Variational Inference for Neyman-Scott Processes
This addresses the computational bottleneck for researchers and practitioners using NSPs in fields like spatial statistics or event modeling, though it is an incremental improvement by applying an existing method (VI) to a new model.
The paper tackled the slow inference problem in Neyman-Scott processes (NSPs) by developing the first variational inference (VI) algorithm, which generates samples much faster than traditional MCMC and achieves better prediction performance under time constraints.
Neyman-Scott processes (NSPs) have been applied across a range of fields to model points or temporal events with a hierarchy of clusters. Markov chain Monte Carlo (MCMC) is typically used for posterior sampling in the model. However, MCMC's mixing time can cause the resulting inference to be slow, and thereby slow down model learning and prediction. We develop the first variational inference (VI) algorithm for NSPs, and give two examples of suitable variational posterior point process distributions. Our method minimizes the inclusive Kullback-Leibler (KL) divergence for VI to obtain the variational parameters. We generate samples from the approximate posterior point processes much faster than MCMC, as we can directly estimate the approximate posterior point processes without any MCMC steps or gradient descent. We include synthetic and real-world data experiments that demonstrate our VI algorithm achieves better prediction performance than MCMC when computational time is limited.