约束采样的动能 Langevin 分裂格式
Kinetic Langevin Splitting Schemes for Constrained Sampling
Neil K. Chada · Lu Yu
中文摘要
约束采样指从限制在凸集上的分布中生成样本,是计算统计中的重要问题。该工作通过对约束目标的惩罚替代项应用动能 Langevin 动力学的对称 BAOAB 与 ABOBA 分裂,提出新的未调整采样算法,并包含其随机梯度版本。由于惩罚替代项的 Hessian 在约束集边界处不连续,依赖 Lipschitz Hessian 的分析不适用;因此仅基于强凸性与光滑性建立分析,证明两种格式的一步 Wasserstein 收缩、非渐近 Wasserstein 界与复杂度界。随机梯度版本的复杂度较传统 Euler 型离散化显著改进。数值实验在玩具问题与贝叶斯约束回归上验证理论;分裂格式在比 Euler 型方法大数倍的步长下保持稳定,达到给定精度所需梯度评估次数少 5 至 17 倍。
关键要点
- 01问题:凸集约束采样缺乏适用于 Hessian 不连续情形的分裂格式分析。
- 02方法:对惩罚替代项应用 BAOAB/ABOBA 分裂,并给出随机梯度版本。
- 03分析:仅用强凸与光滑性证明一步 Wasserstein 收缩与非渐近界。
- 04结果:随机梯度版本复杂度优于 Euler 型,梯度评估少 5 至 17 倍。
- 05验证:玩具问题与贝叶斯约束回归实验支持理论。
解读
尚无解读。
原始英文摘要
arXiv:2603.23397v3 Announce Type: replace-cross Abstract: Constrained sampling, the task of generating samples from a distribution restricted to a convex set, is an important problem in computational statistics. In this paper, we propose new unadjusted sampling algorithms obtained by applying the symmetric BAOAB and ABOBA splittings, of the kinetic Langevin dynamics, to a penalized surrogate of the constrained target. This also includes their stochastic gradient versions. Since the Hessian of the penalized surrogate is discontinuous across the boundary of the constraint set, analyses relying on a Lipschitz Hessian do not apply. We therefore develop an analysis based only on strong convexity and smoothness, proving one-step Wasserstein contraction schemes, non-asymptotic Wasserstein bounds of both schemes and complexity bounds. The resulting complexities for the stochastic gradient versions improve substantially on traditional Euler-type discretizations. Numerical experiments verify our theory, which is tested on a toy problem and Bayesian constrained regression problem. Overall, the splitting schemes demonstrate stability at step sizes several times larger than Euler-type methods, require five to seventeen times fewer gradient evaluations for a given accuracy.