Master Thesis: AI-Powered Simulations for Brain Tumors

13.03.2025, Diplomarbeiten, Bachelor- und Masterarbeiten

Develop AI-driven solvers using advanced transformer architectures to accelerate brain tumor modeling for personalized radiotherapy planning.

Want to work on cutting-edge AI that makes a real-world impact? Scientific simulations drive breakthroughs in medicine and beyond, but they’re often too slow for real-time use. AI can change that deep learning models can replace expensive computations, making simulations faster while keeping accuracy high.

In this thesis, you will explore advanced transformer architectures to develop AI-driven solvers [1,2], with a focus on brain tumor modeling for personalized radiotherapy planning [3]. Your work could help improve treatment strategies by enabling faster, more accurate irradiation.

What You Need:
• Passion for AI and solving real-world problems.
• Strong Python skills including PyTorch.
• Background in CS, math, physics, or a related field.

What we offer:
• Close mentorship and access to state-of-the-art hardware resources.
• Collaboration with a team of experts in AI, computational science, and beyond.
• Opportunity to publish your work in top-tier venues if results are strong.

This thesis is supervised by Jonas Weidner and Prof. Benedikt Wiestler.

Interested?
Send an email to j.weidner@tum.de, including your CV and bachelor's and master's transcript.

References:
[1] Alkin, Benedikt, et al. "Universal Physics Transformers: A Framework For Efficiently Scaling Neural Operators." The Thirty-eighth Annual Conference on Neural Information Processing Systems.
[2] https://epub.jku.at/obvulihs/download/pdf/11565745
[3] Weidner, Jonas, et al. "Rapid Personalization of PDE-Based Tumor Growth using a Differentiable Forward Model." Medical Imaging with Deep Learning. 2024.

Kontakt: j.weidner@tum.de