Stunned disclosure of technology to identify cosmic galaxies

, voice and image recognition applications show promise for astronomers to analyze images of galaxies and understand how they form and grow.

In the new study, the researchers used computers to simulate the formation of cosmic galaxies to create an intensive machine learning algorithm, after which the system analyzed images of galaxies from data.

The hypothetical images used to help identify the three key stages of galaxy evolution. Later, the researchers put the actual Hubble images into the system to conduct the classification.

The results show significant consistency in the classification and identification of simulated and real galaxies.

Co-author Joel Primack, professor of physics and a member of the Santa Cruz County Physics Institute (SCIPP) at UC Santa Cruz, said: 'We know simulations have limitations, so we do not want to make a big statement. But it is a very lucky and significant discovery. "

The hypothetical images used to help identify the three key stages of galaxy evolution.(Image source: Phys).

Galaxies are quite complex, changing astronomical objects that evolve over billions of years and images of galaxies can only provide timely snapshots. Astronomers can look deeper into the universe and from there "turn back time" to see previous galaxies (because of the time it takes light to travel distance in the universe), but after the progress Chemicalization of a single galaxy over time can only be done in simulations. Comparing galaxies simulating with observed galaxies can reveal important details of actual galaxies and their history.

The researchers used advanced galaxy simulations (VELA simulation) developed by Primack and a group of international collaborators, including Daniel Ceverino (University of Heidelberg), who runs simulation technology, and Avishai Dekel (Hebrew University), which leads to analyzing and explaining them and developing new physical concepts based on them. However, all such simulations are limited in their ability to capture complex physics in galaxy formation.

In particular, the simulations used in this study do not include feedback from active galactic nuclei (pumping energy from radiation when gas is accreted by the supermassive black hole center). Many astronomers consider this process an important factor in regulating star formation in galaxies. However, observing the distant galaxies, the young galaxy appears to show evidence of the phenomenon that leads to the blue nugget phase seen in galaxy simulations.

Koo, a CANDELS coordinator, invited Huertas-Company to visit UC Santa Cruz to continue this work. Google has provided support for deep astronomy research through research funds for Koo and Primack, allowing Huertas-Company to spend the summer in Santa Cruz, with plans for a survey of galaxies. another in the summer of 2018.

For observational data, the team used images of galaxies obtained through the CANDELS project - the largest research project in the history of the Hubble Space Telescope. The first author Marc Huertas-Company, an astronomer at the Paris Observatory and Paris Diderot University, has done the pioneering work of applying intensive machine learning methods to classify galaxies using data. whether CANDELS is public.

Koo, a CANDELS coordinator, invited Huertas-Company to visit UC Santa Cruz to continue this work. Google has provided support for deep astronomy research through research funds for Koo and Primack, allowing Huertas-Company to spend the summer in Santa Cruz, with plans for a survey of galaxies. another in the summer of 2018.