Title
A Neural Network Integrated Circuit Supporting Programmable Exponent And Mantissa
Abstract
A neural network architecture is presented utilizing small (8-b) word sizes for data and weights while maintaining flexibility in the dynamic range achieved by using adjustable exponent and mantissa field. Design issues regarding tradeoff between small and large precision number systems are reviewed. Simulation results on network performance are included. In small precision systems (8-b), a fixed dynamic may limit the system capability. For certain applications it is desirable to have a flexible dynamic range implementable with a variable size of exponent and mantissa.
Publication Date
12-1-1990
Publication Title
Proceedings of the Custom Integrated Circuits Conference
Document Type
Article; Proceedings Paper
Personal Identifier
scopus
Copyright Status
Unknown
Socpus ID
0025626655 (Scopus)
Source API URL
https://api.elsevier.com/content/abstract/scopus_id/0025626655
STARS Citation
Brown, H. K.; Cross, D. D.; and Wuerz, D. L., "A Neural Network Integrated Circuit Supporting Programmable Exponent And Mantissa" (1990). Scopus Export 1990s. 1452.
https://stars.library.ucf.edu/scopus1990/1452