A framework for representing action meaning in artificial systems via force dimensions

  • Authors:
  • Paul Hemeren;Peter Gärdenfors

  • Affiliations:
  • Interaction Lab, School of Humanities and Informatics, University of Skövde, Skövde, Sweden;Lund University Cognitive Science, Lund University, Lund, Sweden

  • Venue:
  • AGI'12 Proceedings of the 5th international conference on Artificial General Intelligence
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

General (human) intelligence critically includes understanding human action, both action production and action recognition. Human actions also convey social signals that allow us to predict the actions of others (intent) as well as the physical and social consequences of our actions. What's more, we are able to talk about what we (and others) are doing. We present a framework for action recognition and communication that is based on access to the force dimensions that constrain human actions. The central idea here is that forces and force patterns constitute vectors in conceptual spaces that can represent actions and events. We conclude by pointing to the consequences of this view for how artificial systems could be made to understand and communicate about actions.