ISBO: Scalable Spatio-Temporal Bayesian Optimization with Log Gaussian Cox Process Models via the INLA-SPDE Approach
ISBO introduces INLA–SPDE Spatio-Temporal Bayesian Optimization for spatio-temporal data. It uses a LogGaussian Cox Process to model intensity and employs Integrated Nested Laplace Approximation and Stochastic Partial Differential Equations for stable inference. This framework efficiently locates high-intensity regions with few evaluations by combining a scalable surrogate model with Bayesian optimization, proving effective in 2D spatial settings.