In recent months, everywhere there is talk of artificial intelligence. Sometimes with knowledge of the facts, sometimes by slipping it a little haphazardly into conversations. It's definitely not a new concept, but it's not an easy topic of discussion, for several reasons. It is difficult to understand the basics, it is difficult to understand the concrete areas of use, it is difficult to enter into the ethical implications.
As a technician, however, I feel the need to understand how this branch of information technology can enter into my work routine, so I started training me to be able to understand a little better the limits and potential.
So now I want to start sharing with you what I have learned and I am learning, so as to help those wishing to learn more to get a better idea of the current evolutionary state.
The first question that I had, was what means "machine learning".
What is Machine Learning?
Machine learning (ML for friends) is a discipline of artificial intelligence (AI for friends).
What is the goal of ML?
The goal of ML is the creation of systems capable of learning independently starting from data analysis.
How does it do it?
ML is based on algorithms that simulate the learning process of human beings, going to refine results and performances from time to time.
ML uses mathematical algorithms applied to large amounts of data with the aim of producing models applicable to those same data that can be replicated on the data itself in order to be able to produce predictions.
The need for a similar technology is given by the fact that a problem cannot always be solved by programmed software, either because it is too complex or because it evolves too quickly.
For example for facial recognition, or language translation, but also for medical diagnoses.
To solve these complex problems, a large amount of data is fed to the system, and the algorithms are left to work on it by recognizing patterns, rules, relationships between the data themselves, so as to obtain a model which can then also be applicable in future.
Everything therefore starts from a training phase, precisely because it is not the programmer who defines the algorithm, but it is the system itself which autonomously analyzes the data provided and discovers how to use them to solve specific problems.
Top comments (0)