An introduction to physics-informed neural networks (PINNs), which integrate physical laws expressed as PDEs into neural network training by enforcing differential equation constraints through automatic differentiation.