Setting Things Straight: On the Straightness and Non-Straightness of Two-Step Rectified Flow
Abstract
Rectified flow (RF) is an ODE-based generative modeling framework that has achieved remarkable success in practice over the past few years. By design, RF seeks to produce flows that have a high degree of straightness, in the sense that the resulting ODE trajectories can be obtained with relatively few (ideally, one) sampling steps. However, a rigorous analysis of straightness or near-straightness of RF flows remains missing in the literature. In this paper, we study settings in which the coupling generated by two-step RF is provably straight at the population level. We provide a complete characterization in two popular classes of data distributions: (i) smooth, strongly log-concave densities and (ii) Gaussian mixture models. Our main results yield sharp conditions for the coupling generated by two-step RF to be straight. Non-straight couplings are constructed when these conditions fail, which demonstrate the sharpness of these conditions. Using the theory developed on the straightness of two-step RF, we propose a notion of approximate straightness that measures the straightness of the learned couplings and can be estimated from the data. We describe a procedure for constructing asymptotic confidence intervals for our proposed notion of approximate straightness and propose a data-driven method for selecting the number of reflow rounds in multiple iterations of RF.
Related
- ReCIRC: Rectified Conformal Risk Control
- Computational and Statistical Guarantees of the c-Rectified flow
- A Sharp Bound of Crossing Limit Cycle for Planar piecewise linear Systems with a Straight Switching Line
- Realizing graph automorphisms as geometric symmetries of straight line drawings
- Short Recurrences and Straight-Line Programs for Primes
BibTeX
@misc{pham2026setting,
title = {{Setting Things Straight: On the Straightness and Non-Straightness of Two-Step Rectified Flow}},
author = {Tuan Pham and Saptarshi Roy and Alessandro Rinaldo and Purnamrita Sarkar},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.36090},
eprint = {2610.36090},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.36090},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Setting Things Straight: On the Straightness and Non-Straightness of Two-Step Rectified Flow
AU - Tuan Pham
AU - Saptarshi Roy
AU - Alessandro Rinaldo
AU - Purnamrita Sarkar
PY - 2026
DA - 2026/10/26
PB - arXiv
DO - 10.48550/arxiv.2610.36090
UR - https://arxiv.org/abs/2610.36090
AB - Rectified flow (RF) is an ODE-based generative modeling framework that has achieved remarkable success in practice over the past few years. By design, RF seeks to produce flows that have a high degree of straightness, in the sense that the resulting ODE trajectories can be obtained with relatively few (ideally, one) sampling steps. However, a rigorous analysis of straightness or near-straightness of RF flows remains missing in the literature. In this paper, we study settings in which the coupling generated by two-step RF is provably straight at the population level. We provide a complete characterization in two popular classes of data distributions: (i) smooth, strongly log-concave densities and (ii) Gaussian mixture models. Our main results yield sharp conditions for the coupling generated by two-step RF to be straight. Non-straight couplings are constructed when these conditions fail, which demonstrate the sharpness of these conditions. Using the theory developed on the straightness of two-step RF, we propose a notion of approximate straightness that measures the straightness of the learned couplings and can be estimated from the data. We describe a procedure for constructing asymptotic confidence intervals for our proposed notion of approximate straightness and propose a data-driven method for selecting the number of reflow rounds in multiple iterations of RF.
ER -