Logic Program Synthesis from Incomplete Information - Flener, Pierre

. .

Résumé :
Program synthesis is a solution to the software crisis. If we had a program that develops correct programs from specifications, then program validation and maintenance would disappear from the software life-cycle, and one could focus on the more creative tasks of specification elaboration, validation, and maintenance, because replay of program development would be less costly. This monograph describes a novel approach to Inductive Logic Programming (ILP), which cross-fertilizes logic programming and machine learning. Aiming at the synthesis of recursive logic programs only, and this from incomplete information, we take a software engineering approach that is more appropriate than a pure artificial intelligence approach. This book is suitable as a secondary text for graduate level courses in software engineering and artificial intelligence, and as a reference for practitioners of program synthesis.

Sommaire:
I State of the Art.- 1 Automatic Programming.- 2 Deductive Inference in Automatic Programming.- 3 Inductive Inference in Automatic Programming.- 4 A Logic Program Development Methodology.- 5 Objectives.- II Building Blocks.- 6 A Specification Approach.- 7 A Framework for Stepwise Logic Algorithm Synthesis.- 8 Algorithm Analysis and Algorithm Schemata.- 9 The Proofs-as-Programs Method.- 10 The Most-Specific-Generalization Method.- III A Logic Algorithm Synthesis Mechanism.- 11 Overview of the Synthesis Mechanism.- 12 The Expansion Phase.- 13 The Reduction Phase.- 14 Conclusion.- Appendix: Conventions, Abbreviations, and Symbols.- References.