Learning by Reading for Robust Reasoning in Intelligent Agents

Abstract

Our hypotheses are that analogical processing plays multiple roles in enabling machines to learn by reading, and that qualitative representations provide a foundation for commonsense knowledge that can be used for reasoning and for natural language semantics. Our accomplishments on this project provide strong evidence for these hypotheses.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Apr 24, 2018
Accession Number
AD1050969

Entities

People

  • Ken Forbus

Organizations

  • Northwestern University

Tags

Communities of Interest

  • Engineered Resilient Systems

DTIC Thesaurus Topics

  • Algorithms
  • Analogies
  • Artificial Intelligence
  • Artificial Intelligence Software
  • Cognitive Science
  • Computer Languages
  • Computer Programs
  • Computer Science
  • Computer Vision
  • Construction
  • Electrical Engineering
  • Hypotheses
  • Intelligent Agents
  • Language
  • Learning
  • Natural Language Understanding
  • Natural Languages
  • Ontologies
  • Psychology
  • Reasoning
  • Semantics
  • Standards
  • Universities

Readers

  • Artificial Intelligence