In a recent publication, a multidisciplinary group of scientists, led by a researcher at the University of Manchester, developed a novel AI (artificial intelligence) strategy for translating technical astronomy terminology into plain English.
The new research, which is the result of the international RGZ EMU (Radio Galaxy Zoo EMU) collaboration, is changing the language of radio astronomy from specific terms like FRI (Fanaroff-Riley Type 1) to simple terms like “hourglass” or “traces host galaxy.”
The paper has been accepted for publication in the Royal Astronomical Society’s Monthly Notices.
In astronomy, specific ideas are efficiently and easily understood by professional astronomers through the use of technical terminology. However, using the same terminology can also make it difficult to involve non-experts in the discussion. On the Zooniverse citizen science platform, the RGZ EMU collaboration is developing a project that seeks public assistance in describing and classifying radio telescope images of galaxies.
“Using artificial intelligence to make scientific language more accessible allows us to share science with everyone.” With the plain English terminology we developed, the public may engage in current astronomy research like never before, and witness all of the incredible science being done throughout the world.”
Micah Bowles, Lead author and RGZ EMU data scientist,
Computer analysis can still miss interesting things that are easy for the human eye to see, and modern astronomy projects collect so much data that it is often impossible for scientists to look at it all by themselves.
“Using AI to make scientific language more accessible is helping us share science with everyone,” said lead author and RGZ EMU data scientist Micah Bowles. The public can now engage with modern astronomy research like never before and experience all of the amazing science being done around the world using the simple terms we developed.”
Similar to satellite dishes, radio telescopes can be used to detect radio light from extremely energetic astrophysical objects, such as black holes in other galaxies, as opposed to television signals. For a long time, these “radio cosmic systems” have been classified into various kinds by space experts to assist them with figuring out the starting points and development of the universe.
More and more of these radio galaxies have been discovered as a result of significant advancements in radio telescopes all over the world. This has not only made it impossible for professional astronomers to examine each one individually and classify it, but it has also introduced new variations that have not yet been captured by the types of radio galaxies that are already in existence. The RGZ EMU team saw a different way forward that would allow citizen scientists to participate more fully in their research project rather than attempting to train people to recognize the various types of radio galaxies and inventing new technical terms for them.
The RGZ EMU team first asked experts to explain a few radio galaxies in technical terms, and then they asked non-experts to explain them in simple terms. They then identified the straightforward descriptions that contained the most scientific information by employing a first-of-its-kind AI-based strategy they had developed. Now, anyone can use these descriptions (also known as “tags”) to describe radio galaxies in a way that makes sense to English-speaking people without needing any special training. This work will not only be important to the RGZ EMU project, but it could also be used in a lot of other situations where using simplified language can speed up research, collaboration, and communication. There is an ever-increasing amount of data in many areas of science.
Driven from Manchester, this examination was directed by analysts from the UK, China, Germany, the U.S., the Netherlands, Australia, Mexico, and Pakistan. All of the data, code, and results can be found online.
More information: Micah Bowles et al, Radio galaxy zoo EMU: Towards a semantic radio galaxy morphology taxonomy, Monthly Notices of the Royal Astronomical Society (2023). DOI: 10.1093/mnras/stad1021
