Lengths of Cycle Times in Random Neural Networks

摘要

The report contains a feasibility study of timing circuits for a model of the human brain. It is shown that certain random neural networks have cycle times which increase exponentially with the size of the network, and that these networks may be used as timing devices even for a period of time equal to a human lifetime. The methods used in this report are those of probability, combinatorial analysis, and number theory. (Author)