We want to raise the power of scientific discovery by thoughtful application of machine learning techniques - closely working together with the Machine Learning researchers and the research community of the University, spanning the natural sciences, the social sciences and the humanities.
We are building a team to tackle this challenge from different angles.
- we develop, implement and deploy probabilistic models for varied research problems, from reconstructing bone surfaces from point-cloud data, to finding parameters for complex and realistic physical simulations, to classifying microscopic pollen grains, to analysing ancient texts.
- we train and advise researchers on the best methods to obtain maximum insight from their data, from feature selection to model evaluation.
- we assess best practices in scientific machine learning, and share our progress with the community in conventional as well as interactive formats.
- finally, we distill recent literature into open-source machine learning code to facilitate realistic and unbiased algorithm benchmarking and to empower researchers across disciplines.
Plus d’informations :
[Website Tübingen University]