Gaussian-Channel Regularity and Distribution Learning with Flow Matching
Organizations: Department of Applied Mathematics, The Hong Kong Polytechnic University, Hong Kong SAR, China · School of Mathematics and Statistics and Hubei Key Laboratory of Computational Science, Wuhan University, Wuhan, China · School of Mathematics and Statistics, Northeast Normal University, Changchun, China
Abstract
We study distribution learning by flow matching with linear interpolation from a Gaussian source under Gaussian-channel regularity, a condition on the target distribution. For an integer , we establish spatial regularity of order for the population velocity, together with temporal Lipschitz continuity on the full interval and local time analyticity before the terminal time. We use these properties to construct a finite multiscale ReLU sieve with a spatial Lipschitz bound independent of the truncation radius. For an exact minimizer of the empirical flow matching risk over this sieve, we obtain an upper bound on the expected Wasserstein- error of order , up to a logarithmic factor, where is the sample size and is the dimension. The bound includes Euler discretization up to time one, without early stopping. The conditions hold at for targets with strongly convex potentials and two-sided Hessian bounds. They also hold at every fixed order for distributions obtained by convolving a law of bounded support with a nondegenerate isotropic Gaussian. For sufficiently large fixed constants defining the target class, we prove a minimax lower bound of order .
Figures & tables
| Condition | Consequence used in the proof |
|---|---|
| Real-tilt covariance bound, | Centered posterior concentration, velocity derivatives on , and target and response tails. |
| Linear growth of | Linear velocity growth, the remainder in the posterior mean near , and the continuous endpoint value. |
| Bounds on and | Spatial Lipschitz and first-order time control, together with the covariance scale needed for complex continuation. |
| Bounds on , | Spatial Taylor accuracy of order , hence the spatial approximation exponent. |
| Result | Polynomial rate | Guarantee |
|---|---|---|
| Kunkel (2026) , Theorem 4.3 | High-probability | |
| Kunkel and Trabs (2026) , Remark 3.8(3) | Expected | |
| Stéphanovitch et al. (2025) , Theorem I ( ) | Expected | |
| Theorem 3.1 , | Expected and high-probability |
Flow-Based Generative Modeling
Flow-based generative modeling learns probability transports from simple noise to data for fast sampling, density estimation, and controllable generation.
See all 609 papers