A simple algorithm for learning O(log n)-term DNF

Since defined by Angluin [A88], the ezact learning model attracted a lot of attention. In this model, the learning algorithm tries to identify a target function using membership queries and equivalence queries. Several interesting concept classes were shown to be learnable in this model (for various examples, see [.A87, A87b, AP91, Han91, AP92, BR92, B93, SS93, BK+94, B95, BCV96]). Of particular relevance to the current note are .~ngluin”s algorithm for learning deterministic automata [.487b] and the algorithms for learning k-term DNF formulae. While for the case of constant k, a simple algorithm for learning k-term DNF was known [A87], it took some time before powerful techniques were developed that allow con-