An analysis of first-order logics of probability
Artificial Intelligence
From statistical knowledge bases to degrees of belief
Artificial Intelligence
A decision-theoretic generalization of on-line learning and an application to boosting
Journal of Computer and System Sciences - Special issue: 26th annual ACM symposium on the theory of computing & STOC'94, May 23–25, 1994, and second annual Europe an conference on computational learning theory (EuroCOLT'95), March 13–15, 1995
Improved boosting algorithms using confidence-rated predictions
COLT' 98 Proceedings of the eleventh annual conference on Computational learning theory
Foundations of Logic Programming
Foundations of Logic Programming
Relational Learning Using Constrained Confidence-Rated Boosting
ILP '01 Proceedings of the 11th International Conference on Inductive Logic Programming
ALT '96 Proceedings of the 7th International Workshop on Algorithmic Learning Theory
| Hi-index | 0.00 |
A popular idea is that the longer the proof the riskier the truth prediction. In other words, the uncertainty degree over a conclusion is an increasing function of the length of its proof. In this paper, we analyze this idea in the context of Inductive Logic Programming. Some simple probabilistic arguments lead to the conclusion that we need to reduce the length of the clause bodies to reduce uncertainty degree (or to increase accuracy). Inspired by the boosting technique, we propose a way to implement the proof reduction by introducing weights in a well-known ILP system. Our preliminary experiments confirm our predictions.