What is deep learning?

Have you ever wondered how the human brain works? How is it that what is going on in the world is received, processed and reacted? Man is one of the most fascinating creatures we know, and thus the impulse to create an equally intelligent and self-learning machine is awakened.

The human brain and the computer

If we would like to see what the brain really consists of, the first thing that comes to our mind (or not – look at the right side :)) is the neuron. It is a tree that transmits the signal perfectly. The axon, or the trunk of our nerve cell, receives impulses. It can accept thousands of signals simultaneously. If there are enough signals, the cell can pass the information on. Further impulse transmission takes place in the dendrites– that is, our tree crown. This happens through a synapse – that is, the gap that exists between the dendrite of one cell and the axon of the other.

arktur deep learning

It sounds like beautiful magic, doesn’t it? Why not recreate it? Such thoughts appeared about 100 years ago, and neural networks have their origins in the 1940s. Creating a machine that learns by itself is a human dream. Following the example of the human brain, we have created artificial intelligence. Noting that learning was a complex process, the term deep learning emerged. Deep learning. Deep learning aims to enable the machine to:

As you can see, these are natural activities that our brain learns. How to teach this machine?

Multilayering

Deep learning could have arisen due to the use of neural networks, i.e. connections of artificial neurons, similar to those found in nature. As the name suggests, we are dealing here with learning that is not one-level. The machine uses many layers and learns by analyzing the data made available to it. What is special about deep learning, however, is that this process without human supervision. It is a slow process, but everything which is valuable takes time;)

Technology to bear this sweet burden

A large number of layers and algorithms that enable the machine to learn by itself requires a lot of computing power. This is where supercomputers come in, which are the perfect body for such a developed brain. Although it seems unrealistic to us, the current phones would be called supercomputers twenty years ago. For comparison – the first supercomputer could perform 3 million operations per second, and today’s Iphone has 154.9 GFLOPS, which gives about 54 times the speed of computing power. If we allow ourselves to look into the future, we will see a vision of our life in 20 years, which will be greatly supported by artificial intelligence.

.

super komputer

How will deep learning affect our future?

Artificial intelligence has potential in every area of ​​our lives. Soon the problem of queues to doctors will disappear, and thanks to applications that will monitor our condition, they will predict our crises. Farmers will no longer have to look after their crops, because an intelligent machine will do it for them. There are many examples and there are thousands of them. If you want to be up to date with news in the development world – follow us!