Learning simple recursive theories
The task of learning relations has been concerned, so far, with the acquisition of intensional descriptions of unrelated concepts. However, in many real domains concepts are strictly related to each other and the instances of one of them cannot possibly be recognised without previous recognition of other objects as instances of related concepts. A typical case is the problem of labelling parts of a scene in order to interpret it. This paper extends in several ways the learning relations paradigm; in particular, a new methodology, allowing a recursive theory to be inferred from a set of examples, is presented. The learning algorithm works bottom-up, creating first an acyclic graph that classifies all the instances in the training set. Afterwards, a recursive theory is synthesised from the graph.
Unable to display preview. Download preview PDF.
- 1.H. Shapiro: Algorithmic Program Debugging. MIT Press: 1982.Google Scholar
- 2.R. Quinlan: Learning Logical Definitions from Relations. Machine Learning, 5, 239–266 (1990).Google Scholar
- 3.S. Muggleton and W. Buntine: Machine Invention of First-order Predicates by Inverting Resolution. Proc. Fifth International Conference on Machine Learning. Ann Arbor: 1988, pp. 339–352.Google Scholar
- 4.A.W. Biermann: The Inference of Regular Lisp Programs from Examples. IEEE Trans. on Systems Man and Cybernetics, 8, 585–600 (1978).Google Scholar
- 5.K. Apt, H. Blair and A. Walker: Towards a Theory of Declarative Knowledge. Proc. Workshop on Foundations of Deductive Databases and Logic Programming. Washington: 1986.Google Scholar
- 6.K. Apt and M. Bezem: Acyclic Programs. Proc. 7th. Conference on Logic Programming. Jerusalem: 1990, pp. 617–633.Google Scholar