Bad Universal Priors and Notions of Optimality
Abstract
A big open question of algorithmic information theory is the choice of the universal Turing machine (UTM). For Kolmogorov complexity and Solomonoff induction we have invariance theorems: the choice of the UTM changes bounds only by a constant. For the universally intelligent agent AIXI (Hutter, 2005) no invariance theorem is known. Our results are entirely negative: we discuss cases in which unlucky or adversarial choices of the UTM cause AIXI to misbehave drastically. We show that Legg-Hutter intelligence and thus balanced Pareto optimality is entirely subjective, and that every policy is Pareto optimal in the class of all computable environments. This undermines all existing optimality properties for AIXI. While it may still serve as a gold standard for AI, our results imply that AIXI is a relative theory, dependent on the choice of the UTM.
Cite
Text
Leike and Hutter. "Bad Universal Priors and Notions of Optimality." Annual Conference on Computational Learning Theory, 2015.Markdown
[Leike and Hutter. "Bad Universal Priors and Notions of Optimality." Annual Conference on Computational Learning Theory, 2015.](https://mlanthology.org/colt/2015/leike2015colt-bad/)BibTeX
@inproceedings{leike2015colt-bad,
title = {{Bad Universal Priors and Notions of Optimality}},
author = {Leike, Jan and Hutter, Marcus},
booktitle = {Annual Conference on Computational Learning Theory},
year = {2015},
pages = {1244-1259},
url = {https://mlanthology.org/colt/2015/leike2015colt-bad/}
}