The course introduces some basic statistical modeling methods in pattern recognition and machine learning and their applications to speech and language processing. The covered topics may include:
Bayesian Decision theory, Generative vs Discriminative modeling
Generative Models - multivariate Gaussian, Gaussian mixture model (GMM), hidden Markov model (HMM), Multinomial, Markov chain model, n-gram, graphical models
Discriminative Models - linear discriminant, logistic regression, support vector machines (SVM), neural networks (NN), sparse kernel machines
Statistical Modeling Methods - maximum likelihood estimation (MLE), Expectation-Maximization (EM), discriminative training (DT)
Some selected applications - speech recognition, text categorization, machine translation, spoken language processing
The above methods are equally applicable to other research areas, such as data mining, information retrieval, computer vision, computational linguistics and so on.