Machine Learning in Physics
The application of artificial intelligence, neural networks, and statistical learning models to solve complex physical equations and discover empirical patterns in experimental data.
Governing formula
Physics-Informed Neural Networks (PINN): Loss = L_data + λ L_physics
SI unit
Computational AI Framework
In depth
Physics-Informed Neural Networks (PINNs) embed differential conservation laws directly into neural network loss functions, guaranteeing that AI predictions strictly satisfy physical conservation laws.
Examples in the real world
Accelerating quantum chemistry electronic structure calculations and predicting plasma instabilities in fusion tokamaks.