In computer science, algorithm means a series of well-defined instructions given to a machine to perform certain actions that allow to solve a certain problem. Normally, the algorithm is designed and developed to solve a specific problem and does not change over time.
In an AI system, however, the algorithm learns and modifies itself based on the data it receives. The more data is analyzed by the AI, the better the final result will be. So in an AI system the machine is subjected to the results it has to obtain and it will be the machine itself, through a training path, to find the best way to achieve that result.
Artificial Intelligence is a set of technologies that includes Machine Learning. In turn, machine learning includes Artificial Neural Networks (Artificial Neural Networks). A subset of Artificial Neural Networks are Deep Learning networks.
Example of Artificial Intelligence:
Machine Learning is a subset of artificial intelligence that deals with creating systems that learn from the data they use. In Machine Learning the information or, better, the models are acquired directly from the data, without the use of predetermined equations.
In a problem solved in the traditional way it is the programmer who describes in detail the various steps necessary for the solution.
With Machine Learning, through appropriate mathematical algorithms exposed to a given set of data in a defined initial phase of training (or learning) and passing through a second phase of testing (or evaluation of results) for the optimization of the parameters, the function capable of identifying within a different set of data the most likely solution is obtained independently, possibly indicating a degree of confidence in the estimate. In practice, the system obtains the function itself. This system, composed of trained algorithm, data and parameters, is called a model.
The automatic learning is to the base of the systems of artificial intelligence in how much it enables the machines to learn without being explicitly and preventively programmed.
Machine Learning models are very effective in identifying correlations in huge data sets, taking into account a number of variables that would be unthinkable for a human being.
Machine Learning has two important advantages over other techniques:
Examples of Machine Learning:
The Artificial Neural Network consists of a mathematical model of calculation that simulates the behavior of biological neural networks of the human brain. The latter, unlike machines, is able to perform several independent operations at the same time and is therefore a model to aspire to.
An Artificial Neural Network can be formed both by software programs and hardware and is based on a network of interconnections between three types of nodes that exchange information:
Examples of Artificial Neural Network are for example:
A neural network consisting of two or more series of intermediate nodes allows a very powerful form of processing, called deep learning.
Deep Learning is the ability of an algorithm to learn from input data through successive layers of processing.
Deep learning uses multiple levels of filters to know the significant characteristics of the data.
Example of Deep Learning: