Machine Learning and AI in the Life Sciences: Methods and Applications

This course introduces key machine learning and AI techniques with a strong focus on their applications in the life sciences. You will learn how to analyze complex biomedical data, from pre-processing to model selection and interpretation, using techniques such as dimensionality reduction, feature selection, supervised and unsupervised learning, visual analysis, and deep learning.

Through hands-on projects based on real-world datasets, you will gain practical experience in selecting and applying the right methods to tackle key challenges in the life sciences and bioinformatics, such as disease prediction, biomarker discovery, and patient (status) classification.  These projects will reinforce your learning while also helping you develop both technical expertise and scientific communication skills through presentations.

By the end of the course, you will be able to process and model life science data, evaluate AI-driven insights, and effectively communicate your findings. Prior experience in programming (e.g., in R, Python, Java, or C/C++) is highly recommended.

Dieser Kurs führt in die wichtigsten Techniken des maschinellen Lernens und der künstlichen Intelligenz ein, wobei der Schwerpunkt auf deren Anwendungen in den Lebenswissenschaften liegt. Sie werden lernen, wie man komplexe biomedizinische Daten analysiert, von der Vorverarbeitung bis zur Modellauswahl und Interpretation, unter Verwendung von Techniken wie Dimensionsreduktion, Feature Selection, überwachtes und unüberwachtes Lernen, visuelle Analyse und Deep Learning.

Durch praktische Projekte, die auf realen Datensätzen basieren, werden Sie praktische Erfahrungen in der Auswahl und Anwendung der richtigen Methoden sammeln, um wichtige Herausforderungen in den Lebenswissenschaften und der Bioinformatik zu bewältigen, wie z. B. die Vorhersage von Krankheiten, die Entdeckung von Biomarkern und die Klassifizierung des Patienten(-status). Diese Projekte werden Ihre Kenntnisse vertiefen und Ihnen helfen, sowohl technisches Fachwissen als auch wissenschaftliche Kommunikationsfähigkeiten durch Präsentationen zu entwickeln.

Am Ende des Kurses werden Sie in der Lage sein, biowissenschaftliche Daten zu verarbeiten und zu modellieren, KI-gesteuerte Erkenntnisse zu bewerten und Ihre Ergebnisse effektiv zu kommunizieren. Vorkenntnisse in der Programmierung (z. B. in R, Python, Java oder C/C++) werden dringend empfohlen.

Voraussetzungen:

Vorlesung Statistik im Master Bioinformatik (oder äquivalente Veranstaltungen)

Pre-requisites:
A solid background in differential geometry or geometric computing will be advantageous but is not required.
Students who haven't followed any related courses (Differential Geometry I, Scientific Visualization, ...) can follow the seminar but should be willing to invest more time.

Description:
Geometric deep learning is a broad and emerging research paradigm concerned with the derivation and study of neural network architectures that respect the invariances and symmetries in data.
Indeed, many real-world tasks come with essential pre-defined regularities arising from the underlying low-dimensionality and structure of the physical world.
Capturing these regularities via unified geometric principles has been shown to provide sizable empirical improvements.
Examples of such geometric architectures include graph neural networks as well as models conditioned on data that reside on curved manifolds where vector space operations are not naturally admissible.

The goal of this seminar will be to obtain in-depth knowledge about the core methodology in geometric deep learning as well as an overview of state-of-the-art methods.
Students will acquire practical skills in reading, presenting, explaining, and discussing scientific papers.
The seminar may be used as a preparation for an MSc thesis topic.

Michael M. Bronstein, Joan Bruna, Taco Cohen, Petar Veličković (2021) Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gauges. arXiv:2104.13478

Pre-requisites:
A solid background on geometry processing or differential geometry will be of advantage but is not required. Students who haven't followed any related courses (Differential Geometry I, Scientific Visualization, ...) can follow the seminar but should be willing to invest more time.

Description:
In this seminar we will work on recent topics and results that take its place in the field of geometric statistics: the statistical analysis of data being elements of nonlinear geometric spaces.
Such data arises frequently for example in form of anatomical shapes in medical image processing or human body motion in computer vision.
Often, sizable empirical improvements can be observed when the geometry of the data spaces is incorporated into the design of the model and its algorithmic treatment.
Riemannian geometry has emerged as one of the most powerful mathematical and computational frameworks for analyzing geometric data.
The goal of this seminar will be to obtain an in-depth knowledge about the core methodology in geometric statistics as well as an overview of state-of-the-art methods.
The student will acquire practical skills in reading, presenting, explaining, and discussing scientific papers.
The seminar may be used as a preparation for an MSc thesis topic.

Pennec, X., Sommer, S., & Fletcher, T. (Eds.). (2019). Riemannian geometric statistics in medical image analysis. Academic Press.

Pre-requisites:
Some background on geometry processing or differential geometry will be of advantage but is not required. Students who haven't followed any related courses (Differential Geometry I, Scientific Visualization, ...) can follow the seminar but should be willing to invest more time.

Description:
In this seminar we work on a selection of recent topics and results in one of the areas of medical image analysis targeting geometric data such as shapes, graphs, and abstract data sets alike. The goal will be to obtain in-depth knowledge about an advanced topic on geometric computing for medical applications, in particular in geometric deep learning and manifold-valued statistics. The student will acquire practical skills in reading, presenting, explaining, and discussing scientific papers. The seminar may be used as a preparation for an MSc thesis topic.