Augmented Statistical Models: Using Dynamic Kernels for Acoustic Models | part of Automatic Speech and Speaker Recognition: Large Margin and Kernel Methods | Wiley Telecom books | IEEE Xplore