Accelerating Diffusion Sampling via Speculative Draft Trees
Abstract
Speculative sampling accelerates diffusion model generation by drafting inexpensive candidate states and correcting them under a coupling that preserves the target distribution exactly, reducing the number of expensive target evaluations. Existing diffusion samplers, notably those based on reflection maximal coupling, are topologically constrained: their lookahead drafts form a chain graph, a single linear sequence, which inherently limits the acceptance rate per target evaluation. We connect speculative sampling in diffusion models to relative entropy coding (REC). This perspective shows the lookahead need not be linear and motivates our central contribution, draft trees, which enrich the candidates considered per round and lower the target function evaluations. We further adopt greedy rejection sampling, an REC algorithm, as the draft-target coupling, improving acceptance while guaranteeing exact target samples. Experiments across diverse target and draft models demonstrate up to 8.3% acceleration over the reflection coupling baseline in practical settings.
Related
- Rank-Aware Speculative Sampling for Diffusion Draft Trees
- Accelerating posterior sampling in Bayesian inverse problems with diffusion models: an online-offline perspective
- Training Parallel Speculative Draft Models by Directly Minimizing Expected Decoding Rounds
- Multiple Importance Sampling for Guided Speculative Inference
- Accelerating Non-Smooth and Heavy-Tailed Sampling
BibTeX
@misc{bullo2026accelerating,
title = {{Accelerating Diffusion Sampling via Speculative Draft Trees}},
author = {Marcello Bullo and Yanxiao Liu and Öykü Sıla Güner and Arpan Mukherjee and Deniz Gündüz},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.17691},
eprint = {2609.17691},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.17691},
}
RIS
TY - UNPB
TI - Accelerating Diffusion Sampling via Speculative Draft Trees
AU - Marcello Bullo
AU - Yanxiao Liu
AU - Öykü Sıla Güner
AU - Arpan Mukherjee
AU - Deniz Gündüz
PY - 2026
DA - 2026/09/15
PB - arXiv
DO - 10.48550/arxiv.2609.17691
UR - https://arxiv.org/abs/2609.17691
AB - Speculative sampling accelerates diffusion model generation by drafting inexpensive candidate states and correcting them under a coupling that preserves the target distribution exactly, reducing the number of expensive target evaluations. Existing diffusion samplers, notably those based on reflection maximal coupling, are topologically constrained: their lookahead drafts form a chain graph, a single linear sequence, which inherently limits the acceptance rate per target evaluation. We connect speculative sampling in diffusion models to relative entropy coding (REC). This perspective shows the lookahead need not be linear and motivates our central contribution, draft trees, which enrich the candidates considered per round and lower the target function evaluations. We further adopt greedy rejection sampling, an REC algorithm, as the draft-target coupling, improving acceptance while guaranteeing exact target samples. Experiments across diverse target and draft models demonstrate up to 8.3% acceleration over the reflection coupling baseline in practical settings.
ER -