AI Pushes the Final Frontier in Space

Galaxy clusters are one of the most massive space structures. Most galaxies exist far away from others or in small groups, in what is known as "the field", but, a few of them live in clusters. These rare space constructs are so dense, scientists have speculated that they hold the secrets about the mysterious dark matter.

Unbeknownst to the vast majority of people, searching for galaxy clusters is a task that is extremely labour intensive. The images that are captured by satellites have researchers pour through them, one by one, checking for the presence of any star clusters.

An example of a galaxy cluster. Researchers would look through each part of a massive image, trying to spot signs of something like this.

An example of a galaxy cluster. Researchers would look through each part of a massive image, trying to spot signs of something like this.

An example of how researchers used to search in the past is the "Abell Catalogue". George Abell, the father of cluster searching, searched for galaxy clusters by projecting images from telescopes onto photographic plates. He then got a magnification glass and iterated through each plate, collating results to create a record of galaxy groups in the northern hemisphere.

To avoid this workload, a team of researchers from Lancaster University built a deep-learning model to ease the workload. The model, called "Deep-CEE", uses unconventional deep learning techniques to streamline the process of searching for clusters in space.

The model works in a similar way to convolutional neural networks and Chan, the leading researcher, fed it examples of clusters to train it. The model combines Abell's strategy and modern computer vision techniques to "replace" the human element of the search. However, unlike the pop-fiction scenarios, humans are actively replacing themselves.

The researchers stated that they aim to run the network on new surveys coming in, in the immediate future. They have already run the model on a past set of data to ensure it works, and after good results, they plan to run it on the LSST's data. The LSST, short for Large Synoptic Survey Telescope, will be used in 2021 to measure and capture data from deeper into space as technology advances.

It is important to emphasise how essential such advancements are. Many people might get overwhelmed by the speed AI has advanced, but, it is important to remember such innovations help all. Especially when two extremely innovative fields, artificial intelligence and space, are combined, you should expect a massive amount of progress to follow, and maybe, such a combination may pave the way for future space travel.