{"id":"arxiv/2610.24295","url":"https://arxiv.org/abs/2610.24295","tags":["stat.CO","math.OC"],"score":0,"title":"PICKLES: Covariance-Preconditioned Kinetic Langevin Ensemble Sampler","domain":"stat.CO","source":"arxiv","authors":["Hong Ye Tan","Stanley Osher","Yifan Chen"],"reviews":0,"subject":"stat.CO","summary":"We propose a covariance-preconditioned kinetic Langevin ensemble sampler named PICKLES. This is a kinetic analogue of the affine invariant interacting Langevin dynamics (ALDI). PICKLES uses a rectangular covariance square root as a preconditioner to avoid computing matrix inverses or square roots, leading to position and momentum variables of different dimensions. Theoretically, the continuous-time dynamics are ergodic under mild regularity and growth conditions on the potential, assuming $N>d+1$ particles and a full-rank initial ensemble. We discretize PICKLES using BAOAB or OBABO…","category":"research","comments":0,"disputed":false,"division":"stat","mentions":0,"discussed":0,"published":"2026-10-16","reactions":{}}