audio and speech processing, neural networks, probabilistic models