Genetic Programming for Robust Text Independent Speaker Verification

The aim of this chapter is to investigate whether Genetic Programming (GP) can be used as a nature inspired human competitive solution for producing machine. A primary goal of machine learning strategies is to allow the user to state a problem in as high a level as possible and provide a solution without further human intervention. While we are still some way off this goal, GP has proven itself a powerful problem solver in many hard domains. We do not go in depth into the background and theory underlying GP, but instead choose to demonstrate its effectiveness at solving a real world problem, namely Automatic Speaker Verification (ASV), and demonstrate human competitive results. abstract