We study the Ornstein-Uhlenbeck-Kac (OU-K) process, where instead of a Brownian motion, a finite-speed Kac process defines the driving process. The OU-K process can be analyzed through a rescaled, time-inhomogeneous Kac process, characterized by a switching rate from the Kac part, and a convergence rate from the Ornstein-Uhlenbeck (OU) part. The law of this inhomogeneous Kac process describes the solution of a wave equation, the Euler-Poisson-Darboux (EPD) equation. For Kac switching rates being multiples of the OU convergence rate, we can explicitly derive this EPD solution via the sum of (the derivative of) two Legendre polynomials, restricted between two Dirac points representing the wave fronts. Using the framework of piecewise-deterministic Markov processes (PDMP), we describe the conditional velocity fields of the OU-K process via our polynomial-law-formula. This allows us to apply the OU-K process in a generative flow model. Since the OU-K velocity fields prove to be bounded by the finite Kac speed, the usual time truncations of diffusion models become unnecessary. We demonstrate advantages over diffusion and flow matching models in terms of generation sharpness on singular data sets, while achieving competitive results on CIFAR-10 image generation.
Figures & tables
Figure 1: Generated samples after 15.000 (left), 30.000 (middle) and 100.000 (right) training steps of a polytope in dimension d=32 via the OU diffusion baseline (top) and our OU-K model (bottom), orthogonally projected onto R2 . By its better regularity, the OU-K model is able to approximate the vertices notably more precisely after significantly fewer training steps of the neural network.
Figure 2: The balanced AUC sharpness metric of the data sets Di , i=1,2,3,4 , in dimension d=128 , as a function of the number of training steps, comparing the OU-K model with the FM, OU and Kac baselines.
FID ↓
FM
OU
Kac
OU-K
Euler 100
4.01
4.15
5.05
4.35
Euler 200
3.59
3.50
4.10
3.56
Table 1: CIFAR-10 FID scores evaluated at 50k samples. Lower score is better.
Appendix figures & tables1 asset
Supplementary material from the paper’s appendix.
Appendix
Figure 3: CIFAR-10 generation results of the FM, OU and Kac baselines and the OU-K model.
Sampling from Flow Matching generative models requires solving an ordinary differential equation (ODE) whose computational cost is dominated by neural network forward passes. We derive four classical ODE solvers -- Euler, Explicit Midpoint, Classical Runge-Kutta (RK4), and Dormand-Prince 5(4) -- from first principles via Taylor expansion, implement them from scratch in PyTorch, and systematically benchmark their efficiency on Conditional Flow Matching tasks ranging from 2D toy distributions to MNIST digits. On the quantitative side, we use sliced Wasserstein distance to construct NFE-quality Pareto frontiers,finding that RK4 at 80 function evaluations achieves sample quality comparable to Euler at 200. Beyond reproducing known convergence rates, we report two empirical observations: (1) the Jacobian eigenvalue spectrum of the learned velocity field stiffens sharply near t=1, explaining why the adaptive Dormand-Prince solver automatically concentrates its step budget at the end of the trajectory; (2) the quality gap between low-order and high-order solvers widens for undertrained and smaller models, indicating that solver choice matters most when the model is imperfect. Code and all experiment scripts are publicly available.
Diffusion models and flow-based methods have shown impressive generative capability, especially for images, but their sampling is expensive because it requires many iterative updates. We introduce W-Flow, a framework for training a generator that transforms samples from a simple reference distribution into samples from a target data distribution in a single step. This is achieved in two steps: we first define an evolution from the reference distribution to the target distribution through a Wasserstein gradient flow that minimizes an energy functional; second, we train a static neural generator to compress this evolution into one-step generation. We instantiate the energy functional with the Sinkhorn divergence, which yields an efficient optimal-transport-based update rule that captures global distributional discrepancy and improves coverage of the target distribution. We further prove that the finite-sample training dynamics converge to the continuous-time distributional dynamics under suitable assumptions. Empirically, W-Flow sets a new state of the art for one-step ImageNet 256×256 generation, achieving 1.29 FID, with improved mode coverage and domain transfer. Compared to multi-step diffusion models with similar FID scores, our method yields approximately 100× faster sampling. These results show that Wasserstein gradient flows provide a principled and effective foundation for fast and high-fidelity generative modeling.
We study dynamic measure transport for generative modeling, focusing on transport maps that connect a source measure P0 to a target measure P1 by integrating a velocity field of the form vt(x)=E[X˙t∣Xt=x], where X∙=(Xt)t is a stochastic process satisfying (X0,X1)∼P0⊗P1 and X˙t is its time derivative. We investigate when X∙ induces a straight-line flow: a flow whose pointwise acceleration vanishes and is therefore exactly integrable by any first-order method. First, we develop multiple characterizations of straight-line flows in terms of PDEs involving the conditional statistics of the process. Then, we prove that straight-line flows under endpoint independence exhibit a sharp dichotomy. On the one hand, we construct explicit, computable straight-line processes for arbitrary Gaussian endpoints. On the other hand, we show that straight-line processes do not exist for targets with sufficiently well-separated modes. We demonstrate this obstruction through a sequence of increasingly general impossibility theorems that uncover a fundamental relationship between the sample-path behavior of a process with independent endpoints and the space-time geometry of this process' flow map. Taken together, these results provide a structural theory of when straight-line generative flows can, and cannot, exist.
Panos Tsimpos, Daniel Sharp, Youssef Marzouk
Operations Research Center · Massachusetts Institute of Technology · Cambridge, MA 02139 +2