Thuerey Group at TUM

Our focus is to develop numerical methods for physics simulations with deep learning techniques.

[Thuerey Group at TUM @thuereyGroup]

Time to celebrate our 5 accepted NeurIPS'26 papers 🥳 Covering a wide range of exciting topics: from foundation models, over optimizers, to multi-fidelity learning. Details to follow soon! https://t.co/NRTpdHTBzj

[Thuerey Group at TUM @thuereyGroup]

Ever wondered how to reliably evaluate your stochastic, temporal GenAI-surrogate? 😀 A brilliant Bachelor student at TUM, Sebastian Pfister, has created a stochastic chaotic system benchmark that combines robustness and ease-of-use, StocBench: https://t.co/4orXOL3uq5

[Tailin Wu @tailin_wu]

🚀 The NeurIPS 2026 RealPDE Competition is now live!
Advance Scientific ML for real-world physical systems with paired PIV and CFD data.

🏆 Two tracks:
• Sim2Real
• Long-term Test-time Adaptation

💰 $21,000 prize pool
🎤 NeurIPS workshop presentation opportunities

Join us: https://t.co/pFWRnbdCrL

@NeurIPSConf #NeurIPS2026 #AI4Science #SciML #AI4Physics #AI4PDE

[Thuerey Group at TUM @thuereyGroup]

Congratulations to Mohammad for his ICML paper developing a novel, sensitivity-based approach for generative topology optimization 👍 https://t.co/Gw6t3wqNhY Our work investigates a fundamental question: What determines whether data-driven topology optimization models generalize?

[Thuerey Group at TUM @thuereyGroup]

Congratulations to Bernhard for his SIGGRAPH 2026 Paper 👍 PDF https://t.co/Rpsc8FgsNi , Video https://t.co/lilbSxIMs5

Our key ide: treat particles as samples in four-dimensional space-time. ST-FLIP acts as an effective temporal anti-aliasing mechanism for FLIP-style solvers.

[Thuerey Group at TUM @thuereyGroup]

Congratulations to Hao for his ICML Oral on rotationally equivariant transformers: https://t.co/rhLMdehki3 👍 The key idea is to transform physical fields into local canonical coordinate systems, enabling standard self-attention while preserving physical symmetries

[Thuerey Group at TUM @thuereyGroup]

I'm happy to introduce "CRAFT": a new federated learning optimizer that treats aggregation as a geometric correction problem instead of naive averaging. https://t.co/vOKMKtWFhT

[Thuerey Group at TUM @thuereyGroup]

What if pretraining scientific foundation models didn’t require massive datasets at all?

We show that this is not only possible, but has a range of neat benefits: our "Tadpole" models learn from canonical PDE data that is generated on-the-fly https://t.co/yVWdTaYcqd

[Thuerey Group at TUM @thuereyGroup]

For those who already checked out our AeroTransformer last week: please also try Yunjia's live "WebWing" demo at https://t.co/eil1pLK2yH ✈️

For completeness: full sources and more details available at https://t.co/mZ0SaIhD7W

[Thuerey Group at TUM @thuereyGroup]

I'm very excited to highlight our recent work on how useful NNs are for stability and resolvent analysis of non-linear systems. Chengyun established a firm connection between the theoretical basis and modern AI methods. https://t.co/gUaUPz9lAV

[Thuerey Group at TUM @thuereyGroup]

I'm excited to share our latest work: AeroTransformer — a step toward bringing the foundation model paradigm to real-world aerodynamic design.
Code & models: https://t.co/mZ0SaIhD7W Paper: https://t.co/97Gt64q2m8

[Thuerey Group at TUM @thuereyGroup]

Ever wondered if a Transformer model could successively refine a PDE solution, one scale at a time? 🤔 Mario's latest work shows: a single model auto-regressively infers and refines flow solutions over finer and finer sets of sample points https://t.co/9MCodA9JLK

[Thuerey Group at TUM @thuereyGroup]

I'm excited to highlight the code release of our scalable & efficient PDE Transformer (P3D) at https://t.co/ciFBd3GIiZ Please try out the pretrained models, and let us know how it works! Highlights are, e.g., stable inference of 1024^3 rollouts on a single GPU with 90GB 😁

[Thuerey Group at TUM @thuereyGroup]

We're happy to report that our Physics-based Flow Matching framework got an accept for #ICLR '26! Physics-Based Flow Matching (PBFM) is a principled framework that explicitly targets Pareto-optimal solutions between physics-constraints and data-driven objectives.

[Thuerey Group at TUM @thuereyGroup]

Fast rotational equivariance for physics GNNs — the source code is now available: https://t.co/fwbcQOmdWQ

Please check out the full Physics-of-Fluids paper here: https://t.co/U18FYNqyw1

[Thuerey Group at TUM @thuereyGroup]

In the paper we introduce a scalable hybrid CNN–Transformer architecture that pushes neural surrogate modeling into the regime of truly high-resolution 3D simulations. The P3D architecture overcomes previous memory and compute barriers that have limited Transformers for PDEs.

[Thuerey Group at TUM @thuereyGroup]

We're very excited to report that our P3D Transformer was accepted at ICLR https://t.co/rcsueNzfG1

[Thuerey Group at TUM @thuereyGroup]

I'm very happy to report that our autoregressive predictions with generative diffusion models is _finally_ accepted 😁 Congratulations Georg!

It's been a long journey, this paper was first submitted to NeurIPS'23, and now got into "Neural Networks" https://t.co/UDcE2axYi1

[Thuerey Group at TUM @thuereyGroup]

Great to see our paper on physics-constrained reconstruction / super-res with generative models posted online now at https://t.co/N1qUKlWmw7 😁
- PDE Transformer as backbone architecture
- differentiable physics constraints to guide
- and ConFIG as optimizer to resolve conflicts

[Thuerey Group at TUM @thuereyGroup]

The SuperWing dataset is a large-scale, open dataset of transonic swept-wing aerodynamics, combining thousands of richly parameterized 3D wing geometries with high-fidelity RANS simulations across the operational flight envelope: https://t.co/vdhciXUkno