Artificial intelligence or AI is currently one of the most popular term related to computer science. Certainly, everyone has some associations with the phrase ‘artificial intelligence’. Some people might see an image of happy future full of robots helping people in managing a wide range of tasks, while other imagine something dystopian where artificial intelligence suddenly became mean for people. Of course, this is the simplest view of AI but the majority of us thinks about artificial intelligence exactly like this. Yet, do you really know what AI is?
Algorithms used for artificial intelligence
Although we might tend to imagine AI as some artificial creatures looking in a more or less robotic way, AI itself is simply an algorithm. It is not any special hardware element added to a computer. So far, AI is a pure code. Still, this kind of a code varies from classic algorithms a lot.
A classic algorithm is a set of strict rules that we tell the computer in order to let it create a particular result for us. Classic algorithms describe all the commands step by step and a computer can only execute them for us. The majority of software created in the previous years was based exactly on this approach.
Artificial intelligence is also using algorithms, however, in its case the algorithms are of a particular type that is known as machine learning algorithms. As the name suggests, these algorithms are created in order to let our machine to learn. Based on some principles of learning that is inherent in people, such algorithms allow us to train our computers in order to solve particular tasks. How exactly does it work?
One of the major types of machine learning algorithms is called supervised learning. In case of supervised learning, we can teach our computers how to solve similar problems in the future by supplying them with sets of data. For example, knowing the actual amount of CO2 emission by particular car models, we can predict the possible emission for some other car models. Algorithms of this kind are usually based on regression.
Another popular type of machine learning algorithms is called unsupervised learning. As the name suggests, in its case we won’t supervise the learning process of a computer, albeit it doesn’t mean that we won’t supply it with any data. Data is still the major component of unsupervised learning, but its type will differ from the one used for supervised learning. For example, classification is a very popular category of unsupervised machine learning algorithms. Such algorithms are used for training computers in such a way that they will be able to find some patterns or similar features between data points. Basing on these features, the data can be divided into categories which in many cases is very useful. For instance, a computer using unsupervised classification algorithms can be learnt how to distinguish a malignant cancer cell on photos.
Other popular types of machine learning algorithms are clustering, associations, anomaly detection, sequence mining, dimension reduction and recommendation systems.
The terms related to AI that are usually misused
Now, when you really understand what AI stands for, let’s look on some widespread terms connected to artificial intelligence that are generally misused by many people.
First of all, artificial intelligence is a very broad field that covers a wide variety of other fields. For example, AI chat boxes, visual recognition systems and augmented reality are all belong to AI. Yet, machine learning is also a part of artificial intelligence. It is simply the subject that focuses exactly on the ways in which computers can be trained.
As you can see, it means that machine learning is a part of the artificial intelligence field.
There is also a highly widespread term called deep learning. If you come across it, you should understand that deep learning is a set of advanced techniques used ibn machine learning that are usually using neural networks. So, the parent field is artificial intelligence, machine learning is a part of AI whereas deep learning is a part of machine learning.
There is also an important term called robotics. Certainly, if a particular robot was built using machine learning algorithms, we can say it is using artificial intelligence. Still, you should understand that robotics itself is a very large field that incorporates many other things not connected to machine learning at all.
Now, you can see the differences between the major terms related to artificial intelligence and also know what AI exactly is.