Uncertainty in Artificial Intelligence (B-KUL-H02D2A)

4.0 ECTS English 46.5 First termFirst term Advanced
POC Artificial Intelligence

The student understands and appreciates the role and need for uncertainty in artificial intelligence systems.

The student knows, understands and is able to apply the Bayesian approach for dealing with uncertainty; he is familiar with the key concepts and algorithms underlying graphical models (such as Bayesian networks (directed graphical models), Markov networks (Markov random field, undirected graphical model), Factor graphs, and Hidden Markov models) such as modeling, inference and learning.

The student understands how techniques for reasoning about uncertainty can be integrated with logic for reasoning and learning, that is, he is familiar with statistical relational learning principles and techniques.

The student is able to identify the common, generic concepts and algorithms in papers published in various domains, using different terminology and having different trade-offs in their knowledge systems.

Default requirements for all Master of Artificial Intelligence students.

The  course will be based on (selected) parts of David Barber's forthcoming book on Bayesian Reasoning and Machine Learning, available  from http://www.cs.ucl.ac.uk/staff/d.barber/brml.

Text book

Activities

0.5 ects. Uncertainty in Artificial Intelligence: Exercises (B-KUL-H00H2a)

0.5 ECTS English 15.0 First termFirst term
POC Artificial Intelligence

There are 6 exercises sessions (mostly with pen and pencil) on various aspects of uncertainty reasoning and graphical models.

3.0 ects. Uncertainty in Artificial Intelligence: Lecture (B-KUL-H02D2a)

3.0 ECTS English 16.5 First termFirst term
POC Artificial Intelligence

Bayesian probability theory: modeling, inference, reasoning, decision making

Graphical models -- Bayesian networks, Markov Networks and Factors Graphs

Independence in graphical models

Inference algorithms 

Hidden and observable parameters

Learning


Dynamic systems (such as Hidden Markov Models and Kalman Filters)

Introduction to statistical relational learning, combining logic with graphical models

How logic and graphical models can be combined

Knowledge Based Model Construction

Probabilistic Prolog Systems

Applications

0.5 ects. Uncertainty in Artificial Intelligence: Project (B-KUL-H08M4a)

0.5 ECTS English 15.0 First termFirst term
POC Artificial Intelligence

The project consists of one or more assignments, possibly involving tasks with implementations of graphical models.

Evaluation

Evaluation : Uncertainty in Artificial Intelligence (B-KUL-H22D2a)

Mode of evaluation : Written
Category : interim evaluations plus final examination during examination period
Type of evaluation : Open book, Presentation

The evaluation consists of 
    an open book exam (during the exam period)
    an oral presentation of the results of the project