SEEK-TIME IMPROVEMENT IN A RANDOM-ACCESS FILE BY APPLICATION OF AN ADAPTIVE ELEMENT

Abstract

An Adaline (adaptive linear neuron) can be trained to distinguish between sets of inputs. In general, the quantized output is used. This report investigates the usefulness of the analog output of Adaline for measuring the frequency of occurrence of a number of different events. Each event is more or less arbitrarily associated with a pattern and it is shown that the degree to which Adaline has been trained to recognize any one of these patterns can be used as a measure of the frequency of occurrence of the associated event. The application of this use of Adaline to a random-access file is simulated in order to show its use in reducing average access time.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 01, 1963
Accession Number
AD0436020

Entities

People

  • W. S. Buslik

Organizations

  • Stanford University

Tags

Communities of Interest

  • Advanced Electronics
  • Ground and Sea Platforms
  • Materials and Manufacturing Processes
  • Weapons Technologies

DTIC Thesaurus Topics

  • Access Time
  • Accuracy
  • Air Force
  • California
  • Diagrams
  • Electromagnetic Wave Propagation
  • Electron Tubes
  • Electronics
  • Electronics Laboratories
  • Engineering
  • Frequency
  • Military Research
  • National Security
  • Reliability
  • Semiconductors
  • Simulations
  • United States

Readers

  • Computer Science.
  • Neural Network Machine Learning.
  • Radio communications and signal processing.