A GENERAL NEURAL NET
Abstract
A neural net is discussed which is general in that (1) connectivity and transmission times along links are at the pleasure of a net's constructer, (2) time is not quantized (except as demanded by the grain of the computation), (3) eight of the synaptic mechanisms distinguished by Bullock are permitted, and (4) the assisted synaptic learning theory of Milner is included. The net has been coded for the IBM 7090. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 01, 1962
- Accession Number
- AD0291805
Entities
People
- E.w. Paxson
- J.w. Smith
Organizations
- RAND Corporation