WUNA Lunch Talk:

Weiwei Zhu

University of British Columbia

Selection of Radio Pulsar Candidates Using Artificial Intelligence / 19 Years High Precision Timing of the Millisecond Pulsar J1713+0747

July 18

12:10PM, Room 230, NRAO, Edgemont Road


Millions of pulsar candidates have been produced by recent pulsar surveys; they cannot be inspected by human experts in a practical time frame. Here we present a supervised machine learning program for automatically identifying good pulsar candidates. The program is trained using several different candidate features including images and plots that human experts also look at. This program is being applied to the P-ALFA pulsar survey and can be applied to other future surveys as well.

PSR J1713+0747 is one of the best high-timing-precision pulsars monitored by the NANOGrav project for the purpose of detecting gravitational waves. We report the analysis of a 19-year span of timing data on PSR J1713+0747 taken by the Arecibo and Green Bank telescopes. The new timing solution we found improves the measurement of the pulsar?s mass, its orbital and geometric parameters, and set new limits on alternative gravitational theories.