What is Artificial Intelligence? We can simply define it as that field of science that aims to develop intelligent machines. They are studied to be able to reason, learn, communicate, getting closer and closer to our being and facilitating different tasks in everyday life.
So the question is artificial intelligence? many already see replacements of humans, job losses, overpowering machines on humans. I'm sorry, I don't want to see it like that. I want to think hard that, instead, it will be the best way to free our time and dedicate it to more important activities including thinking more, reducing repetitive and alienating workloads, giving us support to improve the world in which we live.
To overcome any fear, in my opinion, it is necessary to have knowledge of it and we do it by studying the following topics:
- when born
- how does it work
- machine learning
- deep learning
- application examples
When was artificial intelligence born?
much older than it seems in 1943, in fact, Warren McCulloh and Walter Pitt proposed the first artificial neuron. The expression Artificial Intelligence, however, was coined in 1956, it was John McCarty, an American mathematician who launched the term Artificial Intelligence and from that moment the acronym AI became increasingly important.
In recent years, worldwide, interest in the subject, according to Google Trends, the tool that measures research trends by comparing normalized data, has tripled.
In Italy, on the other hand, we are still slightly less interested in enlargement, and if you're wondering what that peak of research is at the expense of the song by Michel Tel Ai if you pego! summer 2011 smash hit. S, I'm smiling.
Returning to the piece, certainly there is an increase in interest in the subject, there are more and more publications on the subject and every day we read news on how artificial intelligence is the protagonist with its multiple applications.
Artificial intelligence: how it works
Artificial intelligence is born to reproduce the cornerstones of human behavior and to be autonomous in dealing with problems and making decisions.
Therefore, the programs are developed for:
- have a non-sterile knowledge of the world;
- knowing how to solve a problem with cognitive functions;
- take advantage of the logic to reach a solution;
- learn from mistakes and not reproduce them later.
The two major categories that I define artificial intelligence and meet these needs are:
- weak AI (weak AI): programs that approach only slightly our intellectual capacity but that help us solve problems and make decisions independently. An immediately understandable example is all the text translation software.
- strong AI (strong AI): systems that are able to emulate the human mind's ability to develop its own. Let's say that here we can give space to the various memories of the computer of Odyssey in Space, or the very sweet and in love Caterina of Alberto Sordi, for now we surely have systems capable of simulating human thought, but not of having one's own thought and consciousness .
There are so many theories about developments on Artificial Intelligence and a potential development of superiority over humans. For now we are using this new technology to simplify our lives and create new opportunities.
Accenture, at the 2018 World Economic Forum, reports a fantastic figure in its report: thanks to Artificial Intelligence, companies' revenues could grow by 38% by 2020, employment increase by 10% and 61% of managers interviewed believe that, in the next 3 years, the number of operators that will use it will grow. Of course, the work will have to be redefined, but I am counting on our intelligence and adaptation to change.
Artificial Intelligence: whatClick To Tweet
Machine Learning: learning method of Artificial Intelligence
They are the managers of Artificial Intelligence and they make it so that it learns, corrects errors and carries out a task independently.
a real method of learning based on algorithms:
- with teaching supervision to make the AI understand how to behave and without, to develop the analysis of the results and generate a new behavior;
- reinforcement by merit, rewarding the AI for having correctly carried out the assigned task. A bit like the gift we give our children when they are promoted.
Specifying, when the machine is in didactic supervision before doing anything it is trained through examples that allow it to understand the inputs and outputs and the results that are obtained.
If it is decided not to use this training, artificial intelligence will not know the outputs and will have to make mistakes and learn from their mistakes to know the inputs perfectly and get the result. This is the case of neural networks, interconnected mathematical models so as to be able to manage different information and calculations and which are the basis of the Deep Learning of which I speak to you later.
Deep Learning and Artificial Intelligence
You know the facial recognition of your smartphone images to the point of cataloging your photos according to the face present in them? Deep Learning.
a sub-category of Machine Learning, the part of models for lin-depth learning. Our brain is an intense and fast exchange of information through the neurons that allow us to learn from our experiences. This practically happens to AI of this type: a set of statistical calculation algorithms that create complex neural networks.
Do you remember Michael Knight's Kit? Deep Learning. We are not yet at that level, but cars with automatic driving are getting better and better results and will be able to run unmanned, if they are able to rework complex actions with the same speed and to predict the behavior of other individuals driving.
Application examples of Artificial Intelligence
Now that we know so the Artificial Intelligence is also the case to intrigue you with some examples of what is already at work and that maybe you are already close but you don't realize it.
Ok Google, turn on the lights in the room you know? Your voice assistant at home is the result of Artificial Intelligence. He must understand your language, even when you are fast or too tired or you talk by tying your shoes, and he must understand your habits to know how to respond like an old friend, after all, he governs your home environment.
chatbot: intelligent systems that answer your questions in place of the classic telephone operator, and that are used for now standard questions that would only steal time from the operators who, instead, have to deal with customer care problems more complex than the classic FAQs.
Patient analysis systems in the hospital: through the collection and processing of real tons of data, I am able to return cures for very precise diagnosed diseases.
Robot: I had the opportunity to meet a very nice one named Paolo Pepper, along with Salvatore Russo, with whom we exchanged funny jokes (agreed with his operator) and it was a very special moment, between fear and curiosity. Robots of this type are used to give information to the public, they answer simple questions such as where I find the hotel bar or in which lane I find the Amarone della Valpollicella.
Journalists and copywriters: there are already online news sites that write news for people like me. I just hope you don't get to the point of experiencing aseptic news without any kind of added value, like a comment, a critique or a beautiful human judgment.
. (TagsToTranslate) Intelligence (t) artificial (t) cos (t) learning