06. Trends in AI Based Design Research - Dynamic Prediction & Multi-Fidelity Learning in Engineering

Namwoo Kang, CEO of Narnia Labs and Professor at KAIST, covers two of the more advanced challenges in predictive AI for engineering: moving from static to dynamic prediction using graph neural networks, and bridging the gap between simulation and real-world test data through multi-fidelity learning.
He walks through Google DeepMind's mesh-based simulation model, how message passing in graph networks enables time-series prediction, and six practical methods for combining low and high fidelity data, including physics-informed neural networks, transfer learning, delta learning, and multi-fidelity machine learning.

Learn more at 👉 https://www.narnia.ai/

#PredictiveAI #EngineeringDesign #ManufacturingAI

Видео 06. Trends in AI Based Design Research - Dynamic Prediction & Multi-Fidelity Learning in Engineering канала Narnia Labs