A Theory of Universal Artificial Intelligence based on Algorithmic Complexity

TL;DR

This paper introduces AIXI, a universal AI model combining decision theory and sequence prediction, which is theoretically optimal but uncomputable, leading to a computable approximation called AIXI-tl.

Contribution

It presents the AIXI model as a formal, unifying theory of universal artificial intelligence, and proposes a computable approximation called AIXI-tl.

Findings

AIXI is the most intelligent unbiased agent possible.

AIXI-tl is effectively more intelligent than any other bounded agent.

AIXI-tl's computation time is of the order t x 2^l.

Abstract

Decision theory formally solves the problem of rational agents in uncertain worlds if the true environmental prior probability distribution is known. Solomonoff's theory of universal induction formally solves the problem of sequence prediction for unknown prior distribution. We combine both ideas and get a parameterless theory of universal Artificial Intelligence. We give strong arguments that the resulting AIXI model is the most intelligent unbiased agent possible. We outline for a number of problem classes, including sequence prediction, strategic games, function minimization, reinforcement and supervised learning, how the AIXI model can formally solve them. The major drawback of the AIXI model is that it is uncomputable. To overcome this problem, we construct a modified algorithm AIXI-tl, which is still effectively more intelligent than any other time t and space l bounded agent. The computation time of AIXI-tl is of the order tx2^l. Other discussed topics are formal definitions of intelligence order relations, the horizon problem and relations of the AIXI theory to other AI approaches.