Kristina Yordanova (Institut für Data Science): Extracting Dynamic Semantic Structures from Heterogenous Data

In this talk we present the research at the newly founded Institute of Data Science, especially artificial intelligence methods that combine symbolic modelling and probabilistic reasoning with machine learning approaches in order to model, learn and infer the state of dynamic systems. Furthermore, we look at methods for learning the model structure from textual sources and sensor observations. Application domains are behaviour monitoring and analysis and dynamic user support with examples in the domain of people with cognitive impairments, healthy ageing, and intelligent students tutoring.