A Dual-Dimer method for training physics-constrained neural networks with minimax architecture
Dehao Liu, Yan Wang
Topics & Concepts
Artificial neural networkMinimaxComputer scienceSaddle pointGradient descentConvergence (economics)Dual (grammatical number)Eigenvalues and eigenvectorsFunction (biology)Artificial intelligenceMathematical optimizationTraining (meteorology)AlgorithmMathematicsPhysicsLiteratureQuantum mechanicsMeteorologyArtEvolutionary biologyEconomicsGeometryBiologyEconomic growthModel Reduction and Neural NetworksAdvanced Image Processing TechniquesHeat Transfer and Optimization