Just wondering, since we've reached 1 teraflop per PC, yet we are still not able to model an insect's brain. Has anyone seen a decent implementation of a self-learning, self-developing neural network?
You can refer to this paper that presents a prototype of a self-developing neural network architecture, named Neuromorphic Autonomous Controller (NAC). Designed to control a discrete event feedback system, this architecture is confirmed by layers of lined up modules, each module encloses a neural network. Each neural network module dynamically creates its structure and permanently adapts to changing the environment without supervision. the previous training is not required: during processing, learning and controlling are performed simultaneously.