Artificial intelligence: basic tasks and methods based on real life examples

People often confuse (and sometimes equate) the concepts of “neural network” and “artificial intelligence”. In this material, I will describe with examples from real life several methodologies and tasks of artificial intelligence and will try to show that artificial intelligence is not only neural networks.

Expert systems
One of the central concepts in many branches of artificial intelligence is knowledge. This is information along with a way of interpreting it. If the system not only possesses information, but is also able to explain it, it means that it “understands” or possesses knowledge.

There are various subject areas, including very narrow ones. Experts in these fields are able to draw conclusions based on inputs and their knowledge. An experienced racer can tell the brand of a passing car by the sound of the engine. The doctor, only by looking at the patient’s symptoms, can diagnose. For ordinary people, not experts, it is more difficult, and sometimes even beyond their power. The more specific and narrower the subject area, the more difficult it is to find an expert in it. But what if you collect the knowledge of experts and replicate it? For example, a doctor knows that if a patient has lupus, they should be given steroids. Unfortunately, not all patients with this disease can get to a large medical center. But what if even the village nurse could prescribe the right treatment by simply starting up the computer and typing in the patient’s symptoms?

For this, expert systems are used. They consist of two main parts: the inference engine and the knowledge base. If knowledge is presented in a certain form, you can use the same inference engine for different areas of knowledge with minor settings.