Alphabet’s Loon, the team responsible for beaming internet down to Earth from stratospheric helium balloons, has achieved a new milestone: its navigation system is no longer run by human-designed software.
A reinforcement learning (RL) system is now in charge of navigation for a fleet of balloons over Kenya, where Loon switched on its first commercial service earlier this year. Loon says this is the first use of an RL model in “a production aerospace system.” It also noted the “development is exciting because it shows that reinforcement learning can be applied to real-world use cases.” RL systems have previously figured out how to beat top Go and Dota 2 players.
Loon’s AI can figure out the optimal route for balloons significantly faster than the previous navigation system. It does so with more efficiency as well. Balloons can travel similar or greater distances than before with less power. Loon’s record flight duration is 312 days it set that benchmark earlier this year. Perhaps the AI system will be able to keep balloons aloft for even longer.
Loon and Google AI used simulations to train the RL model through trial and error before real-world testing began in Peru. The team then assessed its capabilities directly against a human-crafted system called StationSeeker with a 39-day test over the Pacific Ocean. The AI was able to keep balloons in target areas for longer periods while using less energy. That’s important, as it will help to provide more consistent internet coverage to people in a given area.
The AI-controlled system handily outperformed the human one by consistently staying closer to a device the team uses to measure LTE signals in the field, and that test paved the way for more experiments to prove the efficacy of the system before it formally replaced the one the team had spent years building by hand. Loon now thinks its system can “serve as a proof point that RL can be useful to control complicated, real world systems for fundamentally continual and dynamic activity.”
In his closing remarks, Candido touches on the concept of whether this type of AI is worthy of the name, because of how specialized it is and how closely it resembles a traditional but not self-learning, automated system like the ones that operate heavy machinery or control elements of mass transit.
“While there is no chance that a super-pressure balloon drifting efficiently through the stratosphere will become sentient, we have transitioned from designing its navigation system ourselves to having computers construct it in a data-driven manner,” he says. “Even if it’s not the beginning of an Asimov novel, it’s a good story and maybe something worth calling AI.”
