Learning What to Value

  • Daniel Dewey
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 6830)


I. J. Good’s intelligence explosion theory predicts that ultraintelligent agents will undergo a process of repeated self-improvement; in the wake of such an event, how well our values are fulfilled would depend on the goals of these ultraintelligent agents. With this motivation, we examine ultraintelligent reinforcement learning agents. Reinforcement learning can only be used in the real world to define agents whose goal is to maximize expected rewards, and since this goal does not match with human goals, AGIs based on reinforcement learning will often work at cross-purposes to us. To solve this problem, we define value learners, agents that can be designed to learn and maximize any initially unknown utility function so long as we provide them with an idea of what constitutes evidence about that utility function.


Utility Function Reinforcement Learning Expected Utility Full Search Agent Function 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Good, I.J.: Speculations Concerning the First Ultraintelligent Machine. In: Alt, F.L., Rubinoff, M. (eds.) Advances in Computers, vol. 6, pp. 31–88 (1965)Google Scholar
  2. 2.
  3. 3.
    Hutter, M.: Universal algorithmic intelligence: A mathematical top-down approach. In: Artificial General Intelligence, pp. 227–290. Springer, Berlin (2007)CrossRefGoogle Scholar
  4. 4.
  5. 5.
    Omohundro: The Nature of Self-Improving Artificial Intelligence,
  6. 6.
    Omohundro, S.: The basic AI drives. In: Wang, P., Goertzel, B., Franklin, S. (eds.) Proceedings of the First AGI Conference on Frontiers in Artificial Intelligence and Applications, vol. 171. IOS Press, Amsterdam (2008)Google Scholar
  7. 7.
    Russell, S., Norvig, P.: AI A Modern Approach. Prentice-Hall, Englewood Cliffs (1995)Google Scholar
  8. 8.
    Yudkowsky, E.: Artificial intelligence as a positive and negative factor in global risk. In: Bostrom, N. (ed.) Global Catastrophic Risks. Oxford University Press, Oxford (2008)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2011

Authors and Affiliations

  • Daniel Dewey
    • 1
  1. 1.The Singularity Institute for Artificial IntelligenceSan FranciscoUSA

Personalised recommendations