Learning on the Phase Transition Edge
Proceedings of the Seventeenth International Joint Conference on Artificial Intelligence. Pages 921-925.
Abstract
A previous research has shown that most learning strategies fail to learn relational concepts when descriptions involving more than three variables are required. The reason resides in the emergence of a phase transition in the covering test. After an in depth analysis of this aspect, this paper proposes an alternative learning strategy, combining a Monte Carlo stochastic search with local deterministic search. This approach offers two main benefits: on the one hand, substantial advantages over more traditional search algorithms, in terms of increased learning ability, and, on the other, the possibility of an a-priori estimation of the cost for solving a learning problem, under specific assumptions about the target concept.
@inproceedings{ijcai2001p126,
title = {Learning on the Phase Transition Edge},
author = {Serra, Alessandro and Giordana, Attilio and Saitta, Lorenza},
booktitle = {Proceedings of the Seventeenth International Joint Conference on
Artificial Intelligence, {IJCAI-01}},
publisher = {International Joint Conferences on Artificial Intelligence},
pages = {921--925},
year = {2001},
url = {https://ijcai.org/proceedings/2001/126},
}
TY - CPAPER TI - Learning on the Phase Transition Edge AU - Serra, Alessandro AU - Giordana, Attilio AU - Saitta, Lorenza T2 - Proceedings of the Seventeenth International Joint Conference on Artificial Intelligence, IJCAI-01 PB - International Joint Conferences on Artificial Intelligence PY - 2001 SP - 921 EP - 925 UR - https://ijcai.org/proceedings/2001/126 ER -Copied
Machine-readable versions of this paper, for tools and AI agents:
