Recently, we hear more and more aboutartificial intelligence. It is used almost everywhere: from the field of high technology and complex mathematical calculations to medicine, the automotive industry, and even when using smartphones. The technologies that underlie the work of AI in a modern sense, we use every day and sometimes we may not even think about it. But what is artificial intelligence? How does he work? And is it dangerous?
- 1 What is artificial intelligence
- 2 How artificial intelligence works
- 3 Deep Learning and Neural Networks
- 4 Neural networks - is it an artificial human brain?
- 5 Why deep learning and neural networks are used
- 6 The limits of deep learning and neural networks
- 7 The future of deep learning, neural networks, and AI
What is artificial intelligence?
First, let's define the terminology. If you imagine artificial intelligence as something that can independently think, make decisions, and generally show signs of consciousness, then we hasten to disappoint you. Almost all systems existing today do not even “stand” close to such a definition of AI. And those systems that show signs of such activity, in fact, all the same operate within the framework of predefined algorithms.
Sometimes these algorithms are very, very advanced,but they remain the "framework" within which AI works. There are no “liberties”, and even more so, machines have no signs of consciousness. These are just very productive programs. But they are "the best in their field." In addition, AI systems continue to improve. Yes, and they are arranged quite unbanal. Even if we discard the fact that modern AI is far from perfect, it has a lot in common with us.
How does artificial intelligence work?
AI can do its job first(about which a bit later) and to acquire new skills through deep machine learning. We often hear and use this term. But what does he mean? Unlike the “classical” methods, when all the necessary information is loaded into the system in advance, machine learning algorithms make the system develop independently, studying the available information. Which, in addition, in some cases, the machine can also search independently.
For example, to create a program to detectfraud, machine learning algorithm works with a list of banking transactions and with their final result (legal or illegal). The machine learning model examines examples and develops a statistical relationship between legitimate and fraudulent transactions. After that, when you provide the algorithm with the data of a new banking transaction, it classifies it based on the templates that he outlined from examples in advance.
Generally, the more data you provide,the more accurate is the machine learning algorithm when performing its tasks. Machine learning is especially useful in solving problems where the rules are not defined in advance and cannot be interpreted in a binary system. Returning to our example with banking operations: in fact, we have a binary system of calculation at the output: 0 - legal operation, 1 - illegal. But in order to come to this conclusion, the system needs to analyze a whole bunch of parameters and if you enter them manually, it will take more than one year. And to predict all the options anyway will not work. And a system that works on the basis of deep machine learning will be able to recognize something, even if it hasn’t met exactly in such a case before.
Deep Learning and Neural Networks
Whereas classic machine algorithmstraining solve many problems in which there is a lot of information in the form of databases, they do not cope well with, so to speak, "visual and audio" data such as images, videos, sound files, and so on.
For example, creating a cancer prediction modelBreast cancer using classic machine learning approaches will require the efforts of dozens of medical experts, programmers and mathematicians, ”said AI researcher Jeremy Howard. Scientists would have to make many smaller algorithms in order for machine learning to cope with the flow of information. A separate subsystem for the study of x-rays, a separate one for MRI, another for the interpretation of blood tests, and so on. For each type of analysis we would need our own system. Then they would all be combined into one large system ... This is a very difficult and resource-consuming process.
Deep learning algorithms solve the samea problem using deep neural networks, a type of software architecture inspired by the human brain (although neural networks are different from biological neurons, the principle of operation is almost the same). Computer neural networks are communications of "electronic neurons" that are capable of processing and classifying information. They are sort of “layers” and each “layer” is responsible for something of its own, eventually forming the overall picture. For example, when you train a neural network on images of various objects, it finds ways to extract objects from these images. Each layer of the neural network discovers certain features: the shape of objects, colors, the appearance of objects, and so on.
The surface layers of neural networks detectcommon features. Deeper layers already reveal actual objects. The diagram shows a simple neural network. Green indicate input neurons (incoming information), blue indicate hidden neurons (data analysis), yellow indicate output neuron (solution)
Are neural networks an artificial human brain?
Despite the similar structure of the machine andhuman neural network, they do not possess signs of our central nervous system. Computer neural networks are essentially the same auxiliary programs. It just so happened that our brain turned out to be the most highly organized system for performing calculations. You probably heard the expression "our brain is a computer"? Scientists simply “repeated” some aspects of its structure in a “digital form”. This only allowed to speed up the calculations, but not to endow the machines with consciousness.
This is interesting: When does artificial intelligence learn to reason?
Neural networks have existed since the 1950s (according toat least in the form of concepts). But until recently, they did not receive much development, because their creation required enormous amounts of data and computing power. In the past few years, all this has become available, so the neural networks have come to the fore, having received their development. It is important to understand that for their full appearance there was not enough technology. How they are not enough now to bring technology to a new level.
What is deep learning and neural networks used for?
There are several areas where these two technologies have helped make significant progress. Moreover, we use some of them daily in our lives and do not even think about what is behind them.
- Computer vision is an abilitysoftware to understand the content of images and videos. This is one area where deep learning has made great progress. For example, deep learning image processing algorithms can detect various types of cancer, lung disease, heart disease, and so on. And do it faster and more efficiently than doctors. But deep learning is also rooted in many of the applications you use every day. Apple Face ID and Google Photos use deep learning to recognize faces and improve image quality. Facebook uses deep learning to automatically tag people in uploaded photos, and so on. Computer vision also helps companies automatically identify and block questionable content such as violence and nudity. And finally, deep learning plays a very important role in enabling self-driving cars so that they can understand what surrounds them.
- Voice and speech recognition. When you say a command for your Google Assistant, deep learning algorithms transform your voice into text commands. Several online applications use deep learning to transcribe audio and video files. Even when you “shazamit” a song, algorithms of neural networks and deep machine learning come into play.
- Web search: even if you are looking for something in a search engine, so that your request is processed more clearly and the results are as correct as possible, companies have begun to connect neural network algorithms to their search engines. So, the performance of the Google search engine has grown several times after the system switched to deep machine learning and neural networks.
The limits of deep learning and neural networks
Despite all its advantages, deep learning and neural networks also have some disadvantages.
- Data Dependence: In general, deep learning algorithms require a huge amount of training data to accurately complete their tasks. Unfortunately, to solve many problems, insufficient quality training data is needed to create working models.
- Unpredictability: neural networks are developing in some strange way. Sometimes everything goes as planned. And sometimes (even if the neural network does its job well), even the creators struggle to understand how the algorithms work. The lack of predictability makes it extremely difficult to eliminate and correct errors in the algorithms of neural networks.
- Algorithmic Offset: deep learning algorithms are as good as the data on which they are trained. The problem is that the training data often contains hidden or explicit errors or omissions, and the algorithms get them “inherited”. For example, a face recognition algorithm trained mainly in photographs of white people will work less accurately in people with a different skin color.
- Lack of generalization: Deep learning algorithms are good for performing purposeful tasks, but poorly generalize their knowledge. Unlike people, the deep learning model trained to play StarCraft will not be able to play another similar game: say, in WarCraft. In addition, deep learning does not do well in processing data that deviates from its training examples.
The future of deep learning, neural networks and AI
It’s clear that work on deep learning andneural networks are far from complete. Various efforts are made to improve deep learning algorithms. Deep learning is an advanced method in creating artificial intelligence. It has become increasingly popular in the last few years due to the abundance of data and increased computing power. This is the core technology that underlies many of the applications we use every day.
But will it ever be born on the basis of thistechnology consciousness? Real artificial life? Some scientists believe that at the moment when the number of connections between the components of artificial neural networks approaches the same indicator that exists in the human brain between our neurons, something like this can happen. However, this statement is very doubtful. In order for a real AI to appear, we need to rethink the approach to creating systems based on AI. All that is now is only application programs for a strictly limited range of tasks. How would we not want to believe that the future has already come ...
What do you think? Will people create AI? Share your opinion in our telegram chat.