Fifty Years of Candidate Pulsar Selection - What next?

TL;DR

This paper reviews fifty years of pulsar candidate selection, highlighting challenges from increasing candidate volumes and proposing the need for evolving machine learning methods to maintain detection accuracy.

Contribution

It discusses the historical context, current challenges, and future directions for candidate selection in pulsar astronomy, emphasizing community efforts and technological evolution.

Findings

Machine learning has improved candidate selection accuracy.

Candidate volumes are increasing exponentially, challenging current methods.

Evolving techniques are necessary to keep pace with data growth.

Abstract

For fifty years astronomers have been searching for pulsar signals in observational data. Throughout this time the process of choosing detections worthy of investigation, so called candidate selection, has been effective, yielding thousands of pulsar discoveries. Yet in recent years technological advances have permitted the proliferation of pulsar-like candidates, straining our candidate selection capabilities, and ultimately reducing selection accuracy. To overcome such problems, we now apply intelligent machine learning tools. Whilst these have achieved success, candidate volumes continue to increase, and our methods have to evolve to keep pace with the change. This talk considers how to meet this challenge as a community.