MACHINE LEARNING AND THE SEARCH FOR ALIEN LIFE
NASA researchers have been working on an AI system intended to assist exploration missions identify evidence of life on other planets. This will be accomplished primarily by analyzing soil samples and parsing through relevant data. NASA decided to use machine learning for these tasks because they needed ways to automate certain aspects of geochemical analysis that normally takes up a lot of time. This system will first be utilized during the ExoMars mission that is scheduled to launch in mid-2022.
The Mars rover exploration craft will be capable of drilling more than two meters deep into the martian soil. At this depth, there will be more evidence of microbes and single celled organisms that have not been killed off by the sun’s UV light. NASA also believes that drilling this deep may uncover fossilized evidence of former life on Mars. Using machine learning algorithms the Mars rover will also be able to analyze soil samples on the spot.
The samples collected by the rover will be analyzed by a tool called a mass spectrometer. The main purpose of this tool will be to study the distribution of mass in the ions found in a given soil sample. This is accomplished by freeing up the molecules in a sample using a laser, and calculating the atomic mass from the different molecules. Normally you would need human researchers to analyze the patterns that result from the differing levels of atomic mass from different samples across the spectrum. Utilizing machine learning, however, will allow this part of the analysis to be largely automated as the algorithms have been trained on specific datasets allowing them to identify microbes and other single celled organisms and even prior evidence of their existence.
Analyzing the mass spectrum is a puzzle as different compounds produce a wide variety of different results. Without the assistance of AI it can sometimes take a long time to break down the patterns present in the spectrum and understand what they mean in order to identify meaningful findings. The machine learning algorithms are capable of assisting researchers in cutting through the noise and pointing them in a specific direction.
According to the Goddard Planetary Environments lab this ExoMars mission will be an excellent test case for the AI algorithms designed to help interpret the mass spectrums generated by samples.
AI AND ITS FUTURE ROLE IN SPACE EXPLORATION
As humans venture farther and farther away from Earth in search of life and habitable environments the demand for tools that can analyze mass amounts of data is growing. AI has the potential to help scientists and astronauts in a wide variety of ways. For example we will be able to more accurately analyze a planet’s atmosphere to determine whether it is worth exploring. We may also be able to us AI to design more fuel efficient rockets. No matter how or in what industry we decide to use AI the possibilities appear to be endless.