Introduction to Speech and Audio Processing

3 Hours

Explore the theory and practice that underlies modern automatic speech recognition (ASR) systems and related technologies. Analyze the acoustic theory of speech production, digital signal processing for speech signals, Hidden Markov Models, Gaussian Mixture Models, language models, finite state transducers, decoding algorithms, and deep neural networks. Computer Science 342S and 378 (Topic: Intro to Speech and Audio Proces) may not both be counted. Three lecture hours a week for one semester.