User Tools

Site Tools


start

EECS 6327 Probabilistic Models & Machine Learning (Winter 2018)

Description

The course introduces some probabilistic models and machine learning methods. The covered topics may include:

  • Bayesian Decision theory, Generative vs Discriminative modelling
  • Generative Models (1) - multivariate Gaussian, Gaussian mixture model (GMM), Multinomial, Markov chain model, n-gram
  • Generative Models (2) - graphical models, directed vs. indirected graphical models, exact inference, approximate inference (loopy belief propagation, variational inference, Monte Carlo methods)
  • Discriminative Models (1) - linear discriminant, linear regression, lasso, logistic regression, support vector machines (SVM), sparse kernel machines
  • Discriminative Models (2) - neural networks (NN), back-propagation, deep learning, auto-encoder; recurrent neural networks, convolutional neural networks
  • Advanced models: hidden Markov model (HMM), Latent Dirichlet Allocation (LDA), Conditional Random Fields (CRF), Convolutional Neural Nets, Recurrent Neural Nets and LSTMs
  • Advanced topics: Learnability, Gaussian Processes, Ensemble Methods, Reinforcement Learning, etc.

Announcements

Lecture Times

  • Section A: Tuesdays and Thursdays, 4:00pm - 5:30pm, located at BC228 SC 214.

Lecturer

start.txt · Last modified: 2019/08/22 21:28 by hj