A team of astronomers from the University of Hawaii at the Mānoa Institute of Astronomy (IfA) has produced the most comprehensive catalog of astronomical images of stars, galaxies and quasars ever created with the help of an artificially intelligent neural network.
The group of astronomers from the University of Hawaii at the Mānoa Institute of Astronomy (IfA) issued a Catalog containing 3 billion celestial objects in 2016, including stars, galaxies and quasars (the active nuclei of supermassive black holes). It goes without saying that analyzing this large database, filled with 2 petabytes of data, was an unsuitable task for puny humans and even college students. One of the main objectives that emerged from the publication of the 2016 catalog was to better characterize these distant points of light and also to map the arrangement of galaxies in all three dimensions. The Pan-STARRS team can now check these items off their to-do list, thanks to the powers of machine learning. The results of their work were published to the Monthly Notices of the Royal Astronomical Society.
Their PS1 telescope, located atop Haleakalā on the Big Island of Hawaii, is capable of scanning 75% of the sky and currently hosts the largest deep multicolored optical survey in the world, according to a Press release published by the University of Hawai’i. In contrast, the Sloan Digital Sky Survey (SDSS) only covers 25% of the sky.
To provide the computer with a framework and teach it how to distinguish celestial classes of objects from each other, the team turned to publicly available spectroscopic measurements. These color and size measurements of the objects were in the millions, Robert Beck, the study’s lead author and former postdoctoral fellow in cosmology at IfA, explained in the press release.
“Using a state-of-the-art optimization algorithm, we leveraged the spectroscopic training set of nearly 4 million light sources to teach the neural network to predict the types of sources and distances to galaxies, while correcting for light extinction. for the Milky Way, “ Beck said.
These training sessions worked well; the resulting neural network did a great job when tasked with sorting objects, achieving success rates of 98.1% for galaxies, 97.8% for stars and 96.6% for quasar. The system also determined distances to galaxies, which at most were only about 3% off. The resulting work is “the world’s largest catalog of three-dimensional astronomical images of stars, galaxies and quasars,” according to the University of Hawai’i.
“This beautiful map of the universe provides an example of how the power of the Pan-STARRS big data set can be multiplied with artificial intelligence techniques and complementary observations,” explained Kenneth Chambers, team member and co-author of the study. “As Pan-STARRS collects more and more data, we will use machine learning to extract even more information about objects close to Earth, our solar system, our galaxy and our universe.”
The new catalog, made possible by a grant from the National Science Foundation, is publicly available through the Mikulski archive for space telescopes. The database is 300 gigabytes in size and is accessible in multiple formats, including computer downloadable tables.
T.his investigation has already yielded some interesting scientific data, including an explanation for a rather spooky spatial region known as the Cold Spot. Using the PS1 telescope and also NASA’s Wide Field Survey Explorer satellite, Pan-STARRS scientists have identified a huge supervoid, a “vast region of 1.8 billion light years in diameter, in which the density of galaxies is much lower than normal in the known universe,” as the University by Hawai’i described five years ago. It is this supervoid that is causing the cold spot, as seen in the cosmic microwave background, according to the researchers.
The updated map will also be used to study the overall geometry of the universe, to further test our theories on the standard cosmological model, and to analyze ancient galaxies, among many other avenues of astronomical and cosmological research.