pith. sign in

arxiv: cs/9505104 · v1 · pith:4JGWNEPCnew · submitted 1995-05-01 · 💻 cs.AI

Pac-Learning Recursive Logic Programs: Efficient Algorithms

classification 💻 cs.AI
keywords recursiveprogramslearnableclassesclauselogicalgorithmscompanion
0
0 comments X
read the original abstract

We present algorithms that learn certain classes of function-free recursive logic programs in polynomial time from equivalence queries. In particular, we show that a single k-ary recursive constant-depth determinate clause is learnable. Two-clause programs consisting of one learnable recursive clause and one constant-depth determinate non-recursive clause are also learnable, if an additional ``basecase'' oracle is assumed. These results immediately imply the pac-learnability of these classes. Although these classes of learnable recursive programs are very constrained, it is shown in a companion paper that they are maximally general, in that generalizing either class in any natural way leads to a computationally difficult learning problem. Thus, taken together with its companion paper, this paper establishes a boundary of efficient learnability for recursive logic programs.

This paper has not been read by Pith yet.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.