Name
Abstract | ||
|---|---|---|
How could we solve the machine learning and the artificial intelligence problem if we had infinite computation? Solomonoff induction and the reinforcement learning agent AIXI are proposed answers to this question. Both are known to be incomputable. In this paper, we quantify this using the arithmetical hierarchy, and prove upper and corresponding lower bounds for in-computability. We show that AIXI is not limit computable, thus it cannot be approximated using finite computation. Our main result is a limit-computable e-optimal version of AIXI with infinite horizon that maximizes expected rewards. |
Year | Venue | Keywords |
|---|---|---|
2015 | International Conference on Uncertainty in Artificial Intelligence | AIXI, Solomonoff induction, general reinforcement learning, computability, complexity, arithmetical hierarchy, universal Turing machine |
Field | DocType | Volume |
Discrete mathematics,Universal Turing machine,AIXI,Computability,Solomonoff's theory of inductive inference,Arithmetical hierarchy,Infinite horizon,Artificial intelligence,Mathematics,Machine learning,Reinforcement learning,Computation | Journal | abs/1510.05572 |
ISBN | Citations | PageRank |
978-0-9966431-0-8 | 5 | 0.64 |
References | Authors | |
12 | 2 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Jan Leike | 1 | 150 | 15.49 |
| Marcus Hutter | 2 | 1302 | 132.09 |