Lower Bounds for Parallel Diffusion Sampling
Abstract
Standard diffusion samplers generate samples through sequential evaluations of a learned score function, with each query depending on previous answers. Parallel samplers seek to accelerate this process by making multiple score queries simultaneously. We ask how many adaptive rounds are necessary when each round allows polynomially many queries. We establish the first polynomial lower bounds on the round complexity of diffusion sampling with approximate scores satisfying a mean-square accuracy guarantee. Specifically, we prove (1) a Ω̃(d1/3)-round lower bound for sampling smooth, near-isotropic Gaussian mixtures in ℝd, and (2) a tight Ω(d)-round lower bound for uniform sampling from anisotropic axis-aligned boxes contained in the unit ball. Both bounds hold for arbitrary randomized algorithms making polynomially many queries per round at arbitrary locations and noise levels, with inverse-polynomial score error and constant total variation accuracy. Our near-isotropic construction follows the blockwise approach of earlier lower bounds for parallel matroid search and discrete sampling \cite{KUW85,AGR24}. We use independent coordinate blocks containing random codebooks with well-separated rates and construct approximate score oracles whose responses in each block depend on consistency tests in earlier blocks. Together, the rate separation and these tests ensure that each round reveals information about at most one additional block with high probability.
Related
- Serial Depth of Adaptive Parallel Sampling: Lower Bounds and Schedules for Masked Diffusion Models
- Query-Efficient Rejection Sampling for Parallel Masked Diffusion
- Lower bounds for heat kernels of elliptic operators with unbounded diffusion, drift, and potential terms
- pper and Lower Bounds for Blow-up Times in Coupled Reaction--Diffusion Systems with Nonlocal Sources
- Exponential Sampling Lower Bounds for Polynomial Sources
BibTeX
@misc{kou2026lower,
title = {{Lower Bounds for Parallel Diffusion Sampling}},
author = {Yiwen Kou and Yimeng Wang},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.09166},
eprint = {2610.09166},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.09166},
note = {CC BY 4.0},
}
RIS
TY - UNPB
TI - Lower Bounds for Parallel Diffusion Sampling
AU - Yiwen Kou
AU - Yimeng Wang
PY - 2026
DA - 2026/10/06
PB - arXiv
DO - 10.48550/arxiv.2610.09166
UR - https://arxiv.org/abs/2610.09166
AB - Standard diffusion samplers generate samples through sequential evaluations of a learned score function, with each query depending on previous answers. Parallel samplers seek to accelerate this process by making multiple score queries simultaneously. We ask how many adaptive rounds are necessary when each round allows polynomially many queries. We establish the first polynomial lower bounds on the round complexity of diffusion sampling with approximate scores satisfying a mean-square accuracy guarantee. Specifically, we prove (1) a Ω̃(d^1/3)-round lower bound for sampling smooth, near-isotropic Gaussian mixtures in ℝ^d, and (2) a tight Ω(d)-round lower bound for uniform sampling from anisotropic axis-aligned boxes contained in the unit ball. Both bounds hold for arbitrary randomized algorithms making polynomially many queries per round at arbitrary locations and noise levels, with inverse-polynomial score error and constant total variation accuracy. Our near-isotropic construction follows the blockwise approach of earlier lower bounds for parallel matroid search and discrete sampling \cite{KUW85,AGR24}. We use independent coordinate blocks containing random codebooks with well-separated rates and construct approximate score oracles whose responses in each block depend on consistency tests in earlier blocks. Together, the rate separation and these tests ensure that each round reveals information about at most one additional block with high probability.
ER -