Extending Interactive Task Learning

Abstract

In this research, we build on our previous ITL research and propose to significantly expand the generality of ITL beyond games, puzzles, and simple robotic tasks. We will focus on increasing the generality of ITL agents along the following dimensions:1. Expand the types of tasks that can be learned to include not only goal-driventasks, but also optimization, procedural, and composite tasks.2. Expand the types of knowledge that can be learned, achieving ???knowledgecompleteness??? so that all types of knowledge the agents can learn aretheoretically learnable.3. Expand the capabilities for interacting with the environment through richer andbroader manipulation actions including those using tools.The research will be developed and evaluated using real-world robotic platforms,including the Fetch mobile manipulator robot.

Document Details

Document Type
DoD Grant Award
Publication Date
Jul 10, 2018
Source ID
N000141812337

Entities

People

  • John E. Laird

Organizations

  • Board of Regents of the University of Michigan
  • Office of Naval Research
  • United States Navy

Tags

Fields of Study

  • Computer science

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.

Technology Areas

  • AI & ML
  • AI & ML - Autonomous Systems
  • Autonomy