It has recently been established that the numerical solution of ordinary
differential equations can be posed as a nonlinear bayesian inference problem,
which can be approximately solved via gaussian filtering and
考虑定义在图的顶点上的噪声信号,并就高斯、丢失和均匀分布的噪声提出平滑算法。假设信号服从在频域定义的先验分布,它偏好信号在图的边缘上平滑。通过将此先验分布与三种噪声生成模型配对,我们提出噪声数据存在时真实信号的最大后验估计 (M.A.P.),并提供计算 M.A.P. 的算法。最后,我们展示了这些算法在白噪声恢复图像数据和 toy & EHR 数据中严重丢失时的有效性。