Proceedings of 1994 28th Asilomar Conference on Signals, Systems and Computers (1994)

Pacific Grove, CA, USA

Oct. 31, 1994 to Nov. 2, 1994

ISSN: 1058-6393

ISBN: 0-8186-6405-3

pp: 536-539

S. Vassiliadis , Dept. of Electr. Eng., Delft Univ. of Technol., Netherlands

K. Bertels , Dept. of Electr. Eng., Delft Univ. of Technol., Netherlands

ABSTRACT

We investigate small depth and size feed-forward neural networks performing binary addition. We propose a set of equations that can be used to realize small depth inexpensive networks for arbitrary operand lengths. In particular we show that O(n) depth-3 networks for the binary addition can be easily constructed having small weight sizes. We also describe a scheme for the design of 32-bit binary adders. When compared to the addition scheme known to produce the least expensive adders on terms of area, using feed-forward neural networks, our scheme requires only 20% of the area in terms of neurons. Consequently our design provides substantial area reduction.<>

INDEX TERMS

adders, digital arithmetic, feedforward neural nets, multilayer perceptrons

CITATION

S. Vassiliadis and K. Bertels, "O(n) depth-3 binary addition,"

*Proceedings of 1994 28th Asilomar Conference on Signals, Systems and Computers(ACSSC)*, Pacific Grove, CA, USA, 1995, pp. 536-539.

doi:10.1109/ACSSC.1994.471510

CITATIONS