This is the first book dedicated to uniting research related to speech and speaker recognition based on the recent advances in large margin and kernel methods.