Welcome to my personal website. The current intend of this page is to give you a rough insight into my scientific interests. In the future, I plan to extend it and present the results of my research findings.
Research Interest: Machine Learning/Meta-Learning/Algorithm-Selection
Machine Learning: Although ML became a mayor hype in recent years, my first contact with it was born out of pure couriosity. Even before going to university, I played around with simple perceptrons and evolutionary online-training in an artificial environment. At that time, I didn't care about efficient training, I was more interested on the idea to train a AI for computer games without having the hassle to implement it on my own. Nowadays, I recognized, that learning to play a game in the most effective way is not the same as developing a fun artificial intelligence to play against in a computer game...
During the early years at the university, I part-time worked as a game programmer. Later, I worked on scientific topics at Fraunhofer IPK (Knowledge Management in industrial use cases via OWL2, Reasoning) and recently, in my master thesis at the DAI-Labor, unsupervised learning on time-series.
After my master degree, I started to work at GT-ARC on the CODA project. CODA's aim is, to provide solutions for algorithm selection and model selection for large classification problems. Simplified: We want to research, which classification algorithms are best to be used on specific tasks and how to tune them in the most efficient way to improve the results.
If you are interested in the basics of algorithm selection, I would like to forward you to the COSEAL Website, where you can find a lot of useful links and basic papers on the topic.
Or you can contact me directly via christian.geissler[AT]gt-arc.com