p mFirst of all we clarify what AI (Artificial Intelligence) means. It’s the things, processes, programs that lead to create a “system” that reasons like a real human brain. As you can see there is no precise, univocal and equal for all definition. This is because the sector is very broad and there are conflicting opinions among them. however, basically it turns around is what we have described in our words.
Now things become more complicated, let’s see the different types of AI:
La Narrow inteligence is a variant which has generally a limited pre-defined range of functions, it choose what to do itself, it is not no self-awareness. In fact some people call this variant, Weak AI. This for the fact that there must be always somebody that say what to do.
Some example of Narrow AI:
- Google Assistant
The general AI is the “real” artificial inteligen according to many people. It would and will be like the human’s brain. It will be able to solve any questions, problems that the human brain can too. So, the differences beetwen General AI and Narrow Ai that the first have consciousness respect of the second. Moreover the Narrow AI (or Weak) do not have all the human cognitive abilities.
This is precisely a test to check if a machine is inteligence. Formulated by Alan Turing and introduced in 1950.
The Turing test is developed in this way (not to mention that many variants were added):
There is a man A, a woman B and a third person C.
The third one is separated by A and B and its task, through a series of questions, is to understand who is the woman and who is the man. Meanwhile A has the task of deceiving C, to make him believe that he is the woman and B the man. Instead B must help C to arrive at the right conclusion.
Subsequently, according to Turing, the machine would take the place of A and if the total difference of the times that C guess who is the man and the woman when there is the car and no, are similar then the machine is considered inteligente.
The Machine Learning is the set of system used to improve an algorithm independently and constantly over time. In the past 20th century (last decades) this method
in the past This was in the last decades of the 20th century.
Deep Learning makes it possible to find a pattern, a regularity within a series of data placed “randomly”.
This happens thanks to a set of algorithms and statistical techniques.
This allows for example to be able to detect obstacles along a path or visual recognition.
The neural network is a mathematical model composed of artificial “neurons” that simplify a biological neural network (neuron circuit).
The possible applications are multiple:
– Early medical diagnosis
– Meteorological forecasts
– Autonomous driving
– recognition of handwritten scrawls
–Optimization of some industrial processes (eg quality control)
–Social media (better management of posts)
Most likely in the future we will live more and more next to the AI. They will help us in many things, many are afraid that they may overwhelm us and they are right. We need to find the right balance between giving the AI too much power and banning it altogether. Because, in a few years, can you imagine yourself living in a city where cars will be self-driving and fly drones around the city to adjust the roads?