empfohlenes Vorwissen

Skills in programming in Prolog. These may be acquired by parallel attendance of the module „Artificial Intelligence - Prolog “. 

Lernziele

Students

• get to know the principal methods of AI based on deterministic algorithms and logic as well as economic fields of application
• know fundamental concepts of advanced searching strategies
• are able to program a complex agent in a predefined environment
• gain access to modern concepts of logic, esp. higher order logic
• know how uncertainty and learning fit into the context of AI
• understand the limitations and ethical implications of AI

Lehrinhalte

Artificial Intelligence (AI) is concerned with the study, understanding, and application of behavior reserved to humans until recent decades on the one hand, as well as to providing meaningful behavior in the broadest sense surpassing the capabilities of conventional software systems on the other. The lecture will have a focus on the foundations of AI specifically rational agents, search and logic.

One topic are various kinds of search including more advanced methods like simplified memory bounded A*, adversarial or local search, with an emphasis on practical implementation in non-trivial examples.

Another topic are logic based agents and logic in general from the propositional calculus to first order logic and corresponding inference mechanisms.

Learning and uncertainty round out the foundations of agent based AI.

At a more theoretical level, boundaries of formalization in general, like results of Gödel, Turing & Chaitin are taken into consideration, developing a practical understanding of the concepts of complexity, shedding light on their role in the approaches studied before.

In conclusion the ethical implications of artificial intelligence are discussed.

• Outline
• Introduction to AI
• Agents
• Search
• Logic
• Uncertainty
• Learning
• Limitations and ethical implications of AI

Literaturangaben

• Russel, Stuart und Norvig, Peter (2003): Artificial Intelligence. A Modern Approach. 2nd Edition, Prentice Hall.

Prüfungsmodalitäten

Written exam (60%) plus supplementary deliverables (40%).

empfohlenes Vorwissen

Die Veranstaltung setzt Kenntnisse der Programmiersprache Prolog voraus. Diese können durch Belegen des Moduls „KI-Programmierung“ erworben werden.

Lernziele

Die Studierenden

• sind in der Lage, abhängig von der Problemstellung zu entscheiden, welche KI-Methoden zur Problemlösung einzusetzen sind
• implementieren davon ausgehend konkrete Lösungen
• können alle notwendigen Prozesse des Projekts selbständig managen

Lehrinhalte

Die Studierenden entwickeln komplexen KI-Anwendungen im Rahmen von Gruppenarbeiten. Die Anwendungsbeispiele wechseln von Jahr zu Jahr.

Literaturangaben

Die Aktualität des Themas lässt gegenwärtig keine spezifischen Literaturangaben zu.

Prüfungsmodalitäten

Abschließende mündliche Prüfung (in der Regel: 20 bis 40 Minuten).