Quickest Changepoint Detection in General Multistream Stochastic Models: Recent Results, Applications and Future Challenges

Abstract: Modern information systems generate large volumes of data with anomalies that occur at unknown points in time and have to be detected quickly and reliably with low false alarm rates. The paper develops a general theory of quickest multistream detection in non-i.i.d. stochastic models when a change may occur in a set of multiple data streams. The first part of the paper focuses on the asymptotic quickest detection theory. Nearly optimal pointwise detection strategies that minimize the expected detection delay are proposed and analyzed when the false alarm rate is low. The general theory is illustrated in several examples. In the second part, we discuss challenging applications associated with the rapid detection of new COVID waves and the appearance of near-Earth space objects. Finally, we discuss certain open problems and future challenges.