Frey, et al.. ALGONQUIN - Learning Dynamic Noise Models from Noisy Speech for Robust Speech Recognition. 2001.