Physics-Informed Neural Networks (PINNs)
Neural networks trained not just on labeled data, but also constrained to satisfy physical laws (governing PDEs, boundary conditions). The training process incorporates domain knowledge about conservation laws.
Why it matters
Requires fewer training examples and generalizes better to unseen conditions. Suitable for data-scarce engineering domains.
This term appears in
Cite this definition
Finocchiaro, Michael. “Physics-Informed Neural Networks (PINNs).” DemystifyingPLM PLM Glossary, 2026, https://www.demystifyingplm.com/glossary/physics-informed-neural-networks-pinns