Object oriented VLSI design automation for pulse coded neural networks
暂无分享,去创建一个
Pulse coded neural networks employ special purpose mixed signal circuits which emulate to varying degrees the temporal behavior of natural neurons. It has long been difficult to verify pulse coded neural network implementations of significant complexity prior to fabrication using conventional design automation tools. In this paper, we introduce SPINACH, an objected oriented solution to the problems of pulse-coded neural network hardware simulation and synthesis. SPINACH is a C++ object library intended for the specification and simulation of pulse-coded neural network hardware. As such, it fulfils the purpose of a specialized mixed-signal hardware description language (HDL) complete with both large-scale simulation capabilities and also circuit synthesis when combined with a commercial VLSI design system.<<ETX>>
[1] Alan F. Murray,et al. Pulse-stream VLSI neural networks mixing analog and digital techniques , 1991, IEEE Trans. Neural Networks.
[2] Jack L. Meador,et al. Competitive learning in asynchronous-pulse-density integrated circuits , 1992 .
[3] J. Meador. Finite precision effects on dynamic behavior in a pulse-coded winner-take-all mechanism , 1992, [Proceedings 1992] IJCNN International Joint Conference on Neural Networks.
[4] Alan F. Murray,et al. Pulse stream VLSI neural networks , 1994, IEEE Micro.