Spectroscopy and Machine Learning Focus Lens on Far Reaches of the Universe

Facebook X LinkedIn Email
Researchers from the ARC Centre of Excellence for All Sky Astrophysics in 3 Dimensions (ASTRO 3D) and the University of New South Wales (UNSW) Sydney spectroscopically confirmed a number of strong gravitational lenses that were initially identified using convolutional neural networks (CNNs).

“Our spectroscopy allowed us to map a 3D picture of the gravitational lenses to show they are genuine and not merely chance superposition,” professor Kim-Vy Tran said.

A machine learning algorithm that searched for certain digital signatures made the work of Tran and the team possible. The algorithm was developed by researcher Colin Jacobs at Swinburne University of Technology, who sifted through tens of millions of galaxy images to prune the sample down to 5000.
Pictures of gravitational lenses from the AGEL survey. The pictures are centered on the foreground galaxy and include the object name. Each panel includes the confirmed distance to the foreground galaxy (zdef) and distant background galaxy (zsrc). Courtesy of Kim-Vy H. Tran, et al., The ASTROnomical Journal (2022).
Pictures of gravitational lenses from the AGEL survey. The pictures are centered on the foreground galaxy and include the object name. Each panel includes the confirmed distance to the foreground galaxy (zdef) and distant background galaxy (zsrc). Courtesy of Kim-Vy H. Tran et al./The Astronomical Journal (2022).

Using the Keck Observatory in Hawaii and the Very Large Telescope in Chile, the researchers assessed 77 of the roughly 5000 potential gravitational lenses that were identified using the machine learning algorithm. The team confirmed that 68 of the 77 lenses, or 88% of the lenses, are strong gravitational lenses covering vast cosmic distances. Such a high percentage suggests that the algorithm used to detect the lenses is reliable — and that potentially thousands of new gravitational lenses, discovered by the algorithm, are yet to be confirmed.

Gravitational lenses are cosmic magnifying glasses used to explore a range of astrophysical phenomena. Strong gravitational lensing extends the observational reach of scientists to include objects that are too faint for even the most powerful telescopes.

In addition to giving scientists a clearer view of objects that are millions of light years away, the newly validated strong gravitational lenses could help scientists identify the invisible dark matter that makes up most of the universe. “We know that most of the mass is dark,” Tran said. “We know that mass is bending light, and so if we can measure how much light is bent, we can then infer how much mass must be there.”

“Apart from being beautiful objects, gravitational lenses provide a window to studying how mass is distributed in very distant galaxies that are not observable via other techniques,” professor Stuart Wyithe, director of ASTRO 3D, said. “By introducing ways to use these new, large data sets of the sky to search for many new gravitational lenses, the team opens up the opportunity to see how galaxies get their mass.”

Having access to many more gravitational lenses located at various distances from Earth will also give scientists a more complete overview of the timeline of the universe, going back almost to the Big Bang.

“The more magnifying glasses you have, the better you can try to survey these more distant objects,” Tran said. “Hopefully, we can better measure the demographics of very young galaxies. Somewhere between those really early first galaxies and us, there’s a whole lot of evolution that’s happening, with tiny star-forming regions that convert pristine gas into the first stars to the sun, the Milky Way. With these lenses at different distances, we can look at different points in the cosmic timeline to track essentially how things change over time, between the very first galaxies and now.”

Professor Tucker Jones of the University of California, Davis described the new sample as “a giant step forward in learning how galaxies form over the history of the universe.” To date, gravitational lenses have been hard to identify, and only about 100 are used routinely.

The work is part of the ASTRO 3D Galaxy Evolution with Lenses (AGEL) survey. The goal of the international research team for AGEL is to spectroscopically confirm a statistically robust sample of about 100 strong gravitational lenses that can be observed with adaptive optics using telescopes in both hemispheres of Earth throughout the year.

To more accurately model the lens mass distribution, spatially resolve subkiloparsec structure in the sources, and search for dark matter substructure in the deflectors and along the line of sight, the team is acquiring high-resolution imaging with the Hubble Space Telescope for a subset of AGEL systems.

The research was published in The Astronomical Journal (

Published: October 2022
artificial intelligence
The ability of a machine to perform certain complex functions normally associated with human intelligence, such as judgment, pattern recognition, understanding, learning, planning, and problem solving.
deep learning
Deep learning is a subset of machine learning that involves the use of artificial neural networks to model and solve complex problems. The term "deep" in deep learning refers to the use of deep neural networks, which are neural networks with multiple layers (deep architectures). These networks, often called deep neural networks or deep neural architectures, have the ability to automatically learn hierarchical representations of data. Key concepts and components of deep learning include: ...
machine learning
Machine learning (ML) is a subset of artificial intelligence (AI) that focuses on the development of algorithms and statistical models that enable computers to improve their performance on a specific task through experience or training. Instead of being explicitly programmed to perform a task, a machine learning system learns from data and examples. The primary goal of machine learning is to develop models that can generalize patterns from data and make predictions or decisions without being...
The scientific observation of celestial radiation that has reached the vicinity of Earth, and the interpretation of these observations to determine the characteristics of the extraterrestrial bodies and phenomena that have emitted the radiation.
The displacement of spectrum lines, as determined by the increasing distance between, and the relative velocity of, the observer and a light source, causing the lines to move toward the red portion of the spectrum. It is used in astrophysics to determine the rate of recession or expansion of celestial bodies. Also known as the Hubble effect.
Research & TechnologyeducationAsia-PacificARC Centre of ExcellenceUniversity of New South WalesImagingOpticsSensors & Detectorsspectroscopyartificial intelligencedeep learningmachine learningneural networksconvolutional neural networkslensesgravitational lensesastronomytelescopesDark Energy Cameraredshiftdark matterBig Bang

We use cookies to improve user experience and analyze our website traffic as stated in our Privacy Policy. By using this website, you agree to the use of cookies unless you have disabled them.