, respond like this" and then teach these rules

[robiulhasan]

start discussing the exact definition of AI, we'll get off track, so for now, it's enough to just roughly remember that there are three types of AI . Below, we've summarized the three types of AI based on
Professor Matsuo Yutaka of the University of Tokyo's book "Will Artificial australia accountant email lead Intelligence Surpass Humans? What Lies Beyond Deep Learning" (Kadokawa EPUB Selection, March 2015) .

Rule-based AI This is a type of AI where humans decide on rules (rules based on knowledge gained from human experience) for "if something comes like this
to a computer one by one to process it. For example, if you want to distinguish between images of cats from images of many different kinds of animals, you teach the machine detailed rules such as "if the ears are here, the eyes are here, and the mouth is here, then judge it to be a human" to make the judgment. This means teaching the machine the features to look out for and the rules for judging them.
Machine learning AI This type of AI allows humans to specify only the features to focus on in order to arrive at an answer, and the machine itself will discover the rules for using those features to make a judgment
. In the same cat image example as above, if you instruct the machine only to "focus on the features of the ears, eyes, and whiskers" and provide it with a large number of photos (each of which is tagged as cat or not), the AI ​​will automatically learn what rules to use to make a judgment using those features.
Deep learning AI This is a type of AI that can even extract the features to focus on in order to provide an answer