Neural networks: from person to machine

Neural Network In 5 Minutes | What Is A Neural Network? | How Neural Networks Work | Simplilearn

Artificial intelligence creation — a task over which the best minds of mankind are struggling. Its solution will become the starting point of a revolution that can change people’s lives and provide answers to many questions. Find out what opportunities machine learning can bring to humanity when the first device that can fully simulate the human brain appears, and how neural networks of machines will change the world of cryptocurrencies, find out in our material.

What is machine learning

Under the term «machine learning» means a complex for teaching technology to perform intellectual tasks. It is based on the desire to shift the functions that a person traditionally performs on metal «shoulders» technology. Based on the specifics of the work, machine learning belongs to the field of creating artificial intelligence, a subsection of the formation of neural networks.

One of the important problems is getting to multitasking. For this, active development is underway to sublimate all available techniques and then combine them into fully functional machine intelligence..

Why smart cars are needed

The ability of technology to use human-like intelligence can answer many questions. Human resources, as well as his capabilities, are extremely limited. He cannot analyze more than a certain amount of information. In addition, a person is subject to feelings that are capable of creating a barrier of perception. The machines — on the contrary, they are able to absorb an intellectual component, protected from possible negative influences.

Machine resources allow:

  • significantly exceed the intellectual capabilities of a person;
  • form unlimited memory and speed of computing processes;
  • organize full-scale analysis capabilities, etc..

The totality of possibilities can become the key to the complete sublimation of all the experience and knowledge of mankind. However, there are other points of view: «The development of artificial intelligence can be both the most positive and the most terrible factor for humanity. We must be aware of the danger it poses», — spoken by Stephen Hawking.

Drivers for the development of machine intelligence

Human imagination, science and nature are at the helm of the development of machine intelligence. Their combination allows you to create the most variable points of perception, revealing new facets of possible embodiment..

The role of human fantasy, in particular cinema, cannot be underestimated in the development of machine learning. The flight of human imagination has become a support for the advancement of science. Allowed to create new, sometimes completely unexpected incarnations of technologies. The images presented to us by the cinema quite often form the basis of inventions. Science fiction has played a critical role for artificial intelligence.

In addition to trying to draw inspiration from images created by people of art, scientists also take many decisions from nature. An example that provided answers to many questions was the Hogan sisters phenomenon. They are Siamese twins, fused heads (craniopagi). The unusualness of the phenomenon lies in the fact that the girls have fused brain tissue. A unique structure has formed between two healthy brains, called the thalamic bridge (channel of neural processes), thanks to which girls can transmit to each other:

  • arbitrary images;
  • desires;
  • sensations and so on.

Such unique examples provide an opportunity to see the solution to the problem of information transmission. Perhaps in the future, a person will receive the perfect embodiment of artificial intelligence..

Bionics plays an extremely important role in solving the problem of transferring human intelligence to a machine. According to science, the key to solving a problem is understanding the mechanisms of the human mind. So, like Leonardo da Vinci trying «steal» the idea of ​​flying in birds, scientists give the available knowledge a technical embodiment.

Neurophysiology has become the main related direction for solving the problem. The data in this area were the first to be used to solve mathematical problems of artificial intelligence..

Neural networks: from person to machine

The problem lies in the difference in the principles of reproduction, assimilation, processing and storage of information. In an attempt to give humanity to a machine, an actual transfer of the principle of the neural network is necessary. The total resource of the human brain significantly exceeds the capabilities of modern technology. However, Ray Kurzweil in 2006 in his book «Singularity is near» conducted research according to which, while maintaining the current pace of increasing computing power, by the end of 2020, humanity will receive a machine that can fully simulate the brain.

Neural network machine learning capabilities

Since machine learning is not only a way to sublimate all available information, but also the main engine of possible progress, its possibilities are endless. It is impossible to single out one area in which artificial intelligence could excel. To understand the possibilities, it is necessary to give the most scattered examples..

The medicine

A machine capable of matching data as quickly as possible can produce accurate diagnoses in no time. Accordingly, artificial intelligence in medicine will be equal to a decrease in the mortality rate, an increase in the effectiveness of treatment, and an expansion of diagnostic capabilities..


A machine capable of recognizing personality traits can build an effective training program tailored to a specific person. Result — knowledge growth, professional development and widespread progress.


The control transport infrastructure — the basis of security. Lack of human factor — basis for emotionless decision-making on effective management with the ability to calibrate all indicators. Result — increasing the level of comfort of life, increasing safety.

There is no area in which artificial machine intelligence cannot find application. The number of examples is limitless.

One of the most sensational examples of the modern capabilities of neural networks is the DeepFakes phenomenon (the nickname of the user that has become the network’s eponymous name). DeepFakes created an algorithm on the basis of which the computer was able to learn how to analyze Youtube videos (displaying emotions on faces, body position depending on the situation, facial expressions). Later, the technology was applied to create fake XXX videos. The loudness of the phenomenon lies in the fact that videos with celebrities were created in this way, and they looked as believable as possible. So, based on the algorithm, anyone can compose a similar video, having a video source and a photo of the person’s face..

«There is a huge area of ​​possible applications of such algorithms, ranging from elements of augmented reality, which will laconically fit into certain areas of the footage, to large-scale detailed projections. The Prizma application, which has been downloaded over 100 million times in the App Store and Google Play, works on a similar principle. Applying such algorithms sequentially to each frame of the video sequence, we seem to find ourselves in the drawn world. In the future, it will be possible to replace absolutely any objects in the video sequence: people, animals, cars, etc. And despite the fact that the quality of such fakes is still low, the ongoing development of the algorithms used may cast doubt on any video evidence. As you know, action gives rise to opposition, and at present, algorithms are also being developed that can distinguish the generated video from the real one.», — shared with BitCryptoNews Alexander Sizov, an expert in the analysis of big data of the company «Innodata».

Neural networks in cryptocurrencies

The field of cryptocurrencies has also been able to profit from the development of artificial intelligence. New financial instrument — stress for the financial system. In attempts to systematize the work with it, machine learning capabilities have become a key element. Formally, the use of neural networks when working with cryptocurrency can be divided into the following components:

  • forecasting;
  • capital Management.

1) Forecasting cryptocurrencies based on neural networks allows you to eliminate the risks associated with high asset volatility. Classic technical and fundamental analyzes are complemented by analytical data that greatly simplifies the work on disclosing the further price movement.

2) Capital management through artificial intelligence is an opportunity to get away from the problems generated by the psychology of the market. Rationality and mathematical validity open up opportunities for the most effective control over assets.

Neural networks in the world of cryptocurrencies allow expanding human understanding of the very phenomenon of a new financial instrument, its potential and possibilities of practical application. It is likely that in the future, control over all world capital will belong to machines running on programmed neural networks that can independently learn from newly acquired data..

Thus, the development of machine learning can open many doors for us. However, it must be understood that intelligence is equal to power that can be used both for the good and against humanity..

text: Evgeniya Likhodey, photo: iStock, Burst, Unsplash   

Neural networks: from person to machine

Similar articles