论文标题
无限差反问题的合奏采样器
Ensemble sampler for infinite-dimensional inverse problems
论文作者
论文摘要
我们引入了一个新的马尔可夫链蒙特卡洛（MCMC）采样器，以实现无限维逆问题。我们的新采样器基于仿射不变的集合采样器，该集合采样器使用相互作用的步行者适应目标分布的协方差结构。我们首次将此集合采样器扩展到无限维函数空间，产生了高效的无梯度MCMC算法。由于我们的新合奏采样器不需要梯度或后协方差估计，因此实施和广泛适用。
We introduce a new Markov chain Monte Carlo (MCMC) sampler for infinite-dimensional inverse problems. Our new sampler is based on the affine invariant ensemble sampler, which uses interacting walkers to adapt to the covariance structure of the target distribution. We extend this ensemble sampler for the first time to infinite-dimensional function spaces, yielding a highly efficient gradient-free MCMC algorithm. Because our new ensemble sampler does not require gradients or posterior covariance estimates, it is simple to implement and broadly applicable.
