Seminar: Machine Intelligence

Course Number 45366-01
Lecturers Marcel Luethi
Volker Roth
Thomas Vetter
Assistants Mario Wieser
Time and Location Mon 10:15 - 12:00; Seminarraum 00.003, Spiegelgasse 1
Start 26-02-2018
Prerequisites Open to Master and PhD students with basic knowledge in probability theory and statistics, linear algebra and some first introduction to learning, as for example taught in the course pattern recognition.
Objectives At the end of the course, students should have a good overview of different formalisms, methods and strategies to reason about uncertainty. Participants should also have learned to study, summarize and present scientific literature to a peer group in an understandable and instructive way.
Contents In this seminar we will discuss together the main concepts of reasoning under uncertainty. We will discuss questions such as:
How is uncertainty represented and quantified?
How can we use probability theory to reason under uncertainty?
How can we make good decisions if we have uncertain information?
We will also discuss different methods for learning from examples and look at strategies to learn when no examples, but only an assessment of the outcome of a situation is available (reinforcement learning). Finally, we will see these methods in application in the context of natural language processing.

Each participant will be assigned a topic, which he/she will summarize and present during the seminar. The topic is studied individually by the student, with the help of a tutor.
Literature Stuart Russel, Peter Norvig, Artificial Intelligence: A Modern Approach.
Assessment Lehrveranst.-begleitend

Please note: The final grade of the course is composed of the grade from 4 individual parts:
- A presentation about one of the topics discussed in the course book
- A written report, summarizing the same topic
- A review of a report of another student
- An oral exam (31.5. 2018 and 1.6.2018 )
Credit Points 6
Grades 1-6 0,5
Registration Services (Requires login)