Artificial Intelligence May Assist In The Search For Life On Mars And Other Extraterrestrial Planets

The search for life on Mars and other extraterrestrial worlds may be aided by a recently created machine-learning tool.

Currently, scientists are forced to rely on remote sensing techniques to look for signs of extraterrestrial life because the ability to collect samples from other planets is severely constrained. Therefore, any technique that could help focus or hone this search would be very helpful.

With this in mind, a multidisciplinary team of scientists under the direction of Kim Warren-Rhodes of the SETI (Search for Extraterrestrial Intelligence) Institute in California mapped the sparse lifeforms that live in salt domes, rocks, and crystals in the Salar de Pajonales, a salt flat on the border of the Chilean Atacama Desert and Altiplano, or high plateau.

After that, Warren-Rhodes collaborated with Freddie Kalaitzis from the University of Oxford and Michael Phillips from the Johns Hopkins University Applied Physics Laboratory to train a machine learning model to recognize the patterns and guidelines governing the distribution of life across the harsh region. Through this training, the model learned to recognize the same patterns and guidelines for a variety of landscapes, including those that might exist on other planets.

The team found that by fusing statistical ecology with artificial intelligence, their system could locate and detect biosignatures up to 87.5% of the time. This contrasts with the maximum 10 percent success rate attained by random searches. Furthermore, the program could cut down the search area by up to 97%, greatly assisting researchers in their search for potential chemical biosignatures.

In a statement(opens in new tab), Warren-Rhodes said, "Our framework allows us to combine the power of statistical ecology with machine learning to discover and predict the patterns and rules by which nature survives and distributes itself in the harshest landscapes on Earth.". We hope that other astrobiology teams will adopt our strategy to map other habitable environments and biosignatures. ".

The Perseverance rover, operated by NASA, is currently looking for signs of life on the floor of Mars' Jezero Crater. The researchers say that similar machine learning tools could be used on robotic planetary missions like this one.

No matter how obscure or uncommon, "with these models, we can design tailored roadmaps and algorithms to guide rovers to places with the highest probability of harboring past or present life," Warren-Rhodes said.

Choosing an Earth-based analog of Mars

Because Salar de Pajonales is a good representation of the dry, arid landscape of contemporary Mars, the team used it as a testing stage in their machine learning model. The area is a dry salt lakebed at a high altitude that receives intense UV radiation. Salar de Pajonales, however, still supports some life despite being regarded as being very inhospitable to it.

To find photosynthetic microbes residing in the area's salt domes, rocks, and alabaster crystals, the team gathered over 1,000 samples and nearly 8,000 images from Salar de Pajonales. The pigments that these microbes produce are a potential biosignature on NASA's "ladder of life detection"(opens in new tab), which is intended to direct researchers in their search for extraterrestrial life within the practical bounds of robotic space missions.

Using drone imagery similar to the photographs of Martian terrain taken by the High-Resolution Imaging Experiment (HIRISE) camera on NASA's Mars Reconnaissance Orbiter, the team also examined Salar de Pajonales. Using this information, they were able to conclude that the microbial life at Salar de Pajonales is not randomly distributed but is instead concentrated in biological hotspots that are closely related to the presence of water.

Convolutional neural networks (CNNs) were subsequently trained by Warren-Rhodes' team at Salar de Pajonales to identify and forecast significant geologic features. Some of these characteristics, like patterned ground or polygonal networks, are also present on Mars. Additionally, the CNN was trained to identify and forecast the smaller microhabitats most likely to contain biosignatures.

For the time being, the scientists will keep refining their AI at Salar de Pajonales with the goal of testing its capacity to predict the distribution and location of prehistoric stromatolite fossils and salt-tolerant microbiomes in the future. This should enable it to determine whether the principles it employs in this search can also be used to look for biosignatures in other natural systems that are similar.

The team then hopes to teach the AI to focus on potential habitats in other extreme environments on Earth before possibly exploring those on other planets by mapping hot springs, frozen permafrost-covered soils, and the rocks in dry valleys.