In this seminar, our SmallData Associated Researchers will give a 30 minute presentation of their work on small data and explore how it connects with other projects across the Collaborative Research Center. Each seminar will feature an open Q&A session. A talk title and short description will be available closer to the time.
Towards a data and biology driven characterization of MS
Understanding heterogeneity in multiple sclerosis (MS) requires analytical frameworks capable of integrating sparse, irregular and highly multimodal clinical data with highly granular, multidimensional data from blood and CSF cytometric and sequencing analyses. In this lecture, we present biology-driven and probabilistic machine-learning approaches that redefine MS phenotyping beyond traditional clinical subtypes.Together, these approaches outline a data-driven path towards individualized prediction and biology-informed precision neurology.
Heinz Wiendl (SmallData Associated Researcher)
Department of Neurology and Neurophysiology, Medical Center – University of Freiburg
Mathematical deep learning for high dimensional systems in small data regimes
Many scientific and biomedical problems require modelling complex stochastic systems despite having only limited observational data. In this talk, I will present mathematical results that offer some insights toward reliable learning in such small-data settings. I present our work showing that deep neural networks and ResNet-type architectures can approximate solutions of high-dimensional Kolmogorov partial differential equations with only polynomial growth in dimension and accuracy, thus overcoming the classical curse of dimensionality. These findings provide rigorous foundations for using structure-informed neural networks in applications governed by stochastic dynamics. The talk aims to illustrate how mathematical insights from stochastic analysis and approximation theory can support deep robust learning.
Diyora Salimova (SmallData Associated Researcher)
Department for Applied Mathematics, University of Freiburg
Institute of Medical Biometry and Statistics,
Faculty of Medicine and Medical Center –
University of Freiburg
