SubjectsSubjects(version: 945)
Course, academic year 2023/2024
   Login via CAS
Probabilistic graphical models - NAIL104
Title: Pravděpodobnostní grafické modely
Guaranteed by: Department of Theoretical Computer Science and Mathematical Logic (32-KTIML)
Faculty: Faculty of Mathematics and Physics
Actual: from 2012
Semester: winter
E-Credits: 3
Hours per week, examination: winter s.:2/0, Ex [HT]
Capacity: unlimited
Min. number of students: unlimited
4EU+: no
Virtual mobility / capacity: no
State of the course: taught
Language: Czech, English
Teaching methods: full-time
Teaching methods: full-time
Guarantor: Mgr. Marta Vomlelová, Ph.D.
Class: Informatika Mgr. - Teoretická informatika
Classification: Informatics > Theoretical Computer Science
Annotation -
Last update: T_KTI (03.05.2012)
The course extends the basics of probabilistic graphical models introduced in the NAIL070 Artificial Intelligence 2 course: Bayesian networks and their extensions (DBN, OOBN), decision graphs, partially observable markov decision processes (POMDP) and conditional random fields. We focus on the modelling languages and their evaluation methods. We touch also some applications.
Aim of the course -
Last update: Mgr. Marta Vomlelová, Ph.D. (14.05.2021)

The course gives an introduction to probabilistic graphical models. The students will learn the following formal models, evaluation and model learning algorithm, application areas.

Course completion requirements -
Last update: Mgr. Marta Vomlelová, Ph.D. (07.06.2019)

The exam consists of a written preparation and an oral part. The requirements are given by the course syllabus.

Literature -
Last update: Mgr. Marta Vomlelová, Ph.D. (20.04.2016)
  • S. Hojsgaard, D. Edwards, S. Lauritzen: Graphical Models with R, Springer 2012
  • Finn V. Jensen, Thomas D. Nielsen: Bayesian Networks and Decision Graphs, Springer 2007
  • Leslie Pack Kaelbling, Michael L. Littman, and Anthony R. Cassandra. Planning and acting in partially observable stochastic domains. Artificial Intelligence, Volume 101, pp. 99-134, 1998
  • John Lafferty, Andrew McCallum, Rernando Pereira: Conditional random fields: Probabilistic models for segmenting and labeling sequence data, Morgan Kaufmann 2001, pp. 282-289

Requirements to the exam -
Last update: Mgr. Marta Vomlelová, Ph.D. (07.06.2019)

The exam consists of a written preparation and an oral part. The requirements are given by the course syllabus.

Syllabus -
Last update: Mgr. Marta Vomlelová, Ph.D. (09.05.2023)

1) A brief refresh of the Artificial Intelligence 2 course, Causal and Bayesian networks,

2) advanced evaluation methods: d-separation, junction tree, message passing scheme,

3) dynamic Bayesian networks DBNs,

4) learning Bayesian networks,

5) decision graphs,

6) POMDP - partially observable Markov decision problems,

7) variational approximate inference

8) example applications.

Basic introduction into the Python libraries pgmpy, bayespy.

 
Charles University | Information system of Charles University | http://www.cuni.cz/UKEN-329.html