Error-Corrective Recall of Digital Optical Images in Neural Networks Models by Photo-Burning of Spectral Holes,
Abstract
The neural-network-like scheme of data storage and processing of N-bit sequences of information needs a memory of about N(2) elements (interconnections). In digital auto-associative memories the useful data is presented usually as a set of S different words, v(s)(s=l,...,S), each word being a sequence of N bits. Simple mathematical rule, given originally by Hopfield, can serve as an algorithm to calculate the values of the N(2) elements of the memory matrix, T. Recall of the memory gives an output word, v(out), which results from a thresholded inner product between the interrogating input word, v(in), and the memory matrix. The readout procedure can be expressed mathematically, where TRH(...) stands for the thresholding procedure.
Document Details
- Document Type
- Technical Report
- Publication Date
- May 22, 1992
- Accession Number
- ADP008235
Entities
People
- Alexander Rebane
- Karl K. Rebane
- Olavi Ollikainen