A0425
Title: Fast computer model calibration using annealed and transformed variational inference
Authors: Won Chang - University of Cincinnati (United States) [presenting]
Jaewoo Park - Yonsei University (Korea, South)
Abstract: Computer models play a crucial role in numerous scientific and engineering domains. To ensure the accuracy of simulations, it is essential to properly calibrate the input parameters of these models through statistical inference. While Bayesian inference is the standard approach, employing Markov Chain Monte Carlo methods often encounters computational hurdles due to the costly evaluation of likelihood functions and slow mixing rates. Although variational inference (VI) can be a fast alternative to traditional Bayesian approaches, VI has limited applicability due to boundary issues and local optima problems. To address these challenges, flexible VI methods are proposed based on deep generative models that do not require parametric assumptions on the variational distribution. A surjective transformation is embedded in the framework to avoid posterior truncation at the boundary. Additionally, theoretical conditions are provided that guarantee the success of the algorithm. Furthermore, the temperature annealing scheme can prevent being trapped in local optima through a series of intermediate posteriors. The method is applied to infectious disease models and a geophysical model, illustrating that the proposed method can provide fast and accurate inference compared to its competitors.
