Sem­inar Ma­chine Learn­ing / Or­ganic Com­put­ing

Description / Procedure:

Given a sufficient number of participants, the seminar takes place in the form of a conference, i.e.,

  • Participants can choose a topic from the field of machine learning / organic computing from a suggestion list or suggest a suitable topic themselves.
  • A first version of the paper must be submitted by the middle of the semester.
  • The submitted paper will be reviewed in a peer review by the seminar participants.
  • The participants then have the opportunity to revise their paper to make a final version (camera ready).
  • The results and topics of the elaborated papers will be presented by the participants at the end of the semester as part of a small "conference". The lectures should last a maximum of 25 min., followed by a short round of questions (about 5 min.).
  • In order to give the paper a uniform appearance, it should be based on the well-known LNCS style. The use of latex is recommended and supported. A template for Microsoft Word is also available.
  • The focus of the seminar for master students is on the independent development of a current research topic. For undergraduate students, the focus is more on scientific writing as such (structure, content).

Advantages of the Seminar Organization 

Subject selection is free within the scope of the advertised subject area (In addition, a selection of individual topics is offered).

  • Participants get an insight into the processes at academic conferences.
  • Good preparation for upcoming theses, especially if there is no previous knowledge in LaTex or scientific writing.
  • Guided approach to the individual stages of a scientific elaboration.

Choice of topics  

The topics of the seminar papers should come from the area "Machine Learning" or "Organic Computing".

Bachelor participants: The seminar focuses on teaching the basics of scientific work with topics in the respective area (for example, chapters from a textbook).
Sample topics:

  • Support Vector Machines
  • clustering
  • Reinforcement Learning
  • ant algorithms
  • decision trees
  • particle swarm
  • Artificial immune systems
  • ... [own suggestions]

Master participants: The seminar focuses on research-oriented work based on current research topics in the respective field.

Sample topics:

  • Active Learning
  • emergence
  • Relevance Vector Machines
  • Deep Learning
  • SVM
  • Dirichlet Process Mixture Models
  • Nonparametric Bayesian
  • ... [own suggestions]