ORACLES IN
ARE SUFFFICIENT FOR EXACT LEARNING
Abstract
We study the learnability of representation classes in Angluin's exact learning model. In particular, we consider the following three query types: equivalence queries, equivalence and membership queries, and membership queries only. We show in all three cases that polynomial query complexity implies already polynomial-time learnability, provided that the learner additionally has access to an oracle in . It follows that boolean circuits are polynomial-time learnable with equivalence queries and the help of an oracle in
.a
Some of these results were presented at the 8th International Workshop on Algorithmic Learning Theory [12].