Basic Information
The Speech Signal and Its Acoustics Characteristics, Pitch And Formants, Representations of The Speech Signal# Lpc, Spectrum, Mel, Cepstrum, Mfcc, Isolated Word Recognition, Dtw and Keyword Spotting, Automatic Speech Recognition# Acoustic Model, Pronunciation Model, Language Model, Acoustic Modeling# Hidden Markov Models (hmms), Deep Learning Models, Loss Functions and Ctc, Transformers, Speech Synthesis and Vocoders. Learning Outcomes# With The Completion of The Course, The Students Will Know# 1. Understand and Be Familiar With How Speech Is Generated And The Acoustic Components of The Speech Signal, How Speech Is Represented in Deep Neural Network Systems. 2. Be Familiar With Probabilistic Modeling and Deep Learning Algorithms For Estimating, Recognizing, and Detecting a Sequence Of Random Variables and How They Can Be Used With Speech Signals. 3. Be Familiar With The Most Recent Models For Automatic Speech Recognition and How to Train Them. 4. Will Be Familiar With The Most Recent Models For Generative Algorithms For Sequence Generation, Conditional And Unconditional, and How to Apply Them to Synthesize Human-like Speech.
Faculty: Electrical and Computer Engineering
|Undergraduate Studies
|Graduate Studies
Pre-required courses
(44198 - Intro. to Digital Signal Processing and 44202 - Random Signals and 46195 - Machine Learning)