Speaker: Martin Tveten
Title: Changepoint estimation for the modern world?
In this talk, I will share an interesting piece of high-dimensional statistics with you; "High-dimensional changepoint estimation via sparse projection", by Tengyao Wang and Richard Samworth, published in the Journal of the Royal Statistical Society B, June 2017.
One of Big Insight's major goals is to deal with transient and nonstationary phenomena. To handle data from such phenomena, segmenting time series into its stationary parts by estimating changepoints is an often used strategy. The presence of changepoints can for instance be used as evidence of the world not being stationary, and has therefore been used in everything from detecting system anomalies to solar bursts, analysis of climate change to whether humanity is becoming more peaceful or not, and data quality management to internet security. However, reliably detecting changepoints in high-dimensional data, in a computationally efficient manner, is a hard problem because it is often the case that only a small, unknown subset of components contains changepoints. The mentioned paper contains clever ideas for solving the problem, and could be useful in other high-dimensional settings as well.
The lunch starts at 12:00, and the talk will start around 12:20.
NB: Wednesday BigInsight Lunches are open to staff and students from any of the BigInsight partners, including UiO, but not to others.
Location: Spiseriet, Norwegian Computing Center