http://sange.fi/~atehwa/cgi-bin/piki.cgi/ Recent changes in Inductive Logic Programming: from Machine Learning to Software Engineering http://sange.fi/~atehwa/cgi-bin/piki.cgi/Inductive%20Logic%20Programming%3A%20from%20Machine%20Learning%20to%20Software%20Engineering http://sange.fi/~atehwa/cgi-bin/piki.cgi/#1183657230 <p>[...] <p><del>A</del> <ins>(((A</ins> program, its examples, and the procedures that are used to produce one from the other are naturally intimately connected. A procedure that produces a program from examples is an inductive inference machine IIM(''e'') = ''P''; a procedure that produces a set of examples from a program is a test set generation procedure TSG(''P'') = ''e''. ILP usually studies the construction of a known program. Basically, it could study the selection of an optimal example set for a given inductive inference machine, but usually studies the selection of the optimal IIM for a given set of examples (which is usually quite arbitrarily chosen). For induction, we could also study the optimal program to be produced for a given set of examples -- i.e. what the "correct" result of an IIM actually <del>is.</del> <ins>is.)))</ins> <p>[...] Thu, 05 Jul 2007 17:40:30 +0000