Inverse Gaussian Process regression for likelihood-free inference


Speaker:    Jinglai Li, University of Birmingham

Time:         16:00-17:00 , May.20

Location:   Tencent Meeting

    Meeting ID926-418-725


Host:          Yue Qiu



In this work the speaker considers Bayesian inference problems with intractable likelihood functions. They present a method to compute an approximate of the posterior with a limited number of model simulations. The method features an inverse Gaussian Process regression (IGPR), i.e., one from the output of a simulation model to the input of it. Within the method, they provide an adaptive algorithm with a tempering procedure to construct the approximations of the marginal posterior distributions. With examples they demonstrate that IGPR has a competitive performance compared to some commonly used algorithms, especially in terms of statistical stability and computational efficiency, while the price to pay is that it can only compute a weighted Gaussian approximation of the marginal posteriors.



Jinglai Li received the B.Sc. degree in Applied Mathematics from Sun Yat-sen University in 2002 and the PhD degree in Mathematics from SUNY Buffalo in 2007. After his PhD degree, Jinglai did postdoctoral research at Northwestern University (2007-2010) and MIT (2010-2012) respectively. He subsequently worked at Shanghai Jiao Tong University (Associate Professor, 2012-2017) and University of Liverpool (Reader, 2017-2020). Jinglai joined the University Birmingham as a Professor in 2020. His research interest lies in scientific computing, computational statistics, uncertainty quantification, and data science.