Improved training of physics-informed neural networks for parabolic differential equations with sharply perturbed initial conditions
Yifei Zong, Qizhi He, Alexandre M. Tartakovsky
Topics & Concepts
ResidualNormalization (sociology)MathematicsParabolic partial differential equationGaussianArtificial neural networkDegenerate energy levelsGradient descentError functionApplied mathematicsMathematical analysisDifferential equationPhysicsAlgorithmComputer scienceQuantum mechanicsMachine learningAnthropologySociologyModel Reduction and Neural NetworksNuclear reactor physics and engineeringFluid Dynamics and Turbulent Flows