New simulation charts how the early universe developed within seconds of the Big Bang
A new simulation maps the first few seconds after the Big Bang, focusing on what scientists call the intergalactic medium, or the gas and dust between galaxies.
A team led by researchers at the Institute of Astrophysics of the Canary Islands (IAC) used machine learning, a type of algorithm in which a computer is trained to recognize patterns, to complete 100,000 hours of computation. The algorithm for this project is called Hydro-BAM.
This new work allowed researchers to chart phenomena including dark matter, energized gas, neutral hydrogen and other cosmic ingredients that are essential to understanding the structure of our universe, IAC representatives said in a May 20 statement (opens in new tab).
“The research has also made it possible to reproduce with high precision the so-called ‘Lyman-alpha forests,'” they added. That is a particular pattern of lines in a spectrum (light signature) of galaxies and similar objects created when clouds of hydrogen gas in the way absorb the galactic light.
“These ‘virtual universes’ serve as test beds for the study of cosmology,” the researchers added. “However, the simulations are computationally very expensive, and current computing facilities only allow [us] to explore small cosmic volumes.”
Hydro-BAM is designed to include probability, machine learning and cosmology, meaning the history of the universe. “This algorithm has made it possible to obtain very accurate predictions in just a few tens of seconds,” the researchers said.
Charting the absorption lines in the galactic spectra allowed the team to learn about where the clouds of hydrogen gas are located. Location is a proxy for distance, given that the universe is continually expanding. The clouds also give clues as to what is contained in the intergalactic medium of gas and dust.
“The breakthrough came when we understood that the connections between the quantities of intergalactic gas, dark matter and neutral hydrogen that we were trying to model are well organized in a hierarchical way,” Francesco Sinigaglia, a doctoral student at the University of La Laguna in Spain, the IAC and the University of Padua in Italy, and lead author of the research, said in the statement.