作者: Mark Fanty
DOI: 10.1007/978-1-4613-0633-7_9
关键词: Very-large-scale integration 、 Artificial intelligence 、 Connectionism 、 Embodied cognition 、 Pattern matching 、 Computation 、 Simple (abstract algebra) 、 Connectionist network 、 Link weight 、 Computer science
摘要: Connectionist (or neural) networks consist of large numbers relatively simple computational units. The units communicate nonsymbolically over links. computation performed is embodied in the pattern connectivity. researchers hope to explain or reproduce human performance on matching and other tasks which have proved resistant traditional programming techniques. Roughly, they work designing artificial brains. Because needed — brain has more than 10**10 neurons useful for large, complex will require hardware realizations. Work VLSI ongoing [1].