PhD Seminar: Change-Point Detection for Monitoring the State of the Observed Systems

Мероприятие завершено

Speaker:  Maksim Karpov, third-year PhD student, Laboratory of Methods for Big Data Analysis, Faculty of Computer Science 
When: May 17, 18:10–19:30 

Detecting changes in the operating mode of a system is a well-known task that occurs in various applications: quality control of production processes, early detection of failures, predictive maintenance, monitoring the condition of patients in medicine, building language assistants, analyzing texts and videos.

To solve this problem, change-point detection methods are used. A change point is a point in time at which the behavior of the time series that characterizes the observed system is changing. An occurrence of a malfunction means that the state of the system has changed, and its early detection can help prevent undesirable consequences.

In this talk, we will discuss a particular task of detecting change points in time-series data obtained from information systems at LHCb experiment at CERN. Some background information about change-point detection methods and new approaches will be presented.