xep33gmm -- Graphical Markov Models

Overview

Markov models on graphs represent a model class widely applied in many areas of computer science, such as computer networks, data security, robotics, pattern recognition and computer vision. The first part of the course covers inference and learning for Markov models on chains and trees. Almost all these tasks including structure learning can be solved by efficient algorithms with polynomial time complexity. The second part of the course addresses graphical models on general graphs. Here on the contrary, practically all inference and learning tasks are NP-complete. The focus is therefore on efficient approximation algorithms.

News

  • 27.10.2020 There will be no lecture tomorrow (28.10) because of the national holiday. Next lecture and seminar will take place next week (4.11.) The assignments for the seminar will be published on Thursday (29.10.)
  • 23.09.2020 According to the timetable for this winter semester, seminars (labs) are taking place in odd weeks. The first seminar will take place on October, 7. The assignments for it are published.
  • 21.09.2020 The city of Prague has flagged corona status “red” last Friday. This means to our deep regret that teaching at universities has to switch to fully online mode. The lectures/seminars of the course will take place as scheduled in the faculty timetables. We will use the BigBlueButton platform (BBB), which runs safely and securely on local university servers. All you need for joining the virtual teaching rooms is a web-browser and a headset. If you are not familiar with this platform, please have a look at https://bigbluebutton.org/ and the demo videos provided there. We will schedule the meetings to start a few minutes ahead of the regular time, so that you can set up your gear and join them without hurry. You will receive invitation e-mails well before each scheduled meeting. The BBB platform is integrated in the faculty upload system BRUTE. Therefore you can also join the meetings via BRUTE.

Details

  • Teacher: Boris Flach web-page
  • Prerequisites: probability theory and mathematical statistics, graphs and graph algorithms, pattern recognition
  • Course format: (2/1)
  • Schedule: WS20/21
  • Lectures: See here for the syllabus
  • Seminars: Exercises/Assignments will be provided prior to each seminar class. You are supposed to work on them before the seminar. Solutions will be discussed at the seminar class.
  • Grading/Credits: Written exam (see here for example1, example2), 4CP
  • Textbooks and References:
    • Michail I. Schlesinger, Vaclav Hlavač, Ten Lectures on Statistical and Structural Pattern Recognition, chapter 8 [Schlesinger-TLPR2002]
    • Stan Z. Li, Markov Random Field Modeling in Image Analysis [Li-MRFIA2009]
    • Daphne Koller, Nir Friedman, Probabilistic Graphical Models Principles and Techniques [Koller-PGM2009]
    • Christopher M. Bishop, Pattern Recognition and Machine Learning (for additional reading) [Bishop-PRML2006]
    • Gerhard Winkler, Image Analysis, Random Fields and Markov Chain Monte Carlo Methods: A Mathematical Introduction (for additional reading) [Winkler-IARF2006]
courses/xep33gmm/start.txt · Last modified: 2020/10/27 14:26 by flachbor