![[2512.05251] One-Step Diffusion Samplers via Self-Distillation and Deterministic Flow](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2512.05251] One-Step Diffusion Samplers via Self-Distillation and Deterministic Flow
Summary
The paper presents a novel one-step diffusion sampler that utilizes self-distillation and deterministic flow to enhance sampling efficiency in machine learning, achieving high-quality results with fewer computational steps.
Why It Matters
This research addresses the computational inefficiencies in existing sampling algorithms, which often require numerous iterations to produce reliable samples. By introducing a method that reduces the number of steps needed, it has significant implications for speeding up various applications in machine learning and statistics, particularly in scenarios where computational resources are limited.
Key Takeaways
- Introduces a one-step diffusion sampler that enhances sampling efficiency.
- Achieves competitive sample quality with significantly fewer network evaluations.
- Utilizes self-distillation and deterministic flow for improved ELBO estimates.
- Addresses the limitations of existing sampling algorithms in terms of computational cost.
- Demonstrates effectiveness across synthetic and Bayesian benchmarks.
Statistics > Machine Learning arXiv:2512.05251 (stat) [Submitted on 4 Dec 2025 (v1), last revised 25 Feb 2026 (this version, v2)] Title:One-Step Diffusion Samplers via Self-Distillation and Deterministic Flow Authors:Pascal Jutras-Dube, Jiaru Zhang, Ziran Wang, Ruqi Zhang View a PDF of the paper titled One-Step Diffusion Samplers via Self-Distillation and Deterministic Flow, by Pascal Jutras-Dube and 3 other authors View PDF HTML (experimental) Abstract:Sampling from unnormalized target distributions is a fundamental yet challenging task in machine learning and statistics. Existing sampling algorithms typically require many iterative steps to produce high-quality samples, leading to high computational costs. We introduce one-step diffusion samplers which learn a step-conditioned ODE so that one large step reproduces the trajectory of many small ones via a state-space consistency loss. We further show that standard ELBO estimates in diffusion samplers degrade in the few-step regime because common discrete integrators yield mismatched forward/backward transition kernels. Motivated by this analysis, we derive a deterministic-flow (DF) importance weight for ELBO estimation without a backward kernel. To calibrate DF, we introduce a volume-consistency regularization that aligns the accumulated volume change along the flow across step resolutions. Our proposed sampler therefore achieves both fast sampling and stable evidence estimate in only one or few steps. Across challenging s...