Neural Canonical Transformation with Symplectic Flows
Shuo-Hui Li, Chenxiao Dong, Linfeng Zhang, Lei Wang
Abstract
A modern machine learning known as normalizing flow can automate cumbersome canonical transformations of Hamiltonian equations, thereby opening up this time-honored technique for studying dynamics to a wide array of complex systems.
Topics & Concepts
Symplectic geometryComputer scienceTransformation (genetics)Symplectic integratorHamiltonian systemHamiltonian (control theory)Canonical transformationHamiltonian mechanicsSymplectomorphismClassical mechanicsMathematicsMathematical optimizationSymplectic manifoldPhysicsPure mathematicsQuantum mechanicsQuantumBiochemistryGeneChemistryPhase spaceModel Reduction and Neural NetworksNeural Networks and ApplicationsComputational Physics and Python Applications