Unveiling the Universe: AI's Role in Galaxy Exploration
Written on
Chapter 1: The Challenge of Identifying Galaxy Superclusters
Galaxy superclusters represent the most colossal structures in the universe, yet pinpointing their locations poses a significant challenge for astronomers. These vast networks of galaxies can span hundreds of millions of light-years, but their sparse distribution often makes them hard to detect.
To tackle this issue, researchers at Lancaster University have turned to artificial intelligence (AI). They developed a method known as Deep Learning for Galaxy Cluster Extraction and Evaluation (Deep-CEE) to analyze galaxies and help locate these elusive clusters. This innovative system has already assessed galaxies from the Sloan Digital Sky Survey, which provides the most detailed three-dimensional map of the universe to date.
The galaxy cluster Abell 1689 is highlighted, showcasing the gravitational lensing effect, where light from distant galaxies is bent by the massive gravitational field of the cluster. Image credit: NASA/ESA
Matthew Chan, a doctoral candidate at Lancaster University, commented, “We have successfully applied Deep-CEE to the Sloan Digital Sky Survey. In the future, we aim to utilize our model on groundbreaking surveys such as the Large Synoptic Survey Telescope (LSST), which will explore uncharted territories of the universe.”
AI as a Powerful Tool for Discovery
Astronomer George Abell created the first comprehensive catalog of galaxy clusters, known as the Abell catalog, in 1958. This catalog, which has undergone multiple updates, is still utilized today. Image credit: Emilio Segre Visual Archives / American Institute of Physics / Science Photo Library
From the 1950s to the 1980s, astronomer George Abell compiled a catalog of 4,073 galaxy clusters observable from the northern hemisphere, utilizing a magnifying lens to examine around 2,000 photographic plates for clusters of stars.
The Deep-CEE system employs AI to systematically analyze galaxy positions, mirroring Abell's meticulous searches with his photographic tools. “It is a cutting-edge model based on neural networks, designed to emulate the way the human brain learns to recognize patterns and colors by activating specific neurons,” reported the Royal Astronomical Society.
The research team trained Deep-CEE by repeatedly showing it known objects in images until it could independently identify galaxy clusters and superclusters. A pilot study provided the system with images filled with various astronomical objects, and it successfully detected galaxy clusters.
To develop this new AI approach, a probabilistic algorithm was applied, which relies partially on chance. Often referred to as randomized algorithms, this technique is beneficial for studying the evolution of natural systems over time.
The wealth of data on galaxy locations allows researchers to train their AI models using real galaxies, ensuring a realistic training environment. “With abundant labeled data from past classifications in public catalogs, we avoid relying on simulated data, thus keeping our training data realistic, which is crucial for deep learning algorithms,” the researchers noted in their paper published in the Monthly Notices of the Royal Astronomical Society.
Most galaxies exist in low-density regions dubbed “the field” or in smaller groups, such as the Local Group, which contains our Milky Way Galaxy. This Local Group is part of the Laniakea supercluster, boasting a mass 100,000 times that of the Milky Way and stretching over 500 million light-years. The name Laniakea means “immense heaven” in Hawaiian.
Big Questions Demand Big Data
A computer simulation illustrating the Laniakea supercluster of galaxies, highlighting its structure and the movement of galaxies toward its center. Image Credit: R. Brent Tully (U. Hawaii) et al., SDvision, DP, CEA/Saclay
By investigating superclusters, astronomers hope to unravel two of the universe's greatest mysteries: dark matter, an unseen force binding galaxies, and dark energy, which drives the universe's accelerating expansion.
Neil deGrasse Tyson once stated, “I want to know what dark matter and dark energy are comprised of. They remain a mystery, a complete mystery. No one is any closer to solving the problem than when these two things were discovered.”
Recently, an AI algorithm tasked with studying gravitational effects on galaxies demonstrated the ability to simulate various versions of the universe, varying in dark matter concentrations, without explicit programming for these simulations.
Advancements in telescopes and automated techniques are enabling astronomers to rapidly map and analyze the cosmos. The Large Synoptic Survey Telescope (LSST), set to commence operations in 2021, is projected to gather 15 terabytes (TB) of data each night, mapping the southern hemisphere's sky. AI will process this data, uncovering cosmic secrets and potentially raising new questions.
The first video titled "Allowing an AI to take over! | #5 | From The Depths 2022 | Adventure Mode" explores how AI is being integrated into complex simulations, demonstrating its capabilities in decision-making and exploration.
The second video, "Space Vehicle Quick Guide - From the Depths," provides a concise overview of space vehicles and their functionalities, showcasing the intersection of technology and space exploration.
Did you enjoy this article? Subscribe to The Cosmic Companion Newsletter for more insights!