Incorporating Modulation Information into Deep Neural Networks for Robust Speech Processing