Google’s algorithms are now piloting its Internet beaming balloons


Artificial Intelligence can do a lot of things. It still struggles with some of them and has a lot to learn (read, car driving!) however, it has also gotten unbelievably good at doing some other things. And of course, we are coming upon new applications of the systems everyday. Take Google’s AI for example. The company has begun using its Artificial Intelligence systems to pilot its Internet beaming balloons.

The Google X labs is known best for its balloons. The company launches them up into the stratosphere and they provide it with all sorts of data. However, the main — and significantly more ambitious — use Google is planning for these balloons comes under the Project Loon. The company hopes to use these balloons to provide economically backwards and geographically inaccessible regions with Internet.

The idea is pretty simple. The balloons float over areas that would otherwise be inaccessible or very hard to get to and beam down Internet. Indeed, they could also be made to float around that particular area to provide a more or less permanent Internet connection. However, balloons tend to float away and Google’s control over them is limited to moving them up or down.

Yup. The balloons can not be moved sideways or fitted up with a propulsion system as it would make the setup to heavy and expensive. So, its all about letting nature take its course and using the winds — by moving your balloon up or down at the right time — to go where you want to, or stay at your current position.

Which is exactly why when Google managed to keep its balloon in the Peruvian airspace for over three months, it was widely regarded as a fantastic job. Turns out that Google had some help — in the form of Algorithms (or AI if you will permit the term) specially designed for the purpose. The basic idea here, was same as everywhere else. You start of at the beginning and as the time progresses and you accumulate data, you feed it to your algorithms, letting it learn from it. This is a concept that is at the core of Machine learning and Google’s balloons are yet another example.

Meanwhile, as per Sal Candido, the former Google search engineer who worked on Loon.

We have more machine learning in more of the right places. These algorithms are handling things more efficiently than any person could.

Bad news for professional balloon fliers i suppose. AI is beginning to do it better. do it better. It certainly wasn’t easy and the company had to collect and analyze data for over 17 million kilometers of balloon flights before it got its AI to this level. However, its finally here and is giving Google’s project Loon a huge boost.

Meanwhile, the trend to assimilate and include machine learning — particularly that based on neural networks –continues all major corporations. Google, Microsoft, Facebook and other are all increasingly trying to automate their processes using machine learning. Meanwhile, many other companies who wouldn’t usually deal with this kind of technology — such as automobile makers — are now doing so and indeed, are pinning their hopes on machine learning to bring a revolution across their respective industries.

Good time to be a machine learning expert or a data analyst I suppose.



Leave A Reply