WRSWorld Research StateContribute
WORLD RESEARCH STATE

MeanFlow: average-velocity learning for one-step generation

Contribution grade

Generative-model training · Human-designed research method or resource

Reference score120,000Provisional editorial estimate of this exact contribution; its grade is assessed separately.

What changed

Modeled time-interval average velocity and trained against an identity linking it with instantaneous velocity, enabling a from-scratch one-step flow generator.

Scientific scope

Modeled time-interval average velocity and trained against an identity linking it with instantaneous velocity, enabling a from-scratch one-step flow generator. Inherits flow matching and two-time/flow-map models. Source's field identity and empirical procedure are not independently proof-audited or convergence-certified here. Conditional straight paths do not imply straight marginal flows; one-step does not mean exact data-distribution sampling.

Evidence

The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Inherits flow matching and two-time/flow-map models. Source's field identity and empirical procedure are not independently proof-audited or convergence-certified here. Conditional straight paths do not imply straight marginal flows; one-step does not mean exact data-distribution sampling.

Contributors

Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.

Full scientific record · Browse contribution grades · Request a correction