CMPE547 (Past CMPE58K) - Bayesian Statistics and Machine Learning

Spring 2017

Instructor: Ali Taylan Cemgil
(Volunteer) TA: Semih Akbayrak
Bogazici University,
Department of Computer Engineering,
Istanbul, Turkey

Announcements

  • (28.03) Example Exam. You can bring an A4 cheat sheet with your hand writing.

  • (08.02) The PS will be between 9:00-10:00 before the lecture, the starting date will be announced

Github site for project submissions

Jupyter Notebooks (in preperation)

Slides

Exercise Booklet

booklet.pdf

Timetable

08-Feb Introduction, Review, Bayes Theorem, Probability tables Barber Ch1
15-Feb Bayesian Inference, Applications Barber Ch2, Ch3
22-Feb Directed Graphical Models, Conditional Independence, D-Separation, Undirected Graphical Models, Factor Graphs, Barber Ch4
01-Mar Exact inference in chains and trees, Sum-Product Algorithm, Bucket elimination Barber Ch5
08-Mar Probability Models, ML, MAP and Bayesian Learning, Exponential Family, Learning as Inference, the EM Algorithm Barber Ch8.1-3, Ch9.1-2
15-Mar Multivariate Gaussians, Linear Models, Bayesian Linear Models Barber Ch8.4, Ch17.1-2, 18.1
22-Mar Gaussian Processes, Factor Analysis Barber Ch19.1-3, Ch21.1-4
29-Mar Midterm, in class
05-Apr Discrete state space Markov models, Hidden Markov Models, Forward Backward algorithm, Filtering, Smoothing, Correction smoother Barber Ch23.1-2
12-Apr Linear Dynamical Systems (LDS's), Inference in LDS, Kalman Filter and Smoother, Dynamic Bayes nets, Switching state space models Barber Ch24.1-4 Ch25.1-2
19-Apr Approximate inference, Variational Methods, Deadline for project proposal
26-Apr Spring Break
03-May Hierarchical models, Variational Bayes Barber Ch28.1-4
10-May Review

Chapter numbers are given according to the DRAFT June 18, 2013 version of Bayesian Reasoning and Machine Learning by David Barber.

After Final exam week, there will be a half day poster Presentations.

 
$\Large p(\lambda| x) = p(x|\lambda) p(\lambda) / \int d\lambda' p(x|\lambda') p(\lambda') $