作者: Stefan Prange , Heinrich Klar
DOI: 10.1016/B978-0-444-89958-3.50015-2
关键词:
摘要: ABSTRACT On today's digital computers every problem the solution of which is possible to be described by an algorithm can solved. An a series instructions performed one after other. However, there are problems for no could found up now. These tasks especially lie in area adaptive signal processing, recognition and classification noisy patterns, as well sensory motor control. Inspired power living beings solving these new approach was considered, called neural networks. Neural networks massive parallel interconnection simple processing elements (5) . Today's have already proved their possibilities e.g. handwritten characters or associative storage patterns. used today do not much common with biological neuron. Biology-oriented attempt keep more closely behavior hand biology-oriented offer dynamic processing. other they become powerful tool neurophysiological research. The is, that research on systems such simulation serial computer results unacceptable computing times. There need specific hardware perform network modelling emulation instead (4) Based description development integrated circuit networks, future will derived.