Abstract: Online detection of abrupt changes in streaming time series is a challenging problem with many applications, in particular when little prior knowledge of the statistics of the data is ...