German website

study program: Master of Computer Science (M_Inf), 1st/2nd semester

ECTS credits: 4

lecture term: winter semester

prerequesites: qualified bachelor program in computer science (or related), no prerequesite of master courses (bachelor course Fundamentals of AI is also not required)

language: This lecture will be given in English if there is at least one student not knowing German, otherwise in German.

Artificial Intelligence (AI) is the summary for all kinds of software dealing with complex problems and using human-prone techniques. Unfortunately, this very imprecise definition is the best which is universally accepted.

Since computer science has been an own science, the AI community has always been the first to promote innovative software techniques. In the development of new programming languages, AI provided milestones for new programming paradigms such as Prolog (for logical programming), Lisp (for functional programming), and Smalltalk (for object-oriented programming). AI also promoted new softwarearchitectures such as expert systems, knowledge-based systems, and multi-agent systems. A lot of concepts which had their origin in the AI community are now considered standard knowledge for all computer scientists such as object-oriented programming or multi-agent technology in distributed applications. They are now used also in (non-AI) standard software, sometimes even better than in the original AI software.

Concerning algorithmic techniques, there is no sharp distinction between "conventional programming" and AI. AI contributed search techniques which are normally used for problems that cannot be solved both, exactly and efficiently. Typical application domains of these techniques are planning tools in scheduling and logistics. The A* algorithm which is the basis for modern path algorithms in traffic applications as well as in computer games ("Game AI") may be regarded as an AI technique considering its serach strategy. But being a modification of the efficient basic path algorithm of Dijkstra, it may also be regarded as a standard (non-AI) algorithm.

The AI technology most established in practice is the architecture of knowledge-based systems which is a generalization of the older expert system technology.

Neural networks have been invented outside the classical AI community, because they did not evolve from logical programming which was the major AI programming technique when neural networks were invented. Due to the above general definition of AI, they are now also considered part of AI, together with other novel "softcomputing" techniques such as evolutionary or genetic algorithms.

Swarm intelligence is a special form of multi-agent techniques and, thus, does also belong to AI. The most prominent example are ant algorithms which are used in computer networks and in a research stage also in traffic networks. In general, swarm intelligence finds its use in distributed systems with highly dynamic behaviour.

This course covers everything of the bachelor course Fundamentals of AI (thus, the bachelor course is not required as a prerequesite).

In addition, this course discusses the following basic technologies in greater detail:

1. Logic programming
2. Truth maintenance systems, specially assumption-based truth maintenance systems (ATMS)
3. Model-based diagnosis
4. Ant algorithms

The following technologies not covered in the bachelor course are also covered:

1. Multi-agent technologies
2. Semantic networks

Participants will be eligible and are encouraged to do own research and development, e.g. in a master software project or in a master thesis. There is a tight bound between this course and my research and development fields. These fields are explicitely presented in this course. 

During this course, we will discuss at least the following applications: