Inductive Program Synthesis by Concept Learning

A method of inductive program synthesis by concept learning is presented in this paper. Hypothesis formation and validation are the two main processes of induction. In our method, hypotheses are formed and generalized according to the background knowledge. The technique of automatic construction of critical examples is adopted to validate the formed hypotheses.