Most AI's are actually computers designed to solve problems, or rather to make predictions based on the data you have (such as the human brain). You enter a bunch of data (like numbers 1-10) and ask it to model it (X + 1 starting at 0), and then Artificial Intelligence predicts it. It's not magic, just use the data you have to guess what's going to happen next.
What distinguishes a Artificial intelligence from other computer programs that It does not need to be programmed directly in each case. We can teach you things, machine learning), but also can read alone (Intensive Reading). Although there are individual differences, they can usually be described as follows:
- AI (Artificial Intelligence): a machine capable of mimicking human thinking.
- ML (Machine Learning): a low set of Artificial Intelligence where people "train" equipment to find data-based patterns and make their predictions.
- DL (Deep Learning): a low ML set where the machine is able to think and draw its own ends, reading for itself.
AI (Artificial Intelligence)
The most comprehensive explanation of what Artificial Intelligence is is simply a machine capable of thinking like a human being. It can be as simple as following a logical flow diagram, or it can be a nearly human computer that can read many sensors and apply that information to new situations. This last part is the key, of course: the "powerful" AI everyone thinks can connect all the information they've learned to have ability to deal with any situation.
Currently, AI is far from that; For example, Alexa may be a good business person but she won't be able to pass the famous Turing test. We currently have a limited form of Artificial Intelligence, and are moving away from that sci-fi concept of the Termector Skynet movie.
ML (Machine Learning)
Without machine learning, Artificial Intelligence would simply be limited to using a long list of "if X is true, make Y or someone else do Z". However, this innovation is empowering computers to solve things unless they are clearly organized. As an example of machine learning, let's say that you want a system for identifying cats in photos:
- Give your AI a set of what a cat looks like, so it knows how to recognize it. Colors, shapes, etc.
- Show him the pictures (if it's labeled "cat," the AI will be able to easily point it out).
- Once the system has identified enough cats, you should be able to point them to other images: "If the image contains X, Y, and Z elements, then there is 95% of it being a cat."
Even though it sounds complicated, the following can be summed up: "People tell the computer what to look for, and the computer then analyzes its methods until they have a specific model for what we asked for." It's simple, very useful and is what filters SPAM emails, gives you recommendations on Netflix and turns your work into Facebook.
DL (Deep Learning)
As of 2018, this is a breakthrough for Artificial Intelligence. Think of it as machine learning with deep neural networks
- Give her lots of pictures of cats.
- The algorithm will scan the images to see what they match (idea: they are cats).
- Each image will be reconstructed with many levels of detail, from large, standard sizes to small lines. If the composition repeats too much, the algorithm will mark it as an important feature.
- After analyzing enough images, the algorithm will already know how to identify patterns that define the cat's identity and point you to any other situation.
In short: Deep learning is machine learning, where the computer can read on its own (though it is far superior to cats, of course, because machines are currently able to capture many parameters within images, such as landscape view).
Deep Learning requires much more original data and computing power than Machine Learning, of course, but companies like Facebook or Amazon are starting to get you started.
