Funding in the Millions for New AI Research Group

The AI research groups aim to closely integrate AI methods with research fields that use these approaches for interdisciplinary themes. Thus, the integration of AI into basic research and the scientific research of AI itself go hand in hand. The new research group called FOR 5347 Lifespan AI will initially receive funding of more than four million euros for four years.

“I am very pleased about the DFG’s approval,” says Professor Tanja Schultz, the spokesperson for Lifespan AI and a professor at the University of Bremen’s Faculty of Mathematics/Computer Science. “In Lifespan AI, we aim to develop AI methods and tools that model, predict, and explain the development of diseases over the course of someone’s life.” Co-spokesperson Professor Marvin N. Wright from the Leibniz Institute for Prevention Research and Epidemiology – BIPS adds: “We will do this by drawing on high-dimensional life span data compiled from longitudinal epidemiological studies. The data is supplemented by biological, social, and lifestyle information.”

Professor Jutta Günther, Vice President for Research and designated President, is also pleased about the establishment of the high-ranking research group at the University of Bremen and the comprehensive third-party funding that it brings. “In this endeavor, cutting-edge technologies are developed and used to generate scientific knowledge for the benefit of humanity, exemplifying the university’s commitment to answering questions about the future of our society,” Günther says. Professor Iris Pigeot, who is the BIPS Director, adds: “The research group is an excellent example of how fruitful collaboration is between the university and non-university research institutions, in this case with the Leibniz Institute for Prevention Research and Epidemiology – BIPS.”

Lifespan AI: Here’s What It’s All about

In Lifespan AI, sensitive data is used in compliance with ethical and privacy conditions to drive machine learning (ML) and deep learning (DL) models. The aim is to gain causal insights to uncover the causes of complex diseases and optimize prevention strategies.

The work program consists of six projects grouped into three themes that pursue the Lifespan AI vision from different perspectives: data and methods (D), models and interpretation (M), and inference and causality (C). D1 will advance DL strategies to explore and process long-term temporal change based on the integration of high-dimensional data from multiple sources; D2 will combine neural networks and mixed-effects models to predict individual health trajectories over the course of someone’s life; M1 will develop “normalizing flow” methods to derive joint distributions and conditional densities for health data; M2 will create a cognitive digital twin from everyday human activities to predict change across age groups; C1 will develop time-adaptive, explainable AI methods for recurrent neural networks and event times; and C2 will derive a framework for “causal discovery” in longitudinal studies, combining different data sets and accounting for nonlinearities.

Long-Standing Research Cooperation Forms the Basis

The research group is based on the long-standing institutional research cooperation between the University of Bremen and BIPS. It is supported by ten applicants from these two institutions, who together cover the central disciplines: Professor Michael Beetz (AI & Robotics, University of Bremen), Dr. Claudia Börnhorst (Epidemiology, BIPS), Professor Werner Brannath (Statistics & Biometry, University of Bremen), Professor Vanessa Didelez (Causal Inference, BIPS), Professor Horst Hahn (Medical Imaging & AI, University of Bremen), Professor Peter Maaß (Mathematical Analysis of ML, University Bremen), Professor Iris Pigeot (Statistics & Epidemiology, Director of BIPS), Dr. Felix Putze (Adaptive Interaction Systems, University of Bremen), Professor Tanja Schultz (Cognitive Systems & ML, University of Bremen), and Professor Marvin N. Wright (Statistical Learning, BIPS). The research group is also supported by the new cooperative professorship in “Machine Learning in Statistics” at the interface between AI and epidemiology, which serves as a bridge professorship and is filled by Professor Marvin N. Wright.

In addition, the Lifespan AI research group has arranged for Professor Haizhou Li to be a Mercator Fellow. Li has already been closely cooperating with the University of Bremen for years as a professor at the National University of Singapore (NUS), the Chinese University of Hong Kong (CUHK), and as U Bremen Excellence Chair. International collaborators Dr. Rudi GJ Westendorp from the University of Copenhagen and Dr. Michael Wand from the Swiss AI Lab IDSIA are also involved.

The total funding amount will predominantly be used to finance nine PhD student positions and the establishment of an IT infrastructure that will be used jointly by the University of Bremen and BIPS to support the computationally intensive DL models.

Contact:

Professor Tanja Schultz

Faculty of Mathematics/Computer Science

University of Bremen

Tel.: +49 421 218-64270

Email: tanja.schultz@uni-bremen.de