ARTIFICIAL INTELLIGENCE

What is this?

Definition of AI
Artificial intelligence (AI) is a branch of computer science. The term is used to describe “intelligent” computer systems that attempt to simulate human intelligence. The machines follow a certain scheme in which all aspects are taken into account. The machines can learn independently, draw conclusions and correct themselves.
A general definition is difficult, because the terms “intelligence” and “intelligent human behavior” are not very precisely defined themselves.

Types of artificial intelligence
A distinction can be made between strong AI and weak AI

• Strong AI means that human intelligence is “copied”, i.e. the machine receives consciousness
• With the weak AI, the machine simulates human thinking through mathematics and computer science

Furthermore, the AI can be divided into more types:

type 1
Reactive AI
This is a basic form of AI. The environment is directly recognized but has no idea of the outside world. The decisions are based only on the area without any experience.

Type 2
Memory limiter
Systems with limited memory take past information into account and add it to their representations. The memory is sufficient to make correct decisions and perform actions. Chat bots are an example of this.

Type 3
Own consciousness
The systems can develop their own consciousness and recognize and know themselves and others when others act wrong. The systems understand human behavior and feelings and can interact socially.

Type 4
Systems with self-knowledge
The systems can form ideas about themselves and consciously predict the feelings of others and draw consequences. These systems are intelligent and sentient. They are based on type 3.

Methods
The first method consists of machine learning or also called deep learning.
The machine learns independently based on algorithms and codes. The principle is to learn from data collected from many different sources, thereby recognizing patterns and establishing rules.

You take a large number of sample images and label them with what you see on them. The program automatically looks for similarities between the images and applies the learned “rules” to new images. So the more data the machine has, the better the program becomes. By feedback, adjustments can be made and the program improved.

The tasks of a program are usually so complex that one almost always speaks of deep learning: e.g. smartphones or autonomous driving.
Another method of AI to display meaningful content is the search and research help.

It is the attempt to understand the meaning of a task, so that results can be displayed. One form of this is language analysis, which examines the structure and composition of words.

The form of knowledge enhancement is about adding content to specific knowledge, which can be used to identify connections.
There is also the knowledge model, in which references are made and connections between terms can be established.

Vollcom Digital GmbH
Thierschstraße 34, 80538 München
info@vollcom-digital.de