Breakthroughs in the area of artificial intelligence are discussed in the media more and more often. Controlling robots, piloting fighters and medical diagnostics are just a few things you could do with this advanced technology. But what is artificial intelligence really? What is machine learning and deep learning? Let’s sort the whole thing out!
Introduction to Artificial Intelligence
Artificial Intelligence (AI) is a simulation of human intelligence processes by machines (computer systems). AI systems typically exhibit at least some of the following behaviours related to human intelligence: planning, learning, reasoning, problem solving, knowledge representation, movement and manipulation, and to a lesser extent: social intelligence and creativity.
At a very high level, artificial intelligence can be divided into two general types: narrow and general artificial intelligence. The first of the specified types is a system designed and trained for a specific task. Virtual assistants of this sort include such examples as Apple’s Siri or our Toucan Eye, which supports people with visual impairment.
The second type of artificial intelligence is a system with generalized human cognitive abilities. In this case, AI is confronted with an unknown task – it manifests enough intelligence to find the right solution. These tasks can be: from haircuts to creating spreadsheets or to reasoning on complex topics.
When machines learn on their own…
When a computer system is “fed” with large amounts of data which it then uses to learn how to perform certain tasks, machine learning takes place. Thanks to this, an AI-based system can independently carry out tasks of various levels of difficulty and complexity, such as speech comprehension or word recognition.
The key to the machine learning process are neural networks. Imitating the structures of the human brain, they are networks of interconnected layers of “neurons”, which transmit data to one another. Neural networks can be “trained” to perform specific tasks – this is done by modifying the meaning assigned to input data (as it gradually passes between layers).
A subset of machine learning is deep learning, in which neural networks are expanded into a large number of layers and are taught with even larger amounts of data. These deep neural networks are the ones that fuel the current trend of using computers for tasks such as speech recognition and computer vision.
The subject of artificial intelligence develops dynamically and the area itself brims with valuable insight. In this article, I have focused solely on the most general issues. But I assure you, this is just the beginning. AI is a fascinating topic — in the following articles we will learn more of its secrets. I give you my word.
ARTUR ZACNIEWSKI
Head of Research and Development Department