报告人：赵学彬博士

报告题目：Uncertaintiesin seismic imaging and interrogation problems

摘要：The goal of seismic imaging is to create high resolution maps of spatial variations of the Earth's interior, extending in two or three spatial dimensions. Given that seismic imaging often yields non-unique solutions, it is important to find all possible images that are consistent with observed data to estimate the corresponding uncertainties. Typically, this can be achieved under a probabilistic framework by using Bayes' rule to calculate the so-called posterior (post-experimental) probability distribution function – a process referred to as Bayesian inference or Bayesian inversion. In this talk, I will introduce variational inference and show how this method can be used to solve seismic imaging problems and to assess the corresponding uncertainties efficiently and accurately. Finally, I will use the imaging results to interrogate subsurface structures by answering specific scientific questions.

报告时间：2024年7月11日14：00

报告形式：线下：科创K1344；线上：#腾讯会议：729-739-067

报告人简介：赵学彬，英国爱丁堡大学博士后研究员，博士生导师，2023年至今任英国石油（BP）地震成像研究中心访问学者。2023年获爱丁堡大学地质与地球物理学博士学位，2016年和2019年分别获得中国石油大学（北京）学士和硕士学位。赵学彬的主要研究方向包括：地震成像，贝叶斯反演，不确定性分析，以及机器学习在地球物理反演中的应用。迄今为止在Journal of Geophysical Research：solidEarth，Geophysical Journal International和Geophysics等地球物理领域知名期刊发表学术论文20余篇，并长期担任上述期刊的审稿人。其中一篇第一作者论文被评选为JGR: solid Earth主编高亮研究，并在EoS网站进行专题报道。其研究成果在学术界和工业界取得广泛应用。