Zoloflow:基于椭圆插值调度的最优采样算法
Elliptic Sampling: Optimal Interpolant Schedules for Gaussians and Beyond
做扩散模型采样的人应该看看这篇:它把插值调度和 ODE 离散化一起优化,给出椭圆调度族 Zoloflow,理论上对高斯目标拿到 minimax 最优的 score 查询次数。如果你一直在用经典的三角调度,这篇解释了什么时候、怎么换成椭圆调度能省采样步数,还附了图像生成的实证对比。
一篇新论文提出 Zoloflow 采样算法,核心是用保方差的椭圆插值调度(elliptic schedule),由 Jacobi 椭圆函数参数化,通过共形映射解析求解。该调度与离散化方案显式引入目标分布的条件数 κ,并在 κ→1 时退化为经典三角调度。
理论结果方面:对高斯目标,Zoloflow 在所有 oblivious score-span 算法中达到 minimax 最优,以 Θ(log κ · log(√d/ε)) 次精确 score 查询将总变差误差控制在 ε;对 log-concave 分布,用 Õ(√d·polylog(κ,1/ε)) 次 score 评估即可达到 ε 的 2-Wasserstein 误差。
该框架基于复数化 ODE 轨迹的解析性质设计调度与离散化,与 Zolotarev 最优有理函数逼近密切相关。作者还在合成基准和现有图像生成数据集上与非 log-concave 目标的其他调度方案做了对比实验。
Elliptic Sampling: Optimal Interpolant Schedules for Gaussians and Beyond
Generative modeling based on stochastic interpolants requires specifying two key choices: an interpolant schedule bridging the initial and target distributions, and an efficient discretization of the resulting probability-flow ODE. In this work, we introduce a new framework for joint schedule/discretization design based on the analyticity properties of the (complexified) ODE trajectories, which in turn ensure good numerical behavior. Our formulation can be solved exactly via conformal mapping, yielding a natural family of variance-preserving elliptic schedules, parametrized by Jacobi elliptic functions. Pairing an elliptic schedule with a tailored discretization, we obtain Zoloflow, a new sampling algorithm closely connected to Zolotarev's optimal rational function approximants. Crucially, both the elliptic schedule and Zoloflow explicitly incorporate the condition number $κ$ of the target distribution and reduce to the classical trig schedule in the limit $κ\rightarrow 1$. For Gaussian targets, Zoloflow achieves minimax optimality among all oblivious score-span algorithms, attaining total-variation error $\varepsilon$ using $Θ(\logκ\,\log(\sqrt d/\varepsilon))$ (exact) score queries. For the log-concave case, the elliptic schedule allows sampling with at most $\varepsilon$ 2-Wasserstein-error by requiring merely $\widetilde{O}(\sqrt{d} \, \mathrm{polylog}(κ, 1/\varepsilon))$ score evaluations. Complementing these theoretical results, we empirically evaluate the elliptic schedule on general (non-log-concave) target distributions, including detailed comparisons with alternative schedule proposals on both synthetic benchmarks and existing image generation datasets.
This report is categorized under AITOP's "Papers" desk. Key entities and technical topics under continuous tracking include Zoloflow, diffusion models, sampling algorithms, arXiv papers.
Source attribution: curated from arXiv cs.AI (@Cole Becker, Taylan Kargin, Kimia Hassibi, Asuman Ozdaglar, Pablo A. Parrilo).