Universal Principles of Intelligent System Design

Authors

  • Kirill Popov University of California, Berkeley
  • Aamod Shanker
  • Debanjan Bhowmik

Keywords:

Intelligence, Information, Prediction, Causality, Memory, Design, Dynamical systems, Thermodynamics, Learning, epistemology

Abstract

We propose a universal definition of intelligence of a physical system and discuss its implications on design of intelligent systems. The definition proposed is universally valid and does not invoke teleological or anthropic concepts. We discuss the relationship of intelligence to energetic properties by invoking recent results in inequilibrium thermodynamics and computational mechanics. Intelligent system design is reformulated in three natural problems: selective forgetting, memory maintenance and self-recording. We conclude with highlighting the relationship of energetic and informational optimality principles involved in designing intelligent physical systems or their spontaneous emergence.

Downloads

Published

12-01-2018

How to Cite

Popov, K., Shanker, A., & Bhowmik, D. (2018). Universal Principles of Intelligent System Design. Cosmos and History: The Journal of Natural and Social Philosophy, 14(1), 113–122. Retrieved from http://cosmosandhistory.org/index.php/journal/article/view/699