Supernovas with artificial intelligence: Chilean scientists use NASA telescopes to study the universe

Two years ago, the preparation of the astronomical broker Machine learning for rapid event classification (ALERC)a system that is now in charge of processing the data delivered by the four telescopes that make up the ATLAS project, located in Hawaii, South Africa and another recently inaugurated in Chile.

The objective of this observatory is detect and provide early warning of asteroids that pose a risk to people’s lives, but along the way, more information emerges that find data relevant to the scientific community, such as supernova events or black holes. ALeRCE is in charge of this data. To collect them, telescopes scan the night sky four times a night, identifying millions of variable events. it would be impossible to record without the use of advanced computers and artificial intelligence.

“A The broker is an intermediary between two types of telescopes. The former who search for interesting objects in the universe and the latter who, once the phenomenon has been identified, begin to operate in a more focused and detailed manner. Telescopes watch all night long, seeking to identify a change, something that is now in the sky that was not there before. Every 30 seconds, 10,000 objects arrive and once the new phenomenon has been identified, it must be studied in more detail using specialized instruments.“, Explain Francois Forsterprincipal researcher of the ALeRCE and the Center for Mathematical Modeling of the University of Chile (CMM).

As detailed by the also doctor of astronomy and associate researcher of the Millennium Institute of Astrophysics (MAS)the challenge taken up by the interdisciplinary team of ALeRCE “consists of converting this information into a data stream that is cross-checked in real time with observations from the Zwicky Transient Facility (ZTF) in CaliforniaUnited States, for thus being the first broker that combines large data streams in a global multi-telescope system. This combination of data will improve scientific discoveries and increase the resilience of ALeRCE, as an automatic classification system for variable objects in the universe.”

ALeRCE Explorer: top left shows the evolution of the luminosity of a supernova observed by ZTF (green and red) and ATLAS (cyan and orange); top right, an image of the host galaxy before the explosion; bottom left, the most probable classes according to the ALeRCE classifier; and bottom right, supernova discovery images. All of this information is updated in real time as new data is returned from these two telescopes.

“Yes good ATLAS works to analyze possible impacts of meteorites on the earth, the research process provides an infinity of useful information to discover new black holes, supernova explosions and an infinity of phenomena variables that allow us to understand the evolution of our universe. Forster clarified. Furthermore, he added that, in particular, “What interests us is being able to detect younger and younger supernovae that are in the process of exploding.since the first hours and the first days of the explosion are the times when more information can be obtained”.

The ALeRCE was created in 2017 as an inter-institutional collaboration between the Center for Mathematical Modeling of the University of Chile (CMM) and the Millennium Institute of Astrophysics (MAS). They were joined by the Data Observatory (DO) foundation in 2020, and the universities of Concepción (UdeC) and Mayor in 2022. This, in addition to adding researchers from more than 20 domestic and foreign institutions since its inception.

Francois Forster

This astronomical broker currently processes nearly 300,000 events with a dedicated engineering team developing tools for a community of over 6,000 users, in 125 countries. Meanwhile, in 2021, this national project has been selected as one of seven community brokers worldwide that will classify the more than 10 million events per night that the Vera C. Rubin Observatory will report, when it will start operating from Chile in 2024.

It was also a pioneer in several aspects: was the first broker to publicly publish its ranking using artificial intelligence; it is the one that reports the largest and fastest number of supernova explosions (more than 14,000 candidates); is the first to automatically enrich data with historical information on billions of objects. Since June 2022, it is the first broker in ingest data from the network of four ATLAS telescopes, located in Hawaii, South Africa and Chile.

ALeRCE is a system that works with a hybrid infrastructure, data processing mainly in the cloud (AWS-DO), in the National Laboratory for High Performance Computing (NLHPC)the CMM and using its own resources, installed in the National University Network (REUNA). A critical point for the exchange of national and international academic traffic.

Guillaume Cabrera

“For years, we have been developing the ZTF-focused AI tools. This new step poses new challenges for us, in extending these methods to data from different telescopes. It’s like teaching the computer to learn from different sources, just like us humans.“, assures Guillermo Cabrera-Vives, head of the machine learning domain at ALeRCE, director of the Data Science Unit at UdeC and associate researcher at MAS.

Andrew Jordan, director of the Millennium Institute of Astrophysics (MAS) and scientific director of the Data Observatory (DO), pointed out that “the biggest challenges for the ALeRCE in the future are scaling up to handle the huge volume of data from the Vera C. Rubin Observatory, iIncorporate data from more telescopes and develop new artificial intelligence tools that allow the international community to extract the best of science from these observations. In particular, an important challenge is to develop tools to detect these objects of unknown nature, which could revolutionize our understanding of the dynamic universe in the next decade”.

During this year, the ALeRCE processed more than 200 million alerts in real time, including 40 million images. As well as, reported over 6,000 supernovae, 60,000 supermassive black holes and 800,000 variable stars.

#Supernovas #artificial #intelligence #Chilean #scientists #NASA #telescopes #study #universe

Leave a Reply

Your email address will not be published.