Learning Strategies Using Decision Lists

Abstract

This paper presents a new technique for improving the efficiency of problem solvers by using trace data. The learning schema of this kind is usually called speedup learning . Although this learning paradigm has been formalized by [Nat89] based on the Probably Approximately Correct Learning model[Val84], his algorithm does not use operator sequence information of traces at all. We give a new algorithm which makes use of operator sequence information, and its effectiveness is shown by experiments in several domains. We also show theoretical backgrounds of the algorithm.

Cite

Text

Kobayashi. "Learning Strategies Using Decision Lists." International Conference on Algorithmic Learning Theory, 1993. doi:10.1007/3-540-57370-4_61

Markdown

[Kobayashi. "Learning Strategies Using Decision Lists." International Conference on Algorithmic Learning Theory, 1993.](https://mlanthology.org/alt/1993/kobayashi1993alt-learning/) doi:10.1007/3-540-57370-4_61

BibTeX

@inproceedings{kobayashi1993alt-learning,
  title     = {{Learning Strategies Using Decision Lists}},
  author    = {Kobayashi, Satoshi},
  booktitle = {International Conference on Algorithmic Learning Theory},
  year      = {1993},
  pages     = {370-383},
  doi       = {10.1007/3-540-57370-4_61},
  url       = {https://mlanthology.org/alt/1993/kobayashi1993alt-learning/}
}