A Multiscale Strategy for Bayesian Inference Using Transport Maps

A Multiscale Strategy for Bayesian Inference Using Transport Maps
复制标题

使用传输图进行贝叶斯推理的多尺度策略

DOI:
10.1137/15m1032478
复制
发表时间:
2015
期刊:
SIAM/ASA J. Uncertain. Quantification
影响因子:
--
通讯作者:
Y. Marzouk
Y. Marzouk
中科院分区:
--
文献类型:
--
作者:
M. Parno;T. El;Y. Marzouk

文献摘要

参考文献

被引文献

相似文献

在许多反问题中,模型参数不能从观测数据中精确确定。贝叶斯推理提供了一种用于捕获所得到的参数不确定性的机制,但通常具有较高的计算成本。这项工作介绍了一种多尺度分解,利用跨尺度的条件独立性,当存在于某些类别的反问题时,将贝叶斯推理解耦为两个阶段:(1)计算上易于处理的粗尺度推理问题,以及(2)低维粗尺度后验分布到原始高维参数空间的映射。这种分解依赖于通过最佳传输映射的粗尺度和细尺度量的非高斯联合分布的表征。我们证明了我们的方法在地下水流中产生的一系列反问题,使用多尺度有限元方法离散的稳态压力方程。我们比较了多尺度战略与全维…
In many inverse problems, model parameters cannot be precisely determined from observational data. Bayesian inference provides a mechanism for capturing the resulting parameter uncertainty, but typically at a high computational cost. This work introduces a multiscale decomposition that exploits conditional independence across scales, when present in certain classes of inverse problems, to decouple Bayesian inference into two stages: (1) a computationally tractable coarse-scale inference problem, and (2) a mapping of the low-dimensional coarse-scale posterior distribution into the original high-dimensional parameter space. This decomposition relies on a characterization of the non-Gaussian joint distribution of coarse- and fine-scale quantities via optimal transport maps. We demonstrate our approach on a sequence of inverse problems arising in subsurface flow, using the multiscale finite element method to discretize the steady state pressure equation. We compare the multiscale strategy with full-dimensiona...
DOI: 10.1002/9780470685853
发表时间: 1994
期刊: --
影响因子: --
作者:
L. Biegler;G. Biros;O. Ghattas;M. Heinkenschloss;D. Keyes;B. Mallick;Y. Marzouk;L. Tenorio;B. V. B. Waanders-B.-V.-B.-Waanders-1863062;K. Willcox
通讯作者: L. Biegler;G. Biros;O. Ghattas;M. Heinkenschloss;D. Keyes;B. Mallick;Y. Marzouk;L. Tenorio;B. V. B. Waanders-B.-V.-B.-Waanders-1863062;K. Willcox