LeMuR is an MSCA (Marie Skłodowska-Curie Actions) Doctoral Network (DN) 2021 on the topic of "Learning with Multiple Representations". The aim of LeMuR is to develop the theoretical foundations and a first set of algorithms for the new paradigm "Learning with Multiple Representations" (LMR). In addition, corresponding applications will be developed to demonstrate the usefulness of the new approaches.
In particular, LMR algorithms will enable flexible representations (e.g. suitable for explainability, fairness, ...) with multiple objective functions (e.g. including environmental or even social impact) so that the induced models meet the Green Charter and trustworthy AI criteria a priori. The project focuses on weak supervision learning as it addresses one of the major shortcomings of modern ML approaches, namely their data starvation, by using weaker sources for labelling training data. The outcome of the DN will be a group of 10 experts trained to implement the third and subsequent waves of AI in Europe. The highly interdisciplinary and cross-sectoral context in which they will be trained will provide them with research-related and transferable skills relevant for a successful career in key AI fields.
More about the project of our Data Science workgroup can be found here.
