Model-Aware Schedules Improve Generation via Fiberwise Optimal Transport
Quick summary
arXiv:2609.11842v1 Announce Type: cross Abstract: Diffusion and flow-matching schedules control the signal and noise coefficients that mix data and noise along affine probability paths. Minimizing a kinetic action defined on coefficient paths, motivated by optimal transport, helps explain strong baselines but remains model-agnostic and ignores prediction error. Here we introduce a model-aware schedule construction based on fiberwise optimal transport. At a fixed time and state on the probability path, compatible signal/noise decompositions form an affine fiber. We define a fiberwise prediction
Key takeaways
- arXiv:2609.11842v1 Announce Type: cross Abstract: Diffusion and flow-matching schedules control the signal and noise coefficients that mix data and noise along affine probability paths.
- Minimizing a kinetic action defined on coefficient paths, motivated by optimal transport, helps explain strong baselines but remains model-agnostic and ignores prediction error.
- Here we introduce a model-aware schedule construction based on fiberwise optimal transport.
Why it matters
“Model-Aware Schedules Improve Generation via Fiberwise Optimal Transport” should be evaluated beyond branding and benchmark scores. Its practical importance will emerge in task accuracy, latency, unit cost, safety and integration with real workflows.
Member comments