Recasting diffusion noise schedule design as optimal control on Fisher information yields sufficient conditions for O(d/n) sampling error and parametric closed-form schedules that generalize exponential/sigmoid ones and improve empirical performance.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
The authors establish a criterion for strong entropy chaos on path space in the infinite-particle limit for conservative diffusions, given relative entropy bounds, weak convergence of drifts and fixed-time marginals, and regularity of the limit drift.
citing papers explorer
-
Noise Schedule Design for Diffusion Models: An Optimal Control Perspective
Recasting diffusion noise schedule design as optimal control on Fisher information yields sufficient conditions for O(d/n) sampling error and parametric closed-form schedules that generalize exponential/sigmoid ones and improve empirical performance.
-
A criterion for proving entropy chaos on path space
The authors establish a criterion for strong entropy chaos on path space in the infinite-particle limit for conservative diffusions, given relative entropy bounds, weak convergence of drifts and fixed-time marginals, and regularity of the limit drift.