据1992年2月在美国旧金山召开的第39届国际固体电路会议报道,已有4种高水平的超大规模集成神经网络研制成功。美国南加州大学用2微米 CMOS工艺制成用于自组织映射的模拟神经网络处理器,它是一种能够学习的、具有64个输出神经元和1600个突触的模拟神经芯片,采用一种电流型方法来并行地 According to the 39th International Solid State Circuits Conference held in San Francisco in February 1992, four high-level VLSI neural networks have been successfully developed. The University of Southern California, USA, uses a 2-micron CMOS process to create a simulated neural network processor for self-organizing maps. It is a learnable analog neurochip with 64 output neurons and 1600 synapses, Current-type methods come in parallel