Scientists have proposed a new way to optimize the search for aliens on Mars by teaching artificial intelligence to pinpoint sites that are most likely to contain “biosignatures,” or signs of life, reports a new study.
The new deep learning framework was trained to find biosignatures in a harsh Chilean environment that is exposed to high levels of radiation and extremely cold temperatures, creating conditions that are about as close to Mars as you can get on Earth. The AI tool was able to boost the probability of detecting biosignatures in this extreme environment up to 87.5 percent, making it roughly nine times as effective as random searches for signs of life.
Some four billion years ago, Mars was warmer, wetter, and potentially habitable to life. Orbiters and rovers have exposed many regions on the planet that were sculpted by water, the key ingredient for life as we know it, revealing tantalizing glimpses of what may have once been habitats for Martian microbes. For instance, NASA’s Perseverance rover is currently exploring Jezero Crater, an ancient Martian lakebed that brimmed over with water long ago.
While we have a rough idea of where to look for the remains of microbial life on Mars, narrowing down the search to the scale of tiny habitats is a much more difficult task. To address this problem, scientists led by Kimberley Warren-Rhodes, an astrobiologist at the not-for-profit SETI Institute, developed “an adaptable framework that couples statistical ecology with deep learning to recognize and predict biosignature patterns,” according to a study published on Monday in Nature Astronomy.
“In the search for biosignatures on Mars, there is an abundance of data from orbiters and rovers to characterize global and regional habitability, but much less information is available at the scales and resolutions of microbial habitats and biosignatures,” Warren-Rhodes and her colleagues said in the study. “Understanding whether the distribution of terrestrial biosignatures is characterized by recognizable and predictable patterns could yield signposts to optimize search efforts for life on other terrestrial planets.”
“In extreme environments, the distribution of biosignatures is tightly controlled by a complex interdependence of geological, physicochemical and biological interactions,” the team added, noting that access to water is an especially important factor in this calculus. “To date, few studies have systematically examined such linkages across integrated spatial scales or applied machine learning to test the predictive power and probabilities of detecting life at the extremes of habitability.”
With their new paper, the researchers aimed to fill this gap in the research by training a deep learning system to predict the presence of biosignatures at Chile’s Salar de Pajonales, a mountain lake that sits at a high elevation of 3,500 meters at the edge of the hyperarid Atacama Desert.
While no environment on Earth is quite as extreme as modern Mars, Salar de Pajonales “exhibits features from both physical and biological processes highly relevant to biosignature search on Mars” such as “fractal-like ridge networks, patterned ground and shrinkage crack terrains of abiotic and/or biotic origin,” according to the study.
Warren-Rhodes and her colleagues used aerial and ground observations to map out the distribution of the hardy photosynthetic microorganisms that live in this challenging habitat. The team then trained neural networks to predict the odds of biosignatures using a number of geological parameters, such as salinity, rock type, and access to light.
The deep learning approach was able to predict probabilities of biosignature detection at an impressive level of 56.9 to 87.5 percent, making it “a powerful tool to expedite the search and detection of biosignatures in terrestrial analogues,” according to the study. The researchers found that biosignatures were particularly clustered in and around alabaster, a type of rock that retains water for long periods, a discovery that can inform geologic studies on other planets.
The researchers called this study a “proof-of-concept” that offers a new path forward in the search for aliens on Mars, and other worlds, though they noted that many improvements could be made to optimize the tool further. For instance, the team envisions the development of a “library” of data about biosignature distribution on Earth that could be applied to extraterrestrial environments of all kinds.
“Such a library could assist future Mars mission scientists in the selection of facies, mineral assemblages and structures with the highest chance of containing biosignatures,” the team concluded. “Ultimately, we hope the approach will facilitate compilation of a databank of biosignature probability and habitability algorithms, roadmaps and models and serve as a guide for exploration on Mars. The framework may also have applications to other astrobiology targets, such as the surface of Titan, the plumes of Enceladus or the ice cover of Europa.”