Publisher's Synopsis
Papers presented within this volume deal with architectures suitable for, and problems typical of, silicon implementation of neural nets. Analog solutions, as opposed to digital ones, are widely advocated as an attractive implementation, given compactness of the obtained architectures, low power consumption, and ease of emulating the analog behaviour of biological networks. Also, a "mixed-mode" solution, combining the analog/neural and the digital/logical processing aspects is advocated.
