Master Seminar (SS 2009)

Machine Learning: Applications and Recent Directions

General Information

  • You find a general course description at the corresponding pages from the WIAI module guide
  • You find  administrative information at UnivIS.
  • Participants should sign up for the course in the virtual campus.
  • This course is open for master and diploma students.
  • Prerequisites: Basic machine learning knowledge as taught in our Machine Learning course (especially the first three lectures) will be helpful.
  • This course will be held as a blockseminar. We agree on a date in the first session.
  • Presentations and theses may be given/written in German or English.
  • First session on Wednesday, April 22, 10 - 12 am, room F380!

Machine Learning

"The field of machine learning is concerned with the question of how to construct computer programs that automatically improve with experience."

(Tom M. Mitchell (1997). Machine Learning. McGraw-Hill.)

Typically, machine learning problems are regarded as search problems: A large space of concepts or functions is searched for one concept or function that best fits the observed training data.  For an example, an artificial neural network (ANN) can be trained to recognize images of faces of various people. A number of images of faces for each person, with varying background, direction in which the person looks etc., is provided as training data. After training, the ANN outputs the identity of the correct person with high accuracy when given a new image of a face.

Forms of representation of learned concepts include, amongst others:

  • Decision trees
  • Artificial neural networks
  • Rule sets
  • Polynoms
  • Bayesian belief networks

General learning algorithm types include supervised learning (pre-classified training data), unsupervised learning (non-classified data, data exploration), and reinforcement learning (policy learning, delayed reward).

Visit the webpage of our Machine Learning course for further general introduction.


Applications for machine learning include natural language processing, medical diagnosis, bioinformatics, detecting credit card fraud, stock market analysis, object recognition in computer vision, game playing, software engineering, robot locomotion, web mining, social network analysis, inductive programming, and many others.

Recent Trends

Classically, data in machine learning consists of independent vectors over unstructured domains. One recent trend is to consider data where the objects have internal structure or are inter-related and linked together in complex graphs.

Another major trend is to use statistical analysis and methodology to machine learning.

Possible Topics:

Possible topics include particular applications or recent methods, techniques and approaches, as mentioned above.


[SS 05] [SS 06] [SS 08]