Sebastian Weichwald, adjunkt

Siden han kom til KU som postdoc i september 2019, har Sebastian fået et ry som en inspirerende og god lærer. Han stod i spidsen for det CoCaLa-team, der vandt NeurIPS 2019 Causality 4 Climate -konkurrencen.

“During my time as a postdoc here at UCPH, I have worked on and supervised students on projects, inspired by our winning strategy for the NeurIPS competition, that reveal limitations of common additive noise model simulations for benchmarking causal structure learning algorithms. Insights which may hint at limited real-world plausibility of commonly used model classes in causal modelling rather than limitations of structure learning algorithms”, siger Sebastian.

“I will pursue conceptual work in that realm to further probe the plausibility and real-world applicability of models where with causal noise, measurement noise, or noise stemming from aggregating finely-resolved models into macro-models”.

Sebastian er uddannet fra University of Tübingen (BSc matematik) og fra UCL (MSc Computational Statistics and Machine Learning). Han afsluttede sin ph.d. ved Max Planck Institute for Intelligent Systems, Tübingen, med en afhandling om "Pragmatism and Variable Transformations in Causal Modelling".

Du kan møde Sebastian på kontor 4.3.14: ”I'm always happy to hear about and collaborate on real-world applications of statistical causal modelling that practitioners face or, more broadly, to discuss the quirks and limitations of causality frameworks.”