This talk delves into the rules that dictate how systems evolve over time, using what's known as ordinary differential equations (ODEs). These equations provide a blueprint for understanding changes in systems. To kick things off, we'll see an introduction to how these "rules" are deduced from observed behaviors. The talk will encompass various methods, from the more abstract blackbox models to the transparent symbolic techniques. A core aspect of this discussion is understanding the reliability and accuracy of these methods, particularly when faced with both familiar and new scenarios. Finally recent attempts to extract these equations from imperfect real-world data are described.
Niki Kilbertus investigates the interactions between machine learning algorithms and humans with a focus on ethical consequences and trustworthiness. He currently studies identification and estimation of causal effects from observational data in automated decision-making and dynamic environments.
After studying mathematics and physics at the University of Regensburg, Niki Kilbertus obtained his PhD in machine learning in 2020 from the University of Cambridge (UK) in a joint program with the Max Planck Institute for Intelligent Systems. Since 2020 he is a Young Investigator Group Leader with Helmholtz AI at the Helmholtz Center Munich. Since 2021 he is Assistant Professor at TUM and continues to lead his group at Helmholtz AI.
Please help us plan ahead by registrating for the event at our
After the event, there will be a social get-together with food and drinks courtesy of the Division of Medical Image Computing and Interactive Machine Learning Group at the DKFZ.
|What?||Learning Dynamical Laws from Data|
|Who?||Niki Kilbertus, TU Munich|
|When?||September 20th 2023 @ 4pm|
|Where?||DKFZ Communication Center (seminar rooms K1+K2), Im Neuenheimer Feld 280|