Location: MIC-Arena, Maria-Goeppert-Str. 3, 23562 Lübeck
Title: Machine Learning for biomedical data analysis of complex data
Abstract: Modern computer systems are fast and work in high resolutions. They acquire massive amounts of data that have to be analyzed. This so-called "big data" hype can be found literally everywhere as it affects most computer applications, ranging from functional brain imaging to a simple internet search request.
The major issue is that just having lots of data is not the same as understanding it. Therefore, for every new "big data" problem, new (or at least adapted) tools are needed to let machines learn something meaningful from enormous data sets. Especially multivariate learning approaches are very sensitive to non-random noise that regularly appears in big data sets.
Within the last years, we have applied machine learning techniques on various biomedical data problems. In this talk, I will specifically focus on two problems, multimodal cell tracking using geometric hashing, and the encoding of early olfactory processing revealed by non-negative matrix factorization.