We've been hearing more and more about artificial intelligence lately. It is used almost everywhere: from high technology and complex mathematical calculations to medicine, the automotive industry and even smartphones. The technologies that underlie the work of AI in the modern view, 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?
What is artificial intelligence
First, let's define the terminology. If you imagine artificial intelligence as something capable of thinking independently, making decisions, and generally showing signs of consciousness, then we hasten to disappoint you. Almost all systems existing today do not even come close to this definition of AI. And those systems that show signs of such activity, in fact, still operate within the framework of predetermined algorithms.
Sometimes these algorithms are very, very advanced, but they remain the "framework" within which the AI works. Machines have no "liberties" and even more so signs of consciousness. They are just very powerful programs. But they are "the best at what they do." Plus, AI systems continue to improve. And they are not trivial at all. Even putting aside the fact that modern AI is far from perfect, it has a lot in common with us.
How artificial intelligence works
First of all, AI can perform its tasks (about which a little later) and acquire new skills thanks to deep machine learning. We also often hear and use this term. But what does it mean? Unlike “classical” methods, when all the necessary information is loaded into the system in advance, machine learning algorithms force the system to develop independently, studying the available information. Which, moreover, the car in some cases can also search for itself.
Promotional video:
For example, to create a program to detect fraud, a machine learning algorithm works with a list of bank transactions and their end result (legal or illegal). A machine learning model looks at examples and develops a statistical relationship between legitimate and fraudulent transactions. After that, when you provide the algorithm with the details of a new bank transaction, it classifies it based on the patterns that it drew from the examples beforehand.
Typically, the more data you provide, the more accurate the machine learning algorithm becomes when performing its tasks. Machine learning is especially useful for solving problems where the rules are not predefined and cannot be interpreted in binary. Returning to our example with banking operations: in fact, at the output we have a binary numbering system: 0 - legal operation, 1 - illegal. But in order to come to such a 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 based on deep machine learning will be able to recognize something, even if it has never encountered exactly such a case before.
Deep Learning and Neural Networks
While classical machine learning algorithms 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 auditory” data like images, videos, sound files, and so on.
While classical machine learning algorithms 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 auditory” data like images, videos, sound files, and so on.
For example, building a predictive model for breast cancer using classic machine learning approaches will require dozens of medical experts, programmers and mathematicians, says AI researcher Jeremy Howard. Scientists would have to make many smaller algorithms for machine learning to cope with the flow of information. A separate subsystem for studying X-rays, a separate one for MRI, another for interpreting blood tests, and so on. For each type of analysis, we would need its own system. Then they would all be combined into one big system … This is a very difficult and resource-intensive process.
Deep learning algorithms solve the same problem using deep neural networks, a type of software architecture inspired by the human brain (although neural networks are different from biological neurons, they work much the same). Computer neural networks are connections of "electronic neurons" that are capable of processing and classifying information. They are arranged as if in "layers" and each "layer" is responsible for something of its own, eventually forming a general 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 detects certain features: the shape of objects, colors, the appearance of objects, and so on.
The surface layers of neural networks show common features. Deeper layers are already revealing the actual objects. The figure shows a diagram of a simple neural network. Input neurons (incoming information) are shown in green, blue - hidden neurons (data analysis), yellow - output neuron (solution).
Are neural networks an artificial human brain?
Despite the similar structure of the machine and human neural networks, they do not possess the features of our central nervous system. Computer neural networks are essentially all the same auxiliary programs. It just so happens that our brain is the most highly organized system for computing. You've probably heard the expression "our brain is a computer"? Scientists simply "replicated" some aspects of its structure digitally. This allowed only to speed up calculations, but not to endow the machines with consciousness.
Neural networks have been around since the 1950s (at least in the form of concepts). But until recently, they did not receive much development, because their creation required huge amounts of data and computing power. In the last few years, all this has become available, so neural networks have come to the fore, having received their development. It is important to understand that there was not enough technology for their full-fledged appearance. How they are not enough now in order to bring the 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 notable progress. Moreover, we use some of them every day in our life and do not even think about what is behind them.
- Computer vision is the ability of software to understand the content of images and videos. This is one area where deep learning has made a lot of progress. For example, deep learning image processing algorithms can detect various types of cancer, lung disease, heart disease, and so on. And to do it faster and more efficiently than doctors. But deep learning is also ingrained in many of the applications you use every day. Apple Face ID and Google Photos use deep learning for facial recognition and image enhancement. Facebook uses deep learning to automatically tag people in uploaded photos and so on. Computer vision also helps companies to automatically identify and block questionable content such as violence and nudity. And finallydeep learning plays a very important role in making cars self-driving so they can understand their surroundings.
- Voice and speech recognition. When you speak a command to your Google Assistant, deep learning algorithms translate your voice into text commands. Several online applications use deep learning to transcribe audio and video files. Even when you shazam a song, neural networks and deep machine learning algorithms come into play.
- Internet search: even if you are looking for something in a search engine, in order for your request to be processed more clearly and the search results to be as accurate as possible, companies have begun to connect neural network algorithms to their search engines. Thus, the performance of the Google search engine has increased several times after the system switched to deep machine learning and neural networks.
The limits of deep learning and neural networks
Despite all their advantages, deep learning and neural networks also have some disadvantages.
- Data Dependency: In general, deep learning algorithms require huge amounts of training data to accurately perform their tasks. Unfortunately, to solve many problems, there is not enough high-quality training data to create working models.
- Unpredictability: Neural networks evolve in some strange way. Sometimes everything goes as planned. And sometimes (even if the neural network does a good job), 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 bias: Deep learning algorithms are just as good as the data they are trained on. The problem is that training data often contains hidden or obvious errors or flaws, and algorithms inherit them. For example, a facial recognition algorithm trained primarily on photographs of white people will work less accurately on people with a different skin color.
- Lack of generalization: Deep learning algorithms are good for performing targeted tasks, but poorly generalize their knowledge. Unlike humans, a deep learning model trained to play StarCraft would not be able to play another similar game: say, WarCraft. Plus, deep learning does a poor job of handling data that deviates from its training examples.
The future of deep learning, neural networks and AI
It's clear that the work on deep learning and neural networks is far from complete. Various efforts are being made to improve deep learning algorithms. Deep Learning is a cutting-edge technique in artificial intelligence. It has become more and more popular in the past few years due to the abundance of data and the increase in processing power. This is the core technology behind many of the applications we use every day.
But will consciousness ever be born on the basis of this technology? Real artificial life? Some of the 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 claim is highly questionable. For real AI to emerge, we need to rethink the way we build AI systems. All that is now is only applied programs for a strictly limited range of tasks. As much as we would like to believe that the future has already arrived …
What do you think? Will humans create AI?