural Networks 5(1):39–53
Ackley D, Littman M (1991) Interactions between learning and evolution. Artificial Life II, SFI Studies in the Sciences of Complexity, vol X, Addison-Wesley, Reading, Mass
Chalmers DJ (1990) The evolution of learning: an experiment in genetic connectionism. Proceedings of the 1990 Connectionist Models Summer School, San Mateo
Koza JR (1994) Genetic programming II: automatic discovery of reusable programs. MIT Press, Cambridge, Mass
Matsunaga Y, Murase K, Yamakawa O et al. (1996) A modified back-propagation algorithm that automatically removes redundant hidden unit by competition (in Japanese). Trans IEICE, Part D-II, J79-D-II(3):403–412
Murao H, Kitamura S (1996) Evolution of locally defined learning rules and their coodination in feedforward neural networks. In: Sugisaka M (ed) Proceedings of the International Symposium on Artificial Life and Robotics (AROB1), Beppu, Oita, Japan, Feb 18–20, pp 183–186
Author information
Authors and Affiliations
Corresponding author
About this article
Cite this article
Murao, H., Kitamura, S. Evolution of locally defined learning rules and their coordination in feedforward neural networks. Artificial Life and Robotics 1, 89–94 (1997). https://doi.org/10.1007/BF02471120
Received:
Accepted:
Issue Date:
DOI: https://doi.org/10.1007/BF02471120