The universe is full of mysteries which we are yet to solve. For e.g. what are the structure of the universe and why the majority of the universe is made up of dark matter & dark energy of an unknown nature?
The advancement over space telescopes has helped us to capture the very micro details of the universe. But no telescope has ever been able to capture dark matter due to its extremely elusive and invisible nature.
The cosmos is spun together with filaments that bunch galaxies together to form clusters. These filaments resemble from far away, encompassing voids a zone where there appears to be nothing.
This has come to the discovery of the cosmic microwave background that gave us a glimpse of what the universe looked near its start; seeing how its structure developed over time to what it is today would help us to uncover the greatest mystery of modern physics i.e the characteristics of dark matter and dark energy.
A Japanese researcher’s team, including Kyoto University Yukawa Institute for Theoretical Physics Project Associate Professor Takahiro Nishimichi, & Kavli Institute for the Physics & Mathematics of the Universe Principal Investigator Masahiro Takada has built an AI tool that could help to solve mysteries of the universe.
The scientists used world’s fastest astrophysical simulation supercomputers ATERUI & ATERUI II to build the Dark Emulator which is able to ingest vast quantities of data and produce an analysis of the universe in sec.
Reportedly, scientists are calling it the ‘Dark Emulator’ through which data recorded by a few of the world’s largest observatories was run. The team of the researchers was thus able to study the possibilities of the origin of the cosmic structures, emulating an entire virtual universe using machine learning.
With the help of machine learning, the AI system creates hundreds of virtual universes. They were able to alter several important characteristics of the Universe, such as those of dark matter and dark energy, and generate a picture of how they might have changed the cosmos.
The “Dark Emulator” is able to learn from each of those simulations, allowing it to guess how changes in those characteristics would affect the results. That means it does not need to create entirely new simulations, allowing it to generate another virtual universe far more quickly than ever before.
In the paper, the researchers describe how it was able to accurately predict specific effects in their virtual universes in just seconds. Similar simulations would take days without using the Dark Emulator, they say.
According to the researchers, the AI tool is able to study possibilities about the origin of cosmic structures and how dark matter distribution may have changed over time, using data from some of the top observational surveys conducted about space. The database was gathered using a supercomputer over three years.
“We built an extraordinarily large database using a supercomputer, which took us three years to finish, but now we can recreate it on a laptop in a matter of seconds. I feel like there is great potential in data science,” said lead author Takahiro Nishimichi in a statement.
“Using this result, I hope we can work our way towards uncovering the greatest mystery of modern physics, which is to uncover what dark energy is. I also think this method we’ve developed will be useful in other fields such as natural sciences or social sciences.”
When testing the resulting tool with real-life surveys, it was able to successfully predict weak gravitational lensing effects in the Hyper Suprime-Cam survey, along with the three-dimensional galaxy distribution patterns recorded in the Sloan Digital Sky Survey to within 2% to 3% accuracy, in a matter of seconds.
It does this by analyzing the invisible tendrils between galaxies and performing astronomical (literally) feats of mathematics to create more precise simulations.
Eventually, this technology could help flesh out our understanding of the universe and allow scientists to determine exactly what dark matter is and how dark energy works. For now, this means filling in some of the massive blanks we have in our understanding of what the universe actually looks like beyond our front porch.
The researchers hope to apply their tool using data from upcoming surveys in the 2020s, enabling deeper studies of the origin of the universe.
The study is published in The Astrophysical Journal.
More in AI