Abstract
A wave-parallel computing (WPC) technique is proposed to address the interconnection problem in massively interconnected neural architectures. The fundamental concepts are the multiplexing of several signals onto a single line, and their wave-parallel processing without decomposition. An application of WPC to neural networks is discussed, with emphasis on the reduction in the number of interconnections.
Original language | English |
---|---|
Pages (from-to) | 690-691 |
Number of pages | 2 |
Journal | Electronics Letters |
Volume | 33 |
Issue number | 8 |
DOIs | |
Publication status | Published - 1997 Apr 10 |
Keywords
- Computer architecture
- Neural networks