Abstract:Normalizing flows have been successfully modeling a complex probability distribution as an invertible transformation of a simple base distribution. However, there are often applications that require more than invertibility. For instance, the computation of energies and forces in physics requires the second derivatives of the transformation to be well-defined and continuous. Smooth normalizing flows employ infinitely differentiable transformation, but with the price of slow non-analytic inverse transforms. In this work, we propose diffeomorphic non-uniform B-spline flows that are at least twice continuously differentiable while bi-Lipschitz continuous, enabling efficient parametrization while retaining analytic inverse transforms based on a sufficient condition for diffeomorphism. Firstly, we investigate the sufficient condition for Ck-2-diffeomorphic non-uniform kth-order B-spline transformations. Then, we derive an analytic inverse transformation of the non-uniform cubic B-spline transformation for neural diffeomorphic non-uniform B-spline flows. Lastly, we performed experiments on solving the force matching problem in Boltzmann generators, demonstrating that our C2-diffeomorphic non-uniform B-spline flows yielded solutions better than previous spline flows and faster than smooth normalizing flows. Our source code is publicly available at <a class="link-external link-https" href="https://github.com/smhongok/Non-uniform-B-spline-Flow" rel="external noopener nofollow">this https URL</a>.

A gradient flow method for smooth splines versus least-squares fitting on Riemannian manifolds

On a Complex Hessian Flow

Flow by Gauss Curvature to the Aleksandrov and Dual Minkowski Problems

A Note on the Mean Curvature Flow in Riemannian Manifolds

Convergence Analysis of Gradient Algorithms on Riemannian Manifolds Without Curvature Constraints and Application to Riemannian Mass

Hyperbolic Gradient Flow: Evolution of Graphs in R^n+1

Linear Convergence of Subgradient Algorithm for Convex Feasibility on Riemannian Manifolds

Gradient flows in asymmetric metric spaces and applications

A probabilistic method for gradient estimates of some geometric flows

Smoothing splines on Riemannian manifolds, with applications to 3D shape space

Mean Curvature Flow of Arbitrary Codimension in Complex Projective Spaces

A gradient flow on control space with rough initial condition

A Riemannian Gradient Sampling Algorithm for Nonsmooth Optimization on Manifolds

Discretized Gradient Flow for Manifold Learning in the Space of Embeddings

Hessian Based Smoothing Splines for Manifold Learning

Generalized Ricci Flow I: Local Existence and Uniqueness

Neural Diffeomorphic Non-uniform B-spline Flows

Numerical approximation of boundary value problems for curvature flow and elastic flow in Riemannian manifolds

Singularity Models for High Codimension Mean Curvature Flow in Riemannian Manifolds

Stable discretizations of elastic flow in Riemannian manifolds

Some new gradient estimates for two nonlinear parabolic equations under Ricci flow