Professor Michael Coughlin

The current and future generation of wide-field telescopes with high survey cadences is opening the frontier of the study of variable sources. The first light science of the Zwicky Transient Facility (ZTF) has yielded a significant number of discoveries. The cadence of this system makes discoveries of systems with periods less than one hour possible. Due to the significant amount of astronomical time series data provided by ZTF and other all-sky surveys, it is important to have robust selection criteria and algorithms to find objects with periodic lightcurves. The fact that lightcurves are naturally irregularly sampled and have gaps and error bars means that these algorithms must be robust in order to efficiently find true signals. There are a variety of period finding algorithms in the literature, including those based on least-squares fitting to a set of basis functions. These researchers intend to use "conditional entropy," based on an information theoretic approach, optimized for GPUs, to perform the search. Broadly, information theory-based approaches improve on other techniques by capturing higher order statistical moments in the data, which are able to better model the underlying process and are more robust to noise and outliers. The sensitivity to periodic variations depends on characteristics intrinsic to the object such as the amplitude and the period of the signal being searched for. It also depends on the observations, including the noise levels, sampling, and overall length of the lightcurve being used. The group intends to perform period searches of millions of objects identified by the Zwicky Transient Facility and other surveys, in particular looking for rare systems, such as the population of short-period white-dwarf binaries which will be gravitational-wave sources for future space based gravitational-wave detectors.