Asymptotically Optimal Agents

Preprint English OPEN
Lattimore, Tor ; Hutter, Marcus (2011)
  • Subject: Computer Science - Artificial Intelligence | Computer Science - Learning
    arxiv: Computer Science::Multiagent Systems

Artificial general intelligence aims to create agents capable of learning to solve arbitrary interesting problems. We define two versions of asymptotic optimality and prove that no agent can satisfy the strong version while in some cases, depending on discounting, there does exist a non-computable weak asymptotically optimal agent.
  • References (4)

    [Hut02] Marcus Hutter. Self-optimizing and Pareto-optimal policies in general environments based on Bayes-mixtures. In Proc. 15th Annual Conf. on Computational Learning Theory (COLT'02), volume 2375 of LNAI, pages 364-379, Sydney, 2002. Springer, Berlin.

    [Hut04] Marcus Hutter. Universal Artificial Intelligence: Sequential Decisions based on Algorithmic Probability. Springer, Berlin, 2004.

    [Leg06] Shane Legg. Is there an elegant universal theory of prediction? In Jos Balczar, Philip Long, and Frank Stephan, editors, Algorithmic Learning Theory, volume 4264 of Lecture Notes in Computer Science, pages 274-287. Springer Berlin / Heidelberg, 2006.

    [Ors10] Laurent Orseau. Optimality issues of universal greedy agents with static priors. In Marcus Hutter, Frank Stephan, Vladimir Vovk, and Thomas Zeugmann, editors, Algorithmic Learning Theory, volume 6331 of Lecture Notes in Computer Science, pages 345-359. Springer Berlin / Heidelberg, 2010.

  • Metrics
    No metrics available
Share - Bookmark