Radio Galaxy Zoo: EMU -- paving the way for EMU cataloging using AI and citizen science

TL;DR

This paper presents a combined citizen science and machine learning framework to improve the identification of extended radio sources in the EMU survey, enhancing catalog accuracy and completeness.

Contribution

It introduces a novel integrated approach leveraging both citizen science and AI to better classify extended radio sources in large astronomical surveys.

Findings

Framework is designed to identify around 4 million extended sources.

Expected to significantly improve EMUCAT cataloging accuracy.

Can incorporate external survey data for further enhancement.

Abstract

The Evolutionary Map of the Universe (EMU) survey with ASKAP is transforming our understanding of radio galaxies, AGN duty cycles, and cosmic structure. EMUCAT efficiently identifies compact radio sources, yet struggles with extended objects, requiring alternative approaches. The Radio Galaxy Zoo: EMU (RGZ EMU) project proposes a general framework that combines citizen science and machine learning to identify around 4 million extended sources in EMU. This framework is expected to enhance the EMUCAT cataloging on extended sources and can be further empowered with the introduction of cross-matched external data from surveys such as POSSUM and WALLABY.