RAN Newsletter 01/2024Artificial Intelligence and Machine Learning in Heidelberg
Heidelberg has become home to a vibrant scientific community for machine learning and artificial intelligence (AI). Two projects at the Heidelberg Medical Faculty and the Institute of Pharmacy and Molecular Biotechnology at Heidelberg University provide an insight into the potential applications of AI in research.
Given the huge importance of AI, Heidelberg University got off to a slow start with only a handful of scattered labs working in the domain. This changed with the STRUCTURES Cluster of Excellence, which first recognised the need to boost development. In response, the Interdisciplinary Center for Scientific Computing (IWR) decided to establish a new focus on Machine Learning to complement its work on modelling, simulation and optimisation. The Rectorate has established a new professorship for “Mathematical Foundations of Machine Learning” which is being filled right now, and several professorships in mathematics are currently being reoriented towards Machine Learning.
As in other domains, Heidelberg University hugely profits from cooperation with its surrounding institutions, notably the German Cancer Research Center (DKFZ), the European Molecular Biology Laboratory (EMBL), the Heidelberg Institute for Theoretical Studies (HITS) and the Central Institute of Mental Health (CIMH) in Mannheim. Experts from these institutions are working together with labs at Heidelberg University in the context of the AI Health Innovation Cluster and the ELLIS unit Heidelberg, which is part of the European Laboratory for Learning and Intelligent Systems (ELLIS). While further strengthening of the area is needed, Heidelberg now hosts a vibrant Machine Learning and AI community, with strong contributions ranging from foundational research to applications in physics, the life sciences and earth sciences and the humanities.
The research project „Artificial Intelligence for treating Cancer therapy Resistance“ (AI-Care) in which bioinformaticians at Heidelberg University are involved, pursues the goal of forecasting how aggressive brain tumours respond to certain substances with Artificial Intelligence (AI) methods). With the aid of experimental studies in combination with AI-based, mathematical approaches, the scientists want to develop a model with which they can predict the ability of tumour cells to adapt to therapies. These forecasts are intended to avoid possible cases of resistance. The Carl Zeiss Foundation is making five million euros available to fund the research studies for a period of six years. A total of ten research groups are participating in the scientific studies at the three project locations in Heidelberg and Kaiserslautern-Landau.
Highly invasive and aggressive brain tumours, known as glioblastomas, consist of different kinds of cancer cells with a particularly high plasticity. They possess the ability to adapt to therapies and develop resistance, so that conventional treatments like chemo- or radiotherapy prove ineffective, Prof. Dr Carl Herrmann explains. He is head of the Bioinformatics Department at the Institute of Pharmacy and Molecular Biotechnology of Heidelberg University and a partner in the project AI-Care. By combining single cell sequencing technology with AI methods, the participating scientists want to characterise and model the key molecular processes that regulate the plasticity of glioblastomas. Prof. Herrmann’s team will process the data obtained from the single cell analysis of artificially cultivated glioblastoma tissue with the aid of machine-learning methods. In so doing they expect to derive a model able to control the behaviour of cancer cells, predict their response to medication and optimise personalised therapies. The scientists hope that their approach will not only open up new avenues in treating glioblastomas and other types of cancer but also spark new ideas for AI-assisted, customised precision medicine.
Another Heidelberg AI research project to improve prognoses for people with cardiac insufficiency using artificial intelligence will begin work in July 2024. Scientists from Heidelberg and Mainz are collaborating in a joint research project to improve the frequently difficult prognosis of disease progression and hence the treatment options. The goal of the study is to use artificial intelligence methods and robotics to develop individualised therapies for such patients. The project is based jointly at the Medical Faculty Heidelberg of Heidelberg University and at the University Medical Center Mainz, the two lead partners, and it will receive five million euros in funding from the Carl Zeiss Foundation over a period of six years.
Entitled “Multi-dimensionAI: linking scales of information to improve care for patients with heart failure”, the project focuses on a patient group suffering from a frequent form of chronic cardiac insufficiency. This involves a stiffening of the left ventricle although a sufficient volume of blood is ejected. According to the experts, there are no standard forms of treatment able to reverse the changes in the heart muscle and improve the prognosis of the patients. Left untreated, the result might eventually be heart failure, underlines Junior Professor Dr Sandy Engelhardt from the Medical Faculty Heidelberg. She is the project’s -spokesperson and, with her working group “Artificial Intelligence in Cardiovascular Medicine”, does research in the clinical Departments of Cardiology, Angiology and Pneumology and Cardiac Surgery at Heidelberg University Hospital.
The researchers want to train artificial intelligence (AI) multimodally with the health data of several thousand patients.“We hope that the AI support will enable us to select therapies and evaluate their benefits in a much more targeted manner in future,” says Prof. Engelhardt. As a practical example of an AI-based treatment recommendation, an exercise therapy study is to be conducted, developed and supervised by the sports medicine departments of Heidelberg University Hospital and Mainz University. The patients, who soon get out of breath due to their cardiac insufficiency, will be helped to exercise by wearing a personalised exosuit. It was designed by a team around Prof. Dr Lorenzo Masia, head of the Biorobotics and Medical Technology group at the Institute of Computer Engineering at Heidelberg University.