Detection and Control of Epidemiological Hazards (ESEG)
(2018 - 2021, research project GBA)
Investigation of machine learning techniques that are suitable for identifying epidemiological hazards, such as disease outbreaks, from near-realtime clinical data such as those generated by emergency services in hospitals. Project with Robert-Koch-Institut, Epias (management software for emergency departments), Frankfurt public health authority, over 20 hospitals.
Multilabel Rule Learning
(2018 - 2021, DFG)
Investigation of novel rule learning approaches for solving multilabel classification tasks, a problem class particularly challenging regarding the hidden dependencies between class labels. Rules provide a natural form of expressing such dependencies in a human-comprehensible manner.
Adaptive Preparation of Information from Heterogeneous Sources (AIPHES)
(2015 – 2021, Research Training Group)
Investigation of new concepts and methods for the semi-automatic generation of multi-document summaries. Project between 8 university groups with 11 doctoral positions.
RMU Network for Deep Continuous-Discrete Machine Learning
(2018 – 2020, network project)
Networking project on bridging the worlds of symbolic and knowledge based learning approaches and biologically inspired approaches such as artificial neural networks based on numerical optimization and statistics. Project between six groups of three universities.
Knowledge Discovery in Scientific Literature (KDSL)
(2013 – 2016, DIPF Research Training Group)
Indexing scientific literature through text mining. Project between 9 university groups.
Learning by Pairwise Comparison for Problems with Structured Output Spaces (LPCforSOS)
(2007 – 2012, DFG)
Investigation and development of methods to decompose problems with complex structures into less complex sub-problems that are better understood and easier to solve.
Multilabel Classification Methods for Problems with Large Number of Labels
(2009, DAAD/IKYDA exchange project)
Academic exchange project with Greg Tsoumakas on methods for multi-label data with a large number of labels.
Automated Legal Intelligent System (ALIS)
(2006 – 2009, EU)
Creation of a system that can offer automated and computer-assisted advice or mediation to a user with a legal question. Project between 7 industry partners and 3 academic partners.