My mother never went to school. She cannot read or write. And every time she sees me in front of my computer, she is both amazed and stunned by this small screen which is used to write and surf the internet and follow a whole number of events on the web. What is this thing that keeps you busy all the time? daytime? She asks me all the time when I sit down in front of my computer. I tell her it’s a machine that functions like a human brain. She laughed with a look saying to me you are crazy or are you kidding me?
One day she went with my son to a children’s amusement park and she met a robot at reception and speaks like a human, indicating directions to take. She learns that the robot, in fact, works thanks to artificial intelligence. And when she gets home, she rushes towards me and asks me: “you who are always glued to your computer can you explain to me what it is artificial intelligence?”
I admit that explaining artificial intelligence to a layman is not an easy thing. It’s even more complicated when the person has no concept of digital technology. My mother cannot read or write. She doesn’t have never watched science fiction films. I didn’t know where to start. GMAIL for example which gives response choices when you access an email, this is thanks to artificial intelligence. Netflix to give us access to films uses artificial intelligence. The translations of all publication on Facebook is also artificial intelligence. But my mother doesn’t know any of this. So how do I explain it to her? Besides, she doesn’t stop. She asks her question every ten minutes.
Of all the simplest definitions of artificial intelligence, there is that of the French Yann LeCun who is a world authority in the field of artificial intelligence. It is simpler especially because of its clarity. I concentrate and translate it into my language, Kirundi, based on Yann’s definition. LeCun: “artificial intelligence is a set of techniques that allow machines to accomplish tasks and solve problems that are normally reserved for humans and some animals.” In other words, I said, to my mother it is a machine that can recognize your face or a any person, play cards with you, drive a vehicle, translate one language to another by example from Kirundi to Swahili, talking to people by telephone… These are works which are today made by machines even if in some cases they do not yet make them all by hand. perfection.
My mother, very surprised, asked a common sense question. “What do you mean, a machine can carry out tasks humans?” So it’s about the functioning of artificial intelligence that it poses here. Hence the common sense nature of his question. The adventure of explaining how intelligence works artificial poses a risk. It is that of losing your interlocutor. For my mother’s case, the risk of losing her is even greater especially since she has no concept of artificial intelligence, she has it Heard it for the first time a few days ago. So everything must be carefully simplified.
My mother must keep in mind first that knowledge is possessed by humans. So for For machines to behave intelligently, they must also possess knowledge. He must therefore be instilled in them. There are multiple ways to do this. First there is the classic route which is that of manually programming a machine. In this case the machine acts within rules defined advance. So it cannot respond to all situations, it remains a previously programmed object.
This is why to improve the capacity of the machine, we have evolved in artificial intelligence towards this called machine learning. Basically, it is about preparing the machines to learn by herself. For example, when we teach a baby to recognize a cat, we present him with books in which there are several photos of cats. The baby will quickly assimilate what makes a cat. There machine also achieves this, but on condition of presenting it to it, of storing thousands and thousands of images. This is why it is said that big data is the fuel of artificial intelligence.
Machine learning assumes that the machine has small operators doing calculations and having a learning capacity by modifying their connections between them. These small operators are “multilayer artificial neural networks,” because they are inspired by the human brain. And when they come into action, it’s called deep learning.
Today, deep learning is at the center of promising work in Artificial Intelligence. I make mom understand that research in AI continues and evolves very quickly in view improve machine capacity and efficiency. A new approach, in the family of machine learning, and which gives great hope, is called reinforcement learning. In this case, the machine learns through numerous trials and errors, which are “rewarded” when it produces the decision sought.
These explanations thus provided to my mother, she told me she understood. She told me in summary: “a machine receives data from humans and when asked to reproduce data that it possesses, it does so. done and sometimes with the speed that can beat human capacity.”
