Distributed Mesh Processing and Geometric Deep Learning for Wind Prediction

Distributed Mesh Processing and Geometric Deep Learning for Wind Prediction

Explore a technical conference talk showcasing Neural Concept and RWDI's revolutionary approach to urban design through advanced wind prediction technology. Learn how Ray's distributed computing capabilities handle massive Computational Fluid Dynamics simulations containing over 10 million mesh nodes, while discovering the seamless integration of complex mesh preprocessing with model inference through Ray Serve. Follow along as Saswata Chakravarty addresses the challenges of managing large input mesh sizes and implementing sophisticated systems on Kubernetes, offering valuable insights into applying AI techniques for architectural challenges that are transforming urban planning and design processes.