When you hear about deep learning, you might think of the latest app that sprung up this week to do something clever with images or generate humanlike text. Deep learning might be responsible for some behind-the-scenes aspects of such things, but it's also used extensively in different fields of research. Recently, a team at a tech event called a hackathon applied deep learning to weather forecasting. It proved quite effective, and this got doctoral student Keiya Hirashima from the University of Tokyo's Department of Astronomy thinking.
"Weather is a very complex phenomenon but ultimately it boils down to fluid dynamics calculations," said Hirashima. "So, I wondered if we could modify deep learning models used for weather forecasting and apply them to another fluid system, but one that exists on a vastly larger scale and which we lack direct access to: my field of research, supernova explosions."
Supernovae occur when suitably massive stars burn through most of their fuel and collapse in enormous explosions. They are so huge that they can, and do, influence large areas inside their host galaxies. If a supernova had happened a few hundred years ago within a few hundred light-years from Earth, you might not be reading this article right now. So, the better we understand supernovae, the better we can understand why galaxies are the way they are.
"The problem is the time it takes to calculate the way supernovae explode. Currently, many models of galaxies over long time spans simplify things by pretending supernovae explode in a perfectly spherical fashion, as this is relatively easy to calculate," said Hirashima. "However, in reality, they are quite asymmetric. Some regions of the shell of material that forms the boundary of the explosion are more complex than others. We applied deep learning to help ascertain which parts of the explosion require more, or less, attention during a simulation to ensure the best accuracy, whilst also taking the least amount of time overall. This way of dividing a problem is called Hamiltonian splitting. Our new model, 3D-MIM, can reduce the number of computational steps in the calculation of 100,000 years of supernova evolution by 99%. So, I think we'll really help reduce a bottleneck too."
Of course, deep learning requires deep training. Hirashima and his team had to run hundreds of simulations taking millions of hours of computer time (supercomputers are highly parallel, so this length of time would be divided amongst the thousands of computing elements required). But their results proved it was worth it. They now hope to apply their methodology to other areas of astrophysics; for example, galactic evolution is also influenced by large star-forming regions. 3D-MIM models the deaths of stars, and perhaps soon it will be used to model their births as well. It could even find use beyond astrophysics altogether in other fields requiring high spatial and temporal resolutions, such as climate and earthquake simulations.
|Subscribe Free To Our Daily Newsletters|
NASA awards $2.3 million to study growing food in lunar dust
Earth bacteria could make lunar soil more habitable for plants
GreenOnyx's Wanna Greens Makes Space Debut Aboard SpaceX CRS-29 Mission
Big bang: Dutch firm eyes space baby
Heat Shield demo passes the test dubbed 'Just flawless'
UK Space Agency backs Orbit Fab's innovative refueling interface, GRASP
Starship Test Flies Higher: SpaceX Marks Progress Despite Late Test Incident
US 'strongly condemns' N. Korean space launch
NASA uses two worlds to test future Mars helicopter designs
California lawmakers ask NASA not to cut Mars budget
Spacecraft fall silent as Mars disappears behind the Sun
The Long Wait
China's BeiDou and Fengyun Satellites Elevate Global Weather Forecasting Capabilities
New scientific experimental samples from China's space station return to Earth
Shenzhou XVI crew return after 'very cool journey'
Chinese astronauts return to Earth with fruitful experimental results
SpaceX launches more Starlink satellites from Cape Canaveral|
Instruments led by IRF selected for ESA potential future mission to either Mars or Earth's Orbit
Maxar hands over JUPITER 3, to EchoStar
Maritime Launch reports non-brokered private placement of convertible debentures
ReOrbit's Report Highlights Software-First Satellites as Key Growth Drivers in Space Industry
Climate conspiracy theories flourish ahead of COP28
NASA's Deep Space Optical Comm Demo Sends, Receives First Data
Rice researcher scans tropical forest with mixed-reality device
Deformable Mirrors in Space: Key Technology to Directly Image Earth Twins
Hubble measures the size of the nearest transiting Earth-sized planet
NASA data reveals possible reason some exoplanets are shrinking
Webb detects water vapor, sulfur dioxide and sand clouds in the atmosphere of a nearby exoplanet
Juice burns hard towards first-ever Earth-Moon flyby
Fall into an ice giant's atmosphere
Juno finds Jupiter's winds penetrate in cylindrical layers
Salts and organics observed on Ganymede's surface by June
|Subscribe Free To Our Daily Newsletters|