---
title: "Learning on the Phase Transition Edge"
authors:
  - "Alessandro Serra"
  - "Attilio Giordana"
  - "Lorenza Saitta"
doi: ""
url: "https://ijcai.org/proceedings/2001/126"
pdf: "https://ijcai.org/proceedings/2001/0126.pdf"
venue: "Proceedings of the Seventeenth International Joint Conference on Artificial Intelligence"
year: 2001
pages: "921-925"
track: ""
keywords:
---

# Learning on the Phase Transition Edge

Alessandro Serra, Attilio Giordana, Lorenza Saitta

*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.