BEGIN:VCALENDAR
BEGIN:VEVENT
SUMMARY:CS Colloquium: Richard Samworth (University of Cambridge) - High-dimensional changepoint estimation via sparse projection
DESCRIPTION:Speaker: Richard Samworth, University of Cambridge
Talk Title: High-dimensional changepoint estimation via sparse projection
Series: Yahoo! Labs Machine Learning Seminar Series
Abstract: This lecture satisfies requirements for CSCI 591: Computer Science Research Colloquium. Part of Yahoo! Labs Machine Learning Seminar Series. \n
\n
Changepoints are a very common feature of Big Data that arrive in the form of a data stream. We study high-dimensional time series in which, at certain time points, the mean structure changes in a sparse subset of the coordinates. The challenge is to borrow strength across the coordinates in order to detect smaller changes than could be observed in any individual component series. We propose a two-stage procedure called 'inspect' for estimation of the changepoints: first, we argue that a good projection direction can be obtained as the leading left singular vector of the matrix that solves a convex optimisation problem derived from the CUSUM transformation of the time series. We then apply an existing univariate changepoint detection algorithm to the projected series. Our theory provides strong guarantees on both the number of estimated changepoints and the rates of convergence of their locations, and our numerical studies validate its highly competitive empirical performance for a wide range of data generating mechanisms.
Biography: I am a Professor of Statistics in the Statistical Laboratory, a sub-department of the Department of Pure Mathematics and Mathematical Statistics. This is part of the Faculty of Mathematics at the University of Cambridge. I am also a Teaching Fellow at St John's College, and run the Statistics Clinic for members of the university.\n
\n
I currently hold a five-year EPSRC Early Career Fellowship, which began on 1 December 2012. I am also an Alan Turing Institute Faculty Fellow.
Host: Yan Liu
DTSTART:20161129T160000
LOCATION:SAL 101
URL;VALUE=URI:
DTEND:20161129T170000
END:VEVENT
END:VCALENDAR