Applications of Machine Learning to Software Testing

Lionel Claude Briand
Simula Research Laboratory
http://squall.sce.carleton.ca/people/briand
http://www.simula.no/portal_memberdata/briand

Abstract:

A great deal of software testing, especially when testing larger
components, subsystems, or entire systems, is done using
specification-based, black-box techniques. Such techniques require
expertise and are error-prone.  This talk will present two
applications of machine learning in the context of black-box
testing. One will focus on helping testers improve their black-box
test specifications and the other will then exploit such black-box
specifications to help improve existing debugging techniques.