arXiv:2611.26407v1 Announce Type: cross
Abstract: Flow matching is a versatile framework for generative modeling that transports a source distribution to a target data distribution. A key advantage is its flexibility in the choice of source distribution, such as heavy-tailed and multimodal sources. Yet finite-step sampling guarantees for non-Gaussian sources remain limited. We establish a nonasymptotic convergence theory for Euler discretization of flow matching with linear interpolation. Suppose the source distribution on $\mathbb{R}^d$ has an everywhere positive, continuously differentiable density, finite total variance $\mathsf{TotVar}_{\mathsf{base}}$, and an $L$-Lipschitz score. Using the population velocity field and a suitable time grid, this sampler achieves total variation error at most $\varepsilon$ in $\widetilde{O}(Ld\mathsf{TotVar}_{\mathsf{base}}/\varepsilon)$ iterations, provided the target distribution satisfies a mild second-moment condition. These assumptions cover, among many examples, source distributions such as multivariate Student’s $t$ distributions with more than two degrees of freedom and finite Gaussian mixtures with shared covariance, while accommodating target distributions with neither a smooth density nor bounded support.
Trending
- Is It Actually Worth Upgrading To The Pixel 11?
- Flow Matching from Non-Gaussian Sources: A Theory of Efficient Sampling
- How To Calibrate Your 3D Printer (And Why You Should)
- The Stability Gap in Continual Reinforcement Learning
- This Is What Can Happen If Too Many Devices Are Sharing Your Home Network
- A Large Language Model-based Agentic Framework Accelerating Structural Relaxation and Electronic Band Prediction for First-Principle Calculation
- Your PC’s USB Port Can Charge Your Phone, Even When It’s Turned Off — Here’s How
- [2611.26232] Time Series Structure Learning with Local Causal Parent Discovery
