Self-learning neuronal network in a hybrid VLSI technology for monitoring learning patterns and controlling learning processes adjusts automatically to learning patterns.

摘要

An automatism allows independent adjustment of a neuronal network. Learning patterns in a memory store are made available to a neuro-chip. Integrated on the chip is a micro-core for monitoring learning patterns and controlling learning processes. Collecting and storing representative data in a memory store is enough to solve nonlinear classification problems.