Artificial Intelligence And Neural Networks

Artificial intelligence has proved its power, and has many accomplishments under its belt. AI has beaten humans at games like Go, can power search engines, and can even detect emotion in human speech. But AI has a new skill under its belt, easily the most powerful and extraordinary of all.

The ability to learn.

Neural networks are AI programs that mimic the human brain, hence the name, and use that structure to learn how to do things. It is a fairly complex process, with the program starting off extremely basic. For a neural network to learn anything, you need to give it a task. The neural network will try to accomplish that task, and although it begins not doing well, if it does anything at all, eventually it teaches itself how to do something.

This is an overall list of the things AI can do, with many examples I use taken from this list.

Neural networks can theoretically applied to almost anything involving technology, but they can especially be applied to anything where technology may need to operate on its own.

For instance, neural networks could be applied to search engines to know what people will want to look up, and know what advertisements to show to different people. There has even been an instance where a neural network was used to allow a computer to get progressively better at speedrunning Super Mario World, with the computer starting off not even moving, and eventually getting really good at efficiently getting through a level. You could theoretically get a car to teach itself how to drive.

The importance of neural networks is not what they are currently capable of, but what they will be able to do in the future.

AI will become less dependent on people. Machines can operate more efficiently, search engines can be more accurate, computers will in general become smarter. Programs will be far more capable of creating more things, figuring things out, and in general learning.
The possibilities of what AI can do are limitless.