Artificial Intelligence sneak peek
Intelligence Meaning and introduction
Artificial intelligence is one of the most famous computer related fields due to science fiction. Also is one of my favorites.
Intelligence is a wide term that can be understood in many ways, the self definition of intelligence is not even unique and can be described in many kinds of intelligence, which, itself, exposes how blurry the term is. Howard Gardner -Cambridge university- described the multiple intelligence theory, as many ways of seeing intelligence. Which means you can be many kinds of intelligent:
Linguistic (as in kids), the capability of learning and manipulating a language.
Mathematical and logic, the capability of manipulating numbers and logic problems and relations, in games or experiments, too.
Kinetic and corporal, the ease of perceiving your surrounding and control of the body. Like professional dancers, athletes and such.
Musical, actually related to mathematical, the ease of manipulating the sounds and talent to listen propertly.
Interpersonal, the social and communication ease. Also includes the empathy and self-recognition.
Naturalist, the ease to feel close to nature and understand the environment.
So, knowing this, what is actually the “artificial intelligence” term referring to? IN the general term, AI is just meant to solve problems quickly and efficiently, using complex algorithms. But even tho, there are some huge nerds on computer science who aim of something more than that.
Let’s first talk about the 4 main blocks that separate AI projects. We will talk about:
- Systems that ACTS like humans
- Systems that THINK like humans
- Systems that ACT rationality
- Systems that THINK rationality
Systems that thinks like humans are related to cognitive science. They are based in psychological studies and experiments that try to understand how human mind relations work, and then tries to replicate it with computing. In this case we will see the neurological model of algorithm, in which data is treated like information processed in our mind, with -mathematically defined- neurons.
Systems that thinks rationality are more related to logic, in which the laws of math and logic are pursued strictly. This one is very difficult to apply as practical experiments are way different than ideally mathematical environments.
Systems that act like humans are my favorite personal. These kind, are intended to fool a human on making them think they are able to think and even be humans, but not necessarily are. Science Fiction has used this kind of AI models many times, such in Ex Machina. Turing, one of the computing science fathers, developed this idea in his XX s. studies, and defined the Turing Test in his article “Computing machinery and intelligence”.
This test implied a human tester in a computer or a similar machine, chatting to someone without seeing that person. Based on the conversation, the human tester had to decide if his/her pal in the other side of the conversation is an AI or a real person. This test is -still- currently taken in AI competitions annually.
System that acts rationality are called agents. They are told to obtain a goal and they are given tools and experience to reach that goal. Most of the AI main projects taken nowadays are related to this kind. They are required to learn from the environment, take decisions, and perceive.
But, is actually AI possible?
This is a complex discussion taken by computer scientists, engineers, psychologist and philosophers, and there are also a wide range of answers. Turing again, defined a list of possible arguments against AI being a possible thing, and he personally believed it could be possible in a certain point of the future. Here I wrote about how some of those arguments are already invalid.
There’s something called “The Chinese room”, an argument used to disqualify AI as an impossible concept.
There’s a room closed with an input hole and and output hole. Someone puts a Chinese text in the input, and after a while the translation appears in the output. In general it seems to know Chinese, but does it really? Apparently, no, as the artificial subject in the room will be equipped with lots of manuals and dictionaries to assist it, but will not understand the meaning itself.
This is almost already touching philosophy, which is not technically separated from AI, as the bases of these studies are mathematical AND philosophical. But regardless where the idea comes from, the AI itself is an useful tool in many fields such as economy, neuroscience, psychology, computing, linguistic, medicine, etc.