Let’s be honest, it’s much easier for robots to explore space than us humans. Robots don’t need fresh air and water to survive, or to carry loads of food. However, it requires people to manage them and make decisions. Advances in machine learning technology could change that by making computers more active collaborators in planetary science.

At the Fall 2022 Meeting of the American Geophysical Union (AGU) last week, planetary scientists and astronomers discussed how new machine learning techniques are changing the way we learn about our planet. solar systemFrom planning future mission landings to Jupiter’s icy moon Europe until the discovery of volcanoes. tiny Mercury.

Machine learning is a way of teaching computers to detect patterns in data and then use those patterns to make decisions, predictions or classifications. Aside from the fact that computers don’t require life support, another big advantage is their speed. For many tasks in astronomy, it can take months, years, or even decades for people to review all the data they need.

An example is the identification of rocks in photographs of other planets. For a few rocks, say, “Hey, there’s a boulder!” but imagine doing this a thousand times. This task will become quite tedious and consume valuable work time of scientists.

“You can find hundreds of thousands of rocks up to 10,000, and that takes a very long time,” said Niels Prior, a planetary scientist at Stanford University in California during a presentation at AGU. Prier’s new machine learning algorithm can detect rocks in anything months in just 30 minutes. It is important to know where these large boulders are to ensure that new missions can land safely at their destination. The rocks are also useful for geology, providing clues as to how impacts broke the surrounding rocks and formed craters.

Computers can also identify a number of other planetary phenomena: explosive volcanoes on Mercury, eddies in the thick atmosphere Jupiter and craters on the moon, etc.

During the conference, planetary scientist Ethan Duncan of NASA’s Goddard Space Flight Center in Maryland demonstrated how machine learning can identify fragments, not rock fragments. ice Jupiter’s icy moon Europa. The so-called Chaos Land is a dirty-looking area of ​​Europa’s surface with shiny chunks of ice scattered against a darker background. Europa, with its subterranean ocean, is a primary target for astronomers interested in extraterrestrial life, and mapping these ice floes will be key to planning future missions.

Future missions may also include artificial intelligence as part of the team using the technology to enable the probes to respond to hazards in real time and even land autonomously. Landing is a notorious test for spacecraft and is always one of the most dangerous times on a mission.

“Seven minutes of terror” on Mars [під час спуску та приземлення] “It’s something we talk about a lot,” said Bethany Tailing, a planetary scientist at NASA Goddard. “Things get much more complicated as we move forward through the solar system. We have a communication delay of hours.”

A message from a probe that landed on Saturn’s methane-filled moon Titaniumit will take a little less than an hour and a half to get back World. Once the human response reaches its destination, the communication cycle will take almost three hours. In a situation where real-time responses are required, such as landing, this kind of back and forth with Earth won’t cut it. According to Tailing, machine learning and artificial intelligence can help solve this problem by giving the probe the ability to make decisions based on observations of its environment.

“Scientists and engineers, we’re not trying to get rid of you,” Tailing said. “What we’re trying to say is that the time you spend on this data will be the most useful time we can manage.” Machine learning won’t replace humans, but we hope it can be a powerful addition to our scientific discovery toolkit. Source