Real-time quantum change point detection

Qu&Co comments on this publication:

Change point detection is a vast branch of statistical analysis developing techniques for uncovering abrupt changes in the underlying probability distribution of streaming data. This can be done off-line (using time-series data) or online (processing data sequentially). The latter enables real-time decision making, require less memory and is most relevant in machine learning. In this paper, Sentis et al. discuss online detection strategies for identifying a change point in a stream of quantum particles allegedly prepared in identical states. They show that the identification of the change point can be done without error via sequential local measurements, requiring only one classical bit of memory between subsequent measurements.