Calendar of Events

Speaker: Marat Freytsis (Rutgers University)

I'll discuss some work in progress using machine learning to extract information about potential low-frequency gravitational waves in pulsar timing array data. The methods in current use depend on frequentist estimators designed for specific signal forms that cannot be easily generalized and do not allow for quick searches of more complex or unconventional signals. In contrast to more conventional problems involving time-series data, most of the useful information lies in correlations between multiple time series requiring modified approaches and I will discuss a few of our attempts to address this issue.