Optimal Ordered Problem Solver
Abstract
We present a novel, general, optimally fast, incremental way of searching for a universal algorithm that solves each task in a sequence of tasks. The Optimal Ordered Problem Solver (OOPS) continually organizes and exploits previously found solutions to earlier tasks, efficiently searching not only the space of domainspecific algorithms, but also the space of search algorithms. Essentially we extend the principles of optimal nonincremental universal search to build an incremental universal learner that is able to improve itself through experience. In illustrative experiments, our selfimprover becomes the first general system that learns to solve all n disk Towers of Hanoi tasks (solution size 2^n1) for n up to 30, profiting from previously solved, simpler tasks involving samples of a simple context free language.
 Publication:

arXiv eprints
 Pub Date:
 July 2002
 arXiv:
 arXiv:cs/0207097
 Bibcode:
 2002cs........7097S
 Keywords:

 Computer Science  Artificial Intelligence;
 Computer Science  Computational Complexity;
 Computer Science  Machine Learning;
 I.2.2;
 I.2.6;
 I.2.8
 EPrint:
 43 pages, 2 figures, short version at NIPS 2002 (added 1 figure and references