With over 144,370,000 square miles of surface terrain, Mars has a lot of places where signs of potential life could hide. Factor in the ultra-valuable time of current and future rovers, and it makes it even more challenging to scour for evidence of potential ancient microbes and organisms in an efficient way. To even the playing field a bit, SETI is turning again to artificial intelligence and machine learning in an effort to calculate the most likely and promising places for rovers—and, perhaps one day, astronauts—to look for clues of life. And as first detailed on Monday in Nature Astronomy, the team’s new AI machine learning modeling is already showing potential to speed up humanity’s search for alien life.
To build their AI, the interdisciplinary project led by SETI Institute Senior Research Scientist Kim Warren-Rhodes trained a program on datasets drawn from a region called Salar de Pajonales. Located at the border of Chile’s Atacama Desert and Altiplano, Pajonales served as a decent stand-in for Mars, with its high altitude, arid climate, dry salt lakebed, high amounts of ultraviolet light, and sparse, photosynthetic microbial life. The team amassed over 7,765 images and 1,154 samples of the area’s rocks, crystals, and salt domes, then used the information alongside other datasets to teach their program to understand and detect areas featuring small biosignatures. Upon turning the AI/ML program towards a new nearby area, the system managed to locate similar biosignatures nearly 88 percent of the time, versus less than 10 percent for previous random searches. The new method also decreased necessary search areas up to 97 percent.
In a statement, Rhodes explained that, “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.” They went on to express their hope that other astrobiologists will adapt the approach to mapping other environments, as well as to detect additional biosignatures. “With these models, we can design tailor-made roadmaps and algorithms to guide rovers to places with the highest probability of harboring past or present life—no matter how hidden or rare,” she said.
“While the high-rate of biosignature detection is a central result of this study, no less important is that it successfully integrated datasets at vastly different resolutions from orbit to the ground, and finally tied regional orbital data with microbial habitats,” said another team member, Nathalie A. Cabrol.
Over time, the team hopes they and other astrobiologist groups can continue to build collaborative datasets that could aid in the search for alien life via onboarding them to future planetary rovers.