The High Time Resolution Universe Survey VI: An Artificial Neural Network and Timing of 75 Pulsars

TL;DR

This paper reports the discovery of 75 pulsars, including two with notable behaviors, and details the development of an artificial neural network that effectively filters pulsar candidates with high accuracy.

Contribution

It introduces a neural network-based data-processing pipeline that significantly improves pulsar candidate selection in large surveys.

Findings

Neural network rejected over 99% of false candidates

Detected 85% of pulsars blindly in data

Discovered 75 pulsars, including two with unique behaviors

Abstract

We present 75 pulsars discovered in the mid-latitude portion of the High Time Resolution Universe survey, 54 of which have full timing solutions. All the pulsars have spin periods greater than 100 ms, and none of those with timing solutions are in binaries. Two display particularly interesting behaviour; PSR J1054-5944 is found to be an intermittent pulsar, and PSR J1809-0119 has glitched twice since its discovery. In the second half of the paper we discuss the development and application of an artificial neural network in the data-processing pipeline for the survey. We discuss the tests that were used to generate scores and find that our neural network was able to reject over 99% of the candidates produced in the data processing, and able to blindly detect 85% of pulsars. We suggest that improvements to the accuracy should be possible if further care is taken when training an artificial neural network; for example ensuring that a representative sample of the pulsar population is used during the training process, or the use of different artificial neural networks for the detection of different types of pulsars.