AI, Machine Learning & Robotics – Are They Similar?
Today, everyone talks about the advantages and prospects of machine learning and robotics and the possibilities of creating artificial intelligence in the near future. And we in ServReality are not an exception, because we are actively applying the principles of these new high technologies in our projects. So let’s talk a bit about the specifics of the terms, and what they mean, because the concepts of artificial intelligence and robotics and machine learning are far from identical, and for a person unknowledgeable of the intricacies of these concepts, it is very difficult to understand them.
AI and ML – What’s in Common?
The term artificial intelligence is authored by John McCarthy, who by this concept signified the ability of computers to perform operations up to those of human intelligence. That is planning skills and the ability to distinguish and categorize objects and sounds. Machine learning was also included in the list of skills inherent, according to McCarthy, artificial intelligence. And it was scary to imagine, 60 years ago!
The second concept – machine learning – the world received from the mouth of Arthur Samuel all the same 60 years ago. He defined it as the ability of a computer to learn new things without human participation in the studying process. If usually in order to take into account all the subtleties of a specific task, you need hundreds of lines of code that prescribe each possible option, while using machine learning, the programmer’s task is to create special algorithms that allow the computer to recognize the script and act according to it.
At the same time, there are many variants of machine learning techniques in artificial intelligence, such as a particularly popular method of deep learning. It consists of creating artificial neurons similar to human neurons that form a multilayered neural network. The layers are responsible for specific algorithms of the same recognition and categorization of objects. An analogic scheme is used to teach computerized systems to recognize specific things — for example, horses. The training scheme is connected with interaction with horse images, analysis of decisions made by a man when he determines from a variety of images consisting horse – that’s the primitive example of how artificial intelligence problems and techniques are solved.
What About Robots, Then?
At the same time, artificial intelligence and ML are inextricably linked with the industry of robotics and the creation of the Internet of things. “Smart” systems are gradually entering everyday life at various sides of it, from creating nano-robots for surgery and diagnostics in medicine to introducing the functions of electronic assistants to the simplest things, for example, in home appliances (yes, smart home systems are exactly the example).
Large, small and medium-sized businesses are beginning to be interested in this, a large number of start-ups appear, and this enters the minds of ordinary people. News from the series: “AI composed a music album in the style of Bach” will not surprise anyone today. They are no longer perceived as journalistic fakes, fantasies and so on. Although, of course, in fact, there is some level of journalistic notions in these headlines.
Artificial intelligence and machine learning in robotics are becoming part of our reality, which is rapidly changing, and, in connection with this, it’s worth to talk about one fundamental error that, very often, unfortunately, has to face, especially recently. It lies in the fact that most people consider artificial intelligence and machine learning, robotics something, if not very close and similar, then, in general, a single entity. That is one thing. Robots and artificial intelligence are one and the same.
In fact, this is a very large and harmful misconception. And one does not necessarily imply the other. A simple example: robots without artificial intelligence is our reality. Any programmable devices such as drones (unmanned aerial vehicles) are robots, but decision making and object recognition are not within their competence; all commands come only from the one who controls the robot. And the opposite situation: artificial intelligence systems may not be enclosed in the body of the robot, moreover, they can exist only in the global network as, for example, the systems of the same web search.
However, the combo of AI&robotics is already a reality. It is about creating intelligent robots, a dynamic and rapidly growing industry of robotics. And it is precisely such devices capable of responding to the situation and making decisions within their functions without applying of total human control is the future. And it is very close!