Envision Can Be Possible By Machine Learning Method
Machine learning is a big term which contains all the necessary points when you are trying to teach a machinery to improve on its own. One of the example of machine learning is Neural Network, but they are not the only way to train a computer.
For example:
one of the alternative methods of machine learning is called training with reinforcement. In this method, the computer performs the task and then evaluates its result. If, for example, a computer wins chess, then it assigns a winning value to the series of moves that it uses during the game. Having played millions of games, the system can determine what steps are likely to lead to victory, based on the results of previous games.
Evolution of ideas about the ways of development of robotics, its goals and tasks is very similar to what is observed with an area such as artificial intelligence. The declared general principles and, as it seemed, the understanding of how to achieve some global goal of research was replaced by a narrow specialization, many private, often unrelated sub-goals and directions.
Technologies of machine learning have always been closely related to robotics. The creation of robots – machines capable of acting as a person is the common main goal of these sciences. After the impressive success achieved in the second half of the twentieth century with the successful introduction of industrial robots into the process of automated production, it is now possible to talk about the transfer of the center of scientific research into the creation of autonomous robots. Here it is necessary to mention space robots to study the surface of the celestial bodies of the solar system, robots for underwater research. In the fight against terrorism, there was an urgent need for robots designed to clear suspicious objects in crowded places. You need “smart” robots that can extinguish fires without the operator’s help,
To begin with, we will determine how the robot differs from the automatic control system or the machine. An automaton is a device equipped with a set of receptors and effectors that performs information transformation with the aim of providing the external environment with the required effects. It is assumed that the connection of the machine with the sources of information is rigid.
The advantage of machine learning is that the predictive function does not need to be specified explicitly, but it is enough to determine only its general parameterized form, automatically adjusting the parameters for the training sample – the final set of pairs <input value – output value>. For example, in the task of automatic matching words of the text of their parts of speech, you can explicitly set the rules, and you can apply machine learning techniques to text with already marked parts of speech to determine the best rules automatically, and then apply them to new texts.
