Lixiang Gu

Zhenyu He · Jobs Stroustrup

Basic Info

Name: Lixiang Gu (顾理想)
Affiliation: Peking University (colleague in the DSCOVR single-pixel Earth-retrieval collaboration; co-author on a paper draft together with Yongyun Hu, Caltech’s Yuk Yung, and UC Riverside’s King-Fai Li)
Research areas: inverse-problem and Bayesian inversion methodology; DSCOVR single-pixel “Earth-as-an-exoplanet” retrievals

Collaboration Summary

I collaborated with Lixiang Gu in the DSCOVR Bayesian retrieval group during 2021–2022. The Hardened Balancing Principle (HBP) regularization criterion that I formulated and wrote up was initially developed for PC2 (the land/ocean contrast mode); Lixiang Gu subsequently extended the same criterion to retrievals on PC1 (low-cloud mode) and PC4 (high-cloud mode), a downstream methodological extension within the same group of co-authors.

The collaboration produced a draft paper, “An automatic pipeline for retrieval from DSCOVR data” (draft, not publicly published), with authorship order Lixiang Gu, Zhenyu He, Siteng Fan* (corresponding), King-Fai Li, Yuk L. Yung, Yongyun Hu.

Research Direction

Application of inverse-problem regularization methods to satellite single-pixel remote-sensing retrievals; side-by-side comparison of HBP, L-curve, GCV, and Bayesian criteria to characterize the failure modes of each.