Deep Explanation of a Physics Transformer for Aerodynamics

Not everyone has access to a mentor or a co-worker that can explain the most technical topics (in this case an AI model) in the simplest terms. After having the pleasure of hearing Mr. Johannes Brandstetter lecture on this model last week, I wanted to share a synthetic podcast that is based on their paper:

“AB-UPT: Scaling Neural CFD Surrogates for High-Fidelity Automotive Aerodynamics Simulations via Anchored-Branched Universal Physics Transformers”

🔗Link to paper

I hope you enjoy!

physics transformer, neural CFD surrogate, aerodynamic simulation, scientific machine learning, AI for aerodynamics