23
3

Normalizing flows as approximations of optimal transport maps via linear-control neural ODEs

Abstract

The term "Normalizing Flows" is related to the task of constructing invertible transport maps between probability measures by means of deep neural networks. In this paper, we consider the problem of recovering the W2W_2-optimal transport map TT between absolutely continuous measures μ,νP(Rn)\mu,\nu\in\mathcal{P}(\mathbb{R}^n) as the flow of a linear-control neural ODE. We first show that, under suitable assumptions on μ,ν\mu,\nu and on the controlled vector fields, the optimal transport map is contained in the Cc0C^0_c-closure of the flows generated by the system. Assuming that discrete approximations μN,νN\mu_N,\nu_N of the original measures μ,ν\mu,\nu are available, we use a discrete optimal coupling γN\gamma_N to define an optimal control problem. With a Γ\Gamma-convergence argument, we prove that its solutions correspond to flows that approximate the optimal transport map TT. Finally, taking advantage of the Pontryagin Maximum Principle, we propose an iterative numerical scheme for the resolution of the optimal control problem, resulting in an algorithm for the practical computation of the approximated optimal transport map.

View on arXiv
Comments on this paper