On neural network hardware and programming paradigms
暂无分享,去创建一个
Implementation guidelines and performance benchmarks for dedicated neural network hardware and software are proposed. An example of a dedicated neural network processor and the implementation of a popular neural network architecture and learning algorithm on this hardware is presented, emphasizing the specific architecture and programming paradigm. Finally, the performance of the implementation is benchmarked in a single-processor and multiple-processor environment against a supercomputer. Such implementations are aimed for neural network application requiring rapid online learning.
[1] Simon Haykin,et al. Neural Networks: A Comprehensive Foundation , 1998 .
[2] E. van Keulen,et al. Neural network hardware performance criteria , 1994, Proceedings of 1994 IEEE International Conference on Neural Networks (ICNN'94).
[3] Donald J. Patterson,et al. Computer organization and design: the hardware-software interface (appendix a , 1993 .
[4] Karl Mathia,et al. Solutions of linear equations and a class of nonlinear equations using recurrent neural networks , 1996 .