Neural Networks that Create their Own Goals Using Growth Cycles.

Abstract

Over the past year our research on the generation and control of path planning has formed a new theory using neural networks that allows intelligent agents to automate sub-tasking in achieving desired goals. This theory called Growth Cycles provides functionality that will be crucial at the next stage of computing and communication systems with their exponentially increasing overhead. Distributed intelligent agents will begin to autonomously and adaptively maintain the huge and complex National Information Infrastructure. The specific research shows how an intelligent agent can learn to go from anywhere to anywhere around obstacles in novel and contingent environments. Although the research focuses on adaptive path planning, it can also be generalized and applied to adaptive and autonomous problem solving. The following paper details the entire effort for this contract and has been submitted for publication. (KAZR) p. 1

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 02, 1994
Accession Number
ADA298853

Entities

People

  • Michael Kuperstein

Tags

Communities of Interest

  • Autonomy

DTIC Thesaurus Topics

  • Animal Training
  • Classification
  • Communication Systems
  • Computer Simulations
  • Computing System Architectures
  • Control Systems
  • Dynamic Programming
  • Environment
  • Intelligent Agents
  • Learning
  • Maps
  • Motion Planning
  • Navigation
  • Network Science
  • Neural Networks
  • Reinforcement Learning
  • Sensation

Readers

  • Artificial Intelligence
  • Economics
  • Systems Analysis and Design

Technology Areas

  • AI & ML
  • AI & ML - Autonomous Systems
  • AI & ML - DoD AI Strategy