ISBO: Scalable Spatio-Temporal Bayesian Optimization
What it is
ISBO is a Bayesian Optimization framework designed for spatio-temporal data that models the log-intensity using a Log-Gaussian Cox Process (LGCP). It leverages the INLA-SPDE approach to perform fast and accurate posterior inference on sparse Gaussian Markov Random Fields, enabling efficient optimization of expensive black-box objectives in point-process spaces.
The problem it solves
Standard Gaussian Process-based Bayesian Optimization is ill-suited for doubly stochastic Cox Processes and struggles with the scalability and inference requirements of spatio-temporal problem spaces.
What you could use it for
- Optimizing sensor placement in spatio-temporal monitoring networks
- Identifying high-intensity regions in epidemiological or environmental point-process data
- Efficiently locating peaks in latent intensity fields with minimal function evaluations
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.