CMS Team Project Interactive Visual Computing [CMS-TEA]

• Mats Ole Ellenberg M.Sc., Julián Méndez M.Sc., Dr. rer. nat. Rufat Rzayev
• Prof. Dr.-Ing. Raimund Dachselt

Description

In this course, the students will work on a complex research-related project, which requires competences from several areas of visual computing (CMS-VC-TEA) or life sciences (CMS-CLS-TEA).
The project requires literature research and the use of scientific information sources. The students solve in a group (2-4 persons, depending on the topic and amount of registrations) a complex, typically interdisciplinary task where both 1) professional competence is deepened and 2) extensive methodological and social skills in project management and teamwork are trained.

Topic:

This semester, the project entails the creation of a video-game which must then be played by an artificial intelligence (AI) agent:

• The game can be inspired by existing ones, but should meet a predefined level of complexity and own ideas.
  • A set of initial game recommendations will be provided, but own ideas can be discussed.
• A careful analysis of the potential game inputs and goals must be done, to select an adequate set of AI algorithms that learn and optimize the gameplay.
  • The available AI techniques must be researched and the selection must be thoroughly justified (e.g., evolutionary algorithms, supervised, unsupervised, semi-supervised, reinforcement, incremental, etc.).
• The AI agent must learn to play the game, and ideally become good at it:
  • No hardcoded knowledge of in-game behavior can be provided to the agent, only heuristics.
  • The performance of the AI can be demoed (e.g., by comparing it to human players).

More information will be provided in the Introduction meeting and through the OPAL course.

Proposed Schedule