Score-based Neural Ordinary Differential Equations for Computing Mean Field Control Problems
This provides a computational method for high-dimensional mean field control problems, which is incremental as it builds on existing neural ODE frameworks.
The paper tackles the mean field control problem with individual noises by reformulating it as an unconstrained optimization problem using a system of neural ODEs for first- and second-order score functions, and demonstrates effectiveness in examples like regularized Wasserstein proximal operators and linear quadratic problems.
Classical neural ordinary differential equations (ODEs) are powerful tools for approximating the log-density functions in high-dimensional spaces along trajectories, where neural networks parameterize the velocity fields. This paper proposes a system of neural differential equations representing first- and second-order score functions along trajectories based on deep neural networks. We reformulate the mean field control (MFC) problem with individual noises into an unconstrained optimization problem framed by the proposed neural ODE system. Additionally, we introduce a novel regularization term to enforce characteristics of viscous Hamilton--Jacobi--Bellman (HJB) equations to be satisfied based on the evolution of the second-order score function. Examples include regularized Wasserstein proximal operators (RWPOs), probability flow matching of Fokker--Planck (FP) equations, and linear quadratic (LQ) MFC problems, which demonstrate the effectiveness and accuracy of the proposed method.