From massive formless artificial intelligence to Data’s “positronic brain” on Star Trek: The Next Generation, neural networks have long played a key part in science fiction’s imagination. However, neural nets are actually far more ingrained into everyday life today in ways you may not immediately suspect. These are just a few real-world applications for neural networks today.
We use social media so much that it has, in just over a decade, become an integral part of our lives. It’s such second nature at this point that it probably doesn’t even phase you that when you post a picture of yourself and some friends to Facebook, there’s a good chance it already has an idea of who should be tagged along with you. Facebook is able to do this via its facial recognition programs, which are themselves an outgrowth of neural networks.
The same holds true for Instagram. It is able to predict which emoji you want to use next based on your own tendencies and those of others, which it is able to store and access in real-time using neural nets.
Browsing on Amazon can be a nightmare for your wallet. The problem isn’t that you don’t find what you’re looking for, but rather that you often find so many good recommendations based on what you’re looking for that you buy more than you bargained for – which, of course, is part of Amazon’s goal for making its enterprise so profitable.
Your search history is one thing, but what about when you type something into the search bar? All it takes is a couple of letters or a word or two at most to clue the search engine into what you’re looking for. It then presents results which are more often than not very well suited to what you are looking for. That increased buyer recognition is an outgrowth of neural nets being able to predict your tendencies based on your own history and those of shoppers as a whole.
If you buy a bunch of things at once with a credit card, you’ll likely get a notice asking if you did, in fact, approve those transactions and, if you didn’t, to freeze the card immediately so as to protect against fraud. How does your bank know to do this so quickly? The answer, once again, is neural nets, which are able to flag potentially suspicious transactions based on the speed of purchases made in succession, along with the location, price, and other factors.
Far from mere sci-fi fantasy, neural nets are a very real part of our everyday life.
Originally published to peterlagow.com