SIGGRAPH '94 Proceedings of the 21st annual conference on Computer graphics and interactive techniques
Understanding nonlinear dynamics
Understanding nonlinear dynamics
Swarm intelligence
Evolving neural networks through augmenting topologies
Evolutionary Computation
A Neural Network for Creative Serial Order Cognitive Behavior
Minds and Machines
No free lunch theorems for optimization
IEEE Transactions on Evolutionary Computation
| Hi-index | 0.00 |
Significant emphasis has been placed on the use of genetic algorithms for evolving solutions to otherwise intractable problems but they also offer an opportunity to explore the process of evolution itself. Brains represent complex circuits that have arisen in an astronomically large search space of possibilities to solve a variety of problem types, most of them quite complex. To expedite the investigation of the evolution of neural circuitry, we introduce a system that makes it possible to (1) explore the constraints under which various brain functions might have evolved, (2) demonstrate (as a proof of concept) that such evolution is possible, and (3) that lays the foundation for tracking paths from simpler to more complex architectures. The system we present uses a highly flexible genome encoding scheme in conjunction with a sophisticated genetic algorithm (employing multiple operators) and a flexible neural network simulation program. We show preliminary results for a basic neural circuit and discuss implications for subsequent work.