Machine Learning Talks on Campus

NCT Data Science/ELLIS Life/Heidelberg.AI Seminar 

Beyond Supervised Learning: Exploring Novel Machine Learning Approaches For Robust Medical Image Analysis
Prof. Dr. Bernhard Kainz. SMIEEE

Machine learning has been widely regarded as a solution for diagnostic automation in medical image analysis, but there are still unsolved problems in robust modelling of normal appearance and identification of features pointing into the long tail of population data. In this talk, I will explore the fitness of machine learning for applications at the front line of care and high throughput population health screening, specifically in prenatal health screening with ultrasound and MRI, cardiac imaging, and bedside diagnosis of deep vein thrombosis. I will discuss the requirements for such applications and how quality control can be achieved through robust estimation of algorithmic uncertainties and automatic robust modelling of expected anatomical structures. I will also explore the potential for improving models through active learning and the accuracy of non-expert labelling workforces.

However, I will argue that supervised machine learning might not be fit for purpose, as it cannot handle the unknown and requires a lot of annotated examples from well-defined pathological appearance. This categorization paradigm cannot be deployed earlier in the diagnostic pathway or for health screening, where a growing number of potentially hundred-thousands of medically catalogued illnesses may be relevant for diagnosis.

Therefore, I introduce the idea of normative representation learning as a new machine learning paradigm for medical imaging. This paradigm can provide patient-specific computational tools for robust confirmation of normality, image quality control, health screening, and prevention of disease before onset. I will present novel deep learning approaches that can learn without manual labels from healthy patient data only. Our initial success with single class learning and self-supervised learning will be discussed, along with an outlook into the future with causal machine learning methods and the potential of advanced generative models.

https://dkfz-de.zoom.us/meeting/register/cDpIIUzwRsCtZ2x6Pw8zgw#/registration

Machine learning galore!

Lab presentations
Stephanie Hansmann-Menzemer
Jan Stühmer
Frank Zöllner

Rocket science
Christoph Langenbruch (Hansmann-Menzemer lab)
Flavourful Machine Learning at LHCb 

Leif Seute (Stühmer lab)
Generative Machine Learning for the Design of Dynamical Proteins

Speaker-Name follows (Zöllner lab)
 

To help plan the catering, please register for free by April 22th! 

3rd SIMPLAIX Workshop on Machine Learning for Multiscale Molecular Modeling

The Third SIMPLAIX Workshop on “Machine Learning for Multiscale Molecular Modeling” is jointly organized by SIMPLAIX and the RTG 2450.

The first two SIMPLAIX workshops took place in May 2023 and May 2024 (https://simplaix-workshop2023.h-its.org/ and https://simplaix-workshop2024.h-its.org/). Similarly, the aim of this third workshop is to bring together scientists working in this vibrant field to share their research and discuss current challenges in an informal atmosphere.

Confirmed speakers:

• Matteo Dal Peraro, École Polytechnique Fédérale de Lausanne, Switzerland
• Arne Elofsson, Stockholm University, Sweden
• Shirin Faraji, University of Duesseldorf, Germany
• Stefan Grimme, University of Bonn, Germany
• Johannes Kaestner, University of Stuttgart, Germany
• Sandra Luber. University of Zurich, Switzerland
• Antonia Mey, University of Edinburgh, UK
• Carolin Mueller, University of Erlangen, Germany
• Modesto Orozco, Institute for Research in Biomedicine (IRB), Barcelona, Spain
• Elsa Sanchez Garcia, TU Dortmund University
• Lukas Stelzl, Institute of Molecular Biology, Mainz, Germany

Registration for this event is now https://simplaix-workshop2025.h-its.org/registration/.

SIMPAIX and RTG2450
Rebecca Wade (HITS), Marcus Elstner (KIT), Tristan Bereau (Heidelberg University), Pascal Friederich (KIT), David Hoffmann (KIT), Rostislav Fedorov (HITS), Daniel Sucerquia (HITS), Jonathan Teuffel (HITS).

2nd Sorbonne-Heidelberg Workshop on AI in medicine: Machine Learning for multi-modal data

Machine Learning is transforming science, especially the way we do research in medicine. It can analyze non-linear dependencies of structured clinical data, and it is starting to support in the huge amount of existing text and other unstructured information to extract useful information using recent techniques based on large language models. There is also an increasing amount of specific omics data for each patient, which makes it hard to manually inspect all the details. This is where multimodal data analysis comes in, which is the focus of this year's AI in Medicine workshop. Researchers from Sorbonne and Heidelberg will give keynote speeches to provide insight into their research field, which will fuel discussions.

IWR School on Machine Learning for Fundamental Physics

Machine Learning is here to stay and is shaping the future of fundamental physics research. From optimal inference, over theory-inspired network architectures, to anomaly detection, representation learning and foundation models, a new generation of scientists is driving these exciting developments. This school aims to further strengthen technical expertise and foster new connections.

The 2025 IWR School on Machine Learning for Fundamental Physics is aimed at advanced PhD students specializing in scientific machine learning. We particularly encourage registrations from researchers with experience in scientific machine learning, as demonstrated by papers or preprints related to the topic of the school. The school takes place at the Interdisciplinary Center for Scientific Computing (IWR) at Heidelberg University September 15th-19th 2025.