Accession Number : ADA187707

Title :   The Choice of Learning Strategy in an Autonomous System.

Descriptive Note : Interim rept.,

Corporate Author : NAVAL RESEARCH LAB WASHINGTON DC

Personal Author(s) : Gordon, Diana F

PDF Url : ADA187707

Report Date : 05 Oct 1987

Pagination or Media Count : 30

Abstract : Machine learning is a relatively new field within artificial intelligence. Most of the work to date has concentrated on individual learning strategies. Autonomous systems such as robots, however, require a variety of techniques. This report addresses the problem of strategy choice within a multistrategy system. Learning techniques are first categorized along five dimensions. Afterwards, a description is given of potential sources of constraints on the choice of strategy. Finally, existing learning systems that use multiple strategies are critiqued. Keywords: Robotics.

Descriptors :   *LEARNING MACHINES, *ROBOTICS, ARTIFICIAL INTELLIGENCE, LEARNING, METHODOLOGY, ROBOTS, SOURCES, STRATEGY

Subject Categories : Human Factors Engineering & Man Machine System
      Cybernetics

Distribution Statement : APPROVED FOR PUBLIC RELEASE