Fourier Holography Techniques for Artificial Intelligence

Pavlov, Alexander V.

doi:10.1117/3.793309.ch13

Chapter 13:
Fourier Holography Techniques for Artificial Intelligence

Book: Advances in Information Optics and Photonics

Editor(s): Ari T. Friberg; René Dändliker

Author(s): Pavlov, Alexander V.

Published: 2008
https://doi.org/10.1117/3.793309.ch13

Abstract

In the days immediately after the reinvention of holography by Leith and Upatnieks, and independently by Denisyuk, new applications of holography were being proposed. One of those was based on a number of deep analogies between the attributes of the human mind and brain, and also the properties of optical holography. Based on these analogies, the holographic paradigm was proposed and developed in cognitive science by Pribram. Unfortunately, the term âholographic paradigmâ was recently compromised regarding the so-called âholographic brain,â âholographic universe,â etc. in many near-scientific publications. Many of these papers are based on superficial knowledge of the outer effects of artistic holography and hold general conversations instead on the detailed analysis of particular physical phenomena and mathematical models. Thus, first of all it has to be underlined that there is no intention to maintain that the brain is a hologramâour aim is to analyze how holography can be utilized in the framework of artificial intelligence. Any anthropomorphisms andâor human analogs used in this chapter are for intuitive purposes; that is, to make the chapter more vivid and readable. The first analogy between the brain and holography was the associative properties of both biological brain and optical hologram, which have been appreciated since the inception of holography. Based on this analogy, a number of models for holographic associative memory were proposed and implemented. The list of articles dedicated to holographic associative memories includes a few hundreds papers. The concept of associative memory has been developed as a paradigm for information processing by the brain, and incorporated in the paradigm of artificial neural networks. Associative memory is defined as fault-tolerant, content-addressable memory, which recalls a noise-free reference pattern when addressed by an erased andâor distorted input pattern. Associative memory plays a major role in thinking and forms a base of intelligence. The concept of associative memory is closely linked with another attribute of the brain and mind, namely, the brain thinks not by operations on numbers, but by processing of the patterns of neural activity, also known as internal representation patterns. This attribute leads to the second analogy between brain and holography, i.e., both process information by patterns processing.

Online access to SPIE eBooks is limited to subscribing institutions.

CHAPTER 13
20 PAGES

DOWNLOAD PDF + SAVE TO MY LIBRARY