Modeling, automatisation and optimisation of experimental processes in network-structured domains.

In a lot of research fields such as biology, medicine and sociology, experiments are used to learn about the domain of interest. Such experiments are often complex processes and the interpretation of their result should happen in the context of the related knowledge in the domain. A major challenge is to perform the research efficiently, using a minimum of resources to achieve a maximum of new knowledge. This project aims at addressing this challenge from a computer science point of view. First, new concepts will be developed for the automation of experimental research. Second, models and algorithms will be developed to choose experiments to optimally increase the available knowledge. These methods will be validated in several applications such as proteomics mass spectrometry and gene knockout experiments.