Gaussian Process Regression for Surface Interpolation

Explore Gaussian Process Regression (GPR) for surface interpolation in this 48-minute tutorial presented by Zhiqiao Dong and Manan Mehta from the University of Illinois Urbana-Champaign. Learn about the fundamentals of GPR and its applications in manufacturing for generating high-resolution surface estimations from coarser measurement data. Discover a new technique called filtered kriging (FK) that improves interpolation performance using a pre-filter. Follow along with a hands-on demonstration of the Filtered Kriging Lab tool and understand its application to periodic surfaces manufactured by two-photon lithography. Gain insights into spatial interpolation, covariance modeling, and the GPR workflow through practical examples and in-depth explanations. Access additional resources, including the Filtered Kriging Lab tool and related downloads, to further enhance your understanding of this powerful nonparametric regression method.

Syllabus

Gaussian Process Regression for Surface Interpolation
A Motivating Example from Nanomanufacturing
Motivation for Spatial Interpolation
Spatial Interpolation
1-D Example: Motivation
1-D Example: Inference on New Data
1-D Example: Inference on New Data
Gaussian Process GP
Covariance Kernal for GPR
GPR Workflow
Filtered Kriing Lab Demo
Spatial Interpolation Based on GPR
Spatial Interpolation Based on GPR
Spatial Interpolation Based on GPR
Spatial Interpolation Based on GPR
Conventional GPR-Based Methods
Filtered Kriging
Improved Covariance Modeling with FK
Improved Covariance Modeling with FK
Improved Covariance Modeling with FK
Improved Covariance Modeling with FK
Tutorial to Filtered Kriging for Spatial Interpolaton