Data-assisted Algorithms for Inverse Random Source Scattering Problems - DDPS

Explore data-assisted algorithms for inverse random source scattering problems in this 53-minute webinar presented by Ying Liang for the Data-Driven Physical Simulations (DDPS) series. Learn about a novel approach that utilizes boundary measurement data to reconstruct statistical properties of random sources with fewer realizations than traditional methods. Discover how this technique achieves better reconstruction using only 1/10 of the realizations required by conventional approaches. Compare the performance of various data-driven algorithms, with a focus on Image-to-Image translation methods like pix2pix for reconstructing well-separated inclusions. Gain insights into the stability of this approach with respect to observation data noise and its applications in fields such as antenna synthesis, medical imaging, and earthquake monitoring.

DDPS | Data-assisted Algorithms for Inverse Random Source Scattering Problems by Ying Liang