Events I have been invited to present at, shared along with slides and other resources.
This paper proposes an SSGL method that decomposes spatially varying coefficients into global effects and selectable spatial deviations, accurately identifies which predictors require spatial variation, and improves estimation efficiency and computation relative to Bayesian spatial competitors, while simultaneous uncertainty calibration for highly localized effects remains challenging.
A scalable Bayesian SVC framework that performs predictor-level selection of entire coefficient surfaces through a spike-and-slab group lasso (SSGL) prior on basis expansion coefficients
Derive general density for scale mixtures of complex Gaussians and use them as shrinkage priors in Bayesian complex linear regression
A Comparative Study of Bayesian Neural Networks and Machine Learning Based on COVID-19 Image Classification
Bayesian hierarchical modeling for latency estimation of event-related potentials