Evolution and learning in neural networks: the number and distri

Details Evolution and learning in neural networks: the number and distri Evolution and learning in neural networks: the number and distri

1990-11-21

1993-09-14

MacDonald, Allen R.

395 23, 395 24, G06F 1518

Patent

active

052456960

ABSTRACT:
The present invention relates to the interrelationships between nature (as mediated by evolution and genetic algorithms) and nurture (as mediated by gradient-descent supervised learning) in a population of neural networks for pattern recognition. The Baldwin effect is demonstrated that learning can change the rate of evolution of the population's genome - a "pseudo-Lamarkian" process, in which information learned is ultimately encoded in the genome by a purely Darwinian process. Selectivity is shown for this effect: too much learning or too little learning in each generation leads to slow evolution of the genome, whereas an intermediate amount leads to most rapid evolution. For a given number of learning trials throughout a population, the most rapid evolution occurs if different individuals each receive a different number of learning trials, rather than the same number. Because all biological networks possess structure due to evolution, it is important that such interactions between learning and evolution be understood. Hybrid systems can take advantage both of gradient descents (learning) and large jumps (genetic algorithms) in very complicated energy landscapes and hence may play an increasingly important role in the design of artificial neural systems.

REFERENCES:
patent: 4935877 (1990-06-01), Koza
patent: 5048095 (1991-09-01), Bhanu et al.
A Darwinian Approach to Artificial Neural Networks; Dress et al.; Proc. of System, Man, & Cybernetics Conf; 1987; pp. 572-577.
Towards the Evolution of Symbols; Dolan et al.; Proc of the Second Inter. Conf. on Genetic Algorithms; Jul. 28-31, 1987; pp. 123-131.
How Learning Can Guide Evolution; Hinton and Nowlan; Complex Systems 1 (1987) 495-502.
Symbolic Schemata in Connectionist Memories: Role Binding and the Evolution of Structure; Dolan and Dyer; UCLA-AI-87-11; pp. 1-23; 1987.
Designing Neural Networks Using Genetic Algorithms; Miller, Todd, and Hegde; ICGA '89; George Mason University, Fairfax, Va.; pp. 379-383; Jun. 4-7, 1989.
Towards the Genetic Sysntesis of Neural Networks; Harp, Samad, and Guha; ICGA '89; George Mason University, Fairfax, Va.; pp. 360-369; Jun. 4-7, 1989.
Using Genetic Search to Exploit the Emergent Behavior of Neural Networks; Schaffer, Caruana, and Eshelman; Physica D 42 (1990) 244-248.

Affiliated with

Also associated with

Rating

Evolution and learning in neural networks: the number and distri does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Evolution and learning in neural networks: the number and distri, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Evolution and learning in neural networks: the number and distri will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2034538