The Computers Of The Future Will Think Like Brains


Virtually every device you use—from the one you’re using to read this to the pocket calculator growing dust in the back of your desk—relies on the same basic technology: circuits containing many tiny transistors that communicate with each other using electrons. We’ve come a very long way since the room-sized computers of the 1950s, but as computing gets smaller, faster, and more complicated, we get closer to hitting a wall. There’s a physical limit to how powerful traditional computers can get. That’s why scientists are turning to completely new forms of technology for future computers.

The first in our three-part series on the future of computing involves one form you’re familiar with—it’s sitting right inside your skull.

If It Works Like A Brain And Thinks Like A Brain

 These days, we’re not satisfied letting our computers simply run programs and crunch numbers. For tasks like recognizing faces, identifying speech patterns, and reading handwriting, we need artificial intelligence: computers that can think. That’s why scientists figured out a way to build computers that work like brains, using neurons—artificial ones, anyway.

The big difference between an artificial neural network, as it’s called, and a conventional, or algorithmic, computer is the approach it uses to solve problems. An algorithmic computer solves problems based on an ordered set of instructions. The problem is, you have to know what the instructions are first so you can tell the computer what to do. The benefit to this approach is that the results are predictable, but there are definite drawbacks. An algorithmic computer can only do things one step at a time—even though with many components working simultaneously, that can happen surprisingly fast—and you can’t ask it a question you don’t know how to solve.

That’s where neural networks come in. They process information kind of like a brain: a large number of interconnected “neurons” all work at the same time to solve a problem. Instead of following a set of instructions, they do things by following examples. That means that a neural network literally learns how to solve problems based on limited information. Of course, when you don’t know how to solve a problem, you also don’t know what the solution will be. Like your brain, neural networks sometimes arrive at the wrong solutions. That’s the one drawback to neural computing: it’s unpredictable.

Perfect Harmony

 This isn’t to say that artificial neural networks are better than conventional computers. Each system has its own applications. Need some equations solved? Algorithmic computer to the rescue. Need to quickly and accurately detect lung cancer? Neural networks can do that. They can even work together: algorithmic computers are often used to “supervise” neural networks.

Watch And Learn: Our Favorite Content About Neural Networks

Computers That Think Like Humans

Inside A Neural Network

Get deep into the nitty gritty of how a neural network operates.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s