How do the principles of psychology relate to the activities of 1011 nerve cells in the human brain? Large physical systems generally have emergent collective behaviours. Real computers—the brain included—compute by following a change of computer ‘state’ with time. When that change of state can be described by collective variables, robust dynamics emerge whose mathematical description can be entirely unlike that of the underlying microscopic variables. We illustrate the useful emergent computational dynamics of two elementary neural networks. The first deals with the problem of variable cadence when recognizing dynamical sensory patterns spread over time. The second achieves goal-directed behaviour when multiple candidate goals are simultaneously present.
Part of a symposium to celebrate the work of Professor Sir David MacKay FRS. The symposium was held over the period 14-15 March 2016.
[ Ссылка ]
