Simulations of a region of space 100 million light-years square. The leftmost simulation ran at low resolution. Using machine learning, researchers upscaled the low-res model to create a high-resolution simulation (right). That simulation captures the same details as a conventional high-res model (middle) while requiring significantly fewer computational resources.
Y. Li et al./Proceedings of the National Academy of Sciences 2021
Using neural networks, Flatiron Institute research fellow Yin Li and his colleagues simulated vast, complex universes in a fraction of the time it takes with conventional methods.
Using a bit of machine learning magic, astrophysicists can now simulate vast, complex universes in a thousandth of the time it takes with conventional methods. The new approach will help usher in a new era in high-resolution cosmological simulations, its creators report in a study published online May 4 in Proceedings of the National Academy of Sciences.
“At the moment, constraints on computation time usually mean we cannot simulate the universe at both high resolution and large volume,” says study lead author Yin Li, an astrophysicist at the Flatiron Institute in New York City. “With our new technique, it’s possible to have both efficiently. In the future, these AI-based methods will become the norm for certain applications.”
The new method developed by Li and his colleagues feeds a machine learning algorithm with models of a small region of space at both low and high resolutions. The algorithm learns how to upscale the low-res models to match the detail found in the high-res versions. Once trained, the code can take full-scale low-res models and generate ‘super-resolution’ simulations containing up to 512 times as many particles.
The process is akin to taking a blurry photograph and adding the missing details back in, making it sharp and clear.
This upscaling brings significant time savings. For a region in the universe roughly 500 million light-years across containing 134 million particles, existing methods would require 560 hours to churn out a high-res simulation using a single processing core. With the new approach, the researchers need only 36 minutes.
The results were even more dramatic when more particles were added to the simulation. For a universe 1,000 times as large with 134 billion particles, the researchers’ new method took 16 hours on a single graphics processing unit. Existing methods would take so long that they wouldn’t even be worth running without dedicated supercomputing resources, Li says.
Li is a joint research fellow at the Flatiron Institute’s Center for Computational Astrophysics and the Center for Computational Mathematics. He co-authored the study with Yueying Ni, Rupert Croft and Tiziana Di Matteo of Carnegie Mellon University; Simeon Bird of the University of California, Riverside; and Yu Feng of the University of California, Berkeley.
Cosmological simulations are indispensable for astrophysics. Scientists use the simulations to predict how the universe would look in various scenarios, such as if the dark energy pulling the universe apart varied over time. Telescope observations may then confirm whether the simulations’ predictions match reality. Creating testable predictions requires running simulations thousands of times, so faster modeling would be a big boon for the field.
Reducing the time it takes to run cosmological simulations “holds the potential of providing major advances in numerical cosmology and astrophysics,” says Di Matteo. “Cosmological simulations follow the history and fate of the universe, all the way to the formation of all galaxies and their black holes.”
So far, the new simulations only consider dark matter and the force of gravity. While this may seem like an oversimplification, gravity is by far the universe’s dominant force at large scales, and dark matter makes up 85 percent of all the ‘stuff’ in the cosmos. The particles in the simulation aren’t literal dark matter particles but are instead used as trackers to show how bits of dark matter move through the universe.
The team’s code used neural networks to predict how gravity would move dark matter around over time. Such networks ingest training data and run calculations using the information. The results are then compared to the expected outcome. With further training, the networks adapt and become more accurate.
The specific approach used by the researchers, called a generative adversarial network, pits two neural networks against each other. One network takes low-resolution simulations of the universe and uses them to generate high-resolution models. The other network tries to tell those simulations apart from ones made by conventional methods. Over time, both neural networks get better and better until, ultimately, the simulation generator wins out and creates fast simulations that look just like the slow conventional ones.
“We couldn’t get it to work for two years,” Li says, “and suddenly it started working. We got beautiful results that matched what we expected. We even did some blind tests ourselves, and most of us couldn’t tell which one was ‘real’ and which one was ‘fake.’”
Despite only being trained using small areas of space, the neural networks accurately replicated the large-scale structures that only appear in enormous simulations.
The simulations don’t capture everything, though. Because they focus only on dark matter and gravity, smaller-scale phenomena — such as star formation, supernovae and the effects of black holes — are left out. The researchers plan to extend their methods to include the forces responsible for such phenomena, and to run their neural networks ‘on the fly’ alongside conventional simulations to improve accuracy. “We don’t know exactly how to do that yet, but we’re making progress,” Li says.
The Latest Updates from Bing News & Google News
Go deeper with Bing News on:
Astrophysics machine learning
- A Youtube Video Has Led To A $10,000 Wager On The Laws Of Physicson June 18, 2021 at 9:08 am
Popular science communicator and creator of the Veritasium YouTube channel Derek Muller and UCLA professor of physics Alexander Kusenko have a $10,000 bet ...
- AcademicInfluence.com's Inflection Announces the Decade's Most Influential Books in Physics, Geoscience, History, and Political Scienceon June 17, 2021 at 5:42 am
For centuries, books have been our most important medium for the expression of revolutionary ideas. The people who read these books use ...
- An ally for alloyson June 16, 2021 at 9:00 am
The consortium seeks to accelerate the development of improved heat-resistant alloys for various power plant components and to predict the alloys' long-term performance. A thermal power plant's ...
- Cord raises $4.5M to automate computer vision annotation processeson June 15, 2021 at 3:00 pm
Cord, a computer vision annotation automation platform, has raised an estimated $4.5 million in a seed funding round.
- This Wind-Powered Car Goes Fast Enough To Break Physicson June 9, 2021 at 5:15 pm
Whatever, physics is weird and I barely understand regular math. Thankfully the Veritasium channel on YouTube is here to help us understand things a bit better. There is currently no way to ...
Go deeper with Google Headlines on:
Astrophysics machine learning
Go deeper with Bing News on:
Generative adversarial network
- Tool around in a real-time generated AI version of 'GTAV'on June 18, 2021 at 1:49 pm
GameGAN, as the name suggests, is a generative adversarial network. Every GAN consists of two competing neural networks: a generator and a discriminator. The generator is trained on a sample dataset ...
- AI Under the Hood: Object Detection Model Capable of Identifying Floating Plastic Beneath the Surface of the Oceanon June 17, 2021 at 6:00 am
A group of researchers, Gautam Tata, Sarah-Jeanne Royer, Olivier Poirion, and Jay Lowe, have written a new paper "DeepPlastic: A Novel Approach to Detecting Epipelagic Bound Plastic Using Deep Visual ...
- human in the loopon June 14, 2021 at 5:01 pm
While a generative adversarial network (GAN), where two neural networks compete in a game, has been demonstrated, this is the first time adversarial human users have been used in a learning effort.
- Generating Synthetic Data Using a Generative Adversarial Network (GAN) with PyTorchon June 9, 2021 at 2:37 am
Dr. James McCaffrey of Microsoft Research explains a generative adversarial network, a deep neural system that can be used to generate synthetic data for machine learning scenarios, such as generating ...
- There’s More To Machine Learning Than CNNson June 9, 2021 at 12:01 am
Different learning structures provide optimizations based on variables such as time, accuracy, and what's considered important in the data.