Internship Project
Biology and Life Sciences

Explainable AI and Causality for Digital Pathology

Institution
Charité – Universitätsmedizin
Institute of Pathology
Subject Area
Biology, Medicine, Mathematics, Data Science, AI, Engineering & Technology
Availability
03 May – 23 July
17 May – 06 August
31 May – 20 August
14 June – 03 September
 
Internship Modality:
On-site internship in Berlin

Project Supervisor(s)
Prof. Dr. Teresa G. Krieger
Amos Muench (PhD student)
Academic Level
  • Advanced undergraduate students (from the second year)
  • Master’s students
  • Doctoral candidates
Language
English
Further Information
Project Type
Academic Research
Project Content
The project is situated at the intersection of multimodal AI, spatial transcriptomics, and explainable AI. We aim to investigate foundation models that can predict spatial transcriptomics from routine H&E images. Building on existing work, we want to use explainability methods to explore biological networks that underpin morphology and gene expression.

Subject areas and working fields:
  • Biology (Genetics, Spatial Transcriptomics, Biological Networks)
  • Life Sciences & Medicine (Cancer Genetics)
  • Mathematics, Computer Science & IT (Structural Causal Models, Topology, Graph Theory)
  • Data Science / AI (Explainable AI, Knowledge Graphs, Deep Learning, PyTorch)
  • Engineering & Technology (Software Engineering, High Performance Computing)



Tasks for Interns
The internship will entail both theoretical and practical work, including literature research and design of experiments, implementation, and evaluation of novel explainability methods on large biomedical datasets in a high-performance computing environment.

Academic Level
  • Advanced undergraduate students (from the second year)
  • Master’s students
  • Doctoral candidates
Requirements
To join this project team, you should have a background in mathematics and deep learning or a strong quantitative science background.
Furthermore, you should be interested in biological networks and explainability, and be familiar with PyTorch and Git.

Expected Preparation
Readings:
Concept-based explanation of gene expression prediction from H&E images
Muench et al. 2026
https://doi.org/10.48550/arXiv.2608.16669

DeepSpot-M: a multimodal foundation model for transcriptome-wide virtual spatial transcriptomics from histology
Nonchev et al. 2026
https://doi.org/10.64898/2026.06.19.26356060

Zurück zur Projektliste

For more information on the Humboldt Internship Program or the project, please contact the program coordinator.