No-regret Algorithms for Capturing Events in Poisson Point Processes


Inhomogeneous Poisson point processes are widely used models of event occurrences. We address emphadaptive sensing of Poisson Point processes, namely, maximizing the number of captured events subject to sensing costs. We encode prior assumptions on the rate function by modeling it as a member of a known emphreproducing kernel Hilbert space (RKHS). By partitioning the domain into separate small regions, and using heteroscedastic linear regression, we propose a tractable estimator of Poisson process rates for two feedback models: emphcount-record, where exact locations of events are observed, and emphhistogram feedback, where only counts of events are observed. We derive provably accurate anytime confidence estimates for our estimators for sequentially acquired Poisson count data. Using these, we formulate algorithms based on optimism that provably incur sublinear count-regret. We demonstrate the practicality of the method on problems from crime modeling, revenue maximization as well as environmental monitoring.

Proc. International Conference for Machine Learning (ICML)