Parameterized Action Representation and Natural Language Instructions for Dynamic Behavior Modification of Embodied Agents

Norman I. Badler, Rama Bindiganavale, Jan Allbeck, William Schuler, Liwei Zhao, Seung-Joo Lee, Hogeun Shin and Martha Palmer

We introduce a prototype for building a strategy game. A player can control and modify the behavior of all the characters in a game, and introduce new strategies, through the powerful medium of natural language instructions. We describe a Parameterized Action Representation (PAR) designed to bridge the gap between natural language instructions and the virtual agents who are to carry them out. We will illustrate PAR through an interactive demonstration of a multi-agent strategy game.


This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.