Loading...
Please wait, while we are loading the content...
Similar Documents
Genetic Algorithms for Genetic Neural Nets
| Content Provider | Semantic Scholar |
|---|---|
| Author | Sharp, David H. Reinitzt, John Mjolsnesst, Eric |
| Abstract | In contrast to most synthetic neural nets, biological neural networks have a strong component of genetic determination which acts before and during experiential learning. Three broad levels of phenomena are present: long-term evolution, involving crossover as well as point mutation; a developmental process mapping genetic information to a set of cells and their internal states of gene expression (genotype to phenotype): and the subsequent synaptogenesis. We describe a very simple mathematical idealization of these three levels which combines the crossover search method of genetic algorithms with the developmental models used in our previous work on "genetic" or 'recursively generated" artificial neural nets [18] (and elaborated into a connectionist model of biological development [19]). Despite incorporating all three levels (evolution on genes; development of cells; synapse formation) the model may actually be far cheaper to compute with than a comparable search directly in synaptic weight space. 1\Vork supported by the United States Department of Energy. 2Supported in part by grant 5-T15-LM07056 from the National Library of Medicine. 3Supported in part by the Air Force Office of Scientific Research under grant AFOSR 88-0240. Genetic Algorithms for Genetic Nets 2 1 GENES, CELLS, AND NETWORKS Biology has motivated research on genetic algorithms as well as on synthetic neural nets. a What biological phenomena could motivate a synthesis of these two approaches? We propose that such a synthesis can be based on the interplay of objects at three levels of organization: genes, cells, and networks of cells. Dynamics at the level of genes (point mutations, crossover, inversion and so forth) has been schematized for use in genetic algorithms. Likewise the dynamics of networks of neurons, including learning, are abstracted in synthetic neural nets. The missing dynamical system is development: the elaboration of genetic information to produce neurons and their initial connections. A phenomenological modelling framework for development was presented in [19]. A simplified version of this model, together with simple genetic algorithms and neural nets, results in a synthesis of these ideas in a unified connectionist model. |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://www.dtic.mil/get-tr-doc/pdf?AD=ADA256223 |
| Language | English |
| Access Restriction | Open |
| Subject Keyword | Artificial neural network Biologic Development Biological Phenomena Cancer Center Support Grant Connectionism Crossover (genetic algorithm) Developmental process Dynamical system Gene Expression Genetic algorithm Integrative level Mathematics Neural Network Simulation Neural oscillation Physical object Point Mutation Recursion SyNAPSE Synaptic Package Manager Synaptic weight Synthetic intelligence synaptogenesis |
| Content Type | Text |
| Resource Type | Article |
