In this article I would like to tell you something that I am particularly passionate about and it seems to me one of the foundations of current and future technology. I put the focus on the Machine Learning and the Artificial Intelligence (AI). The fundamental pillars of other trends such as BigData or DataScience.
A topic that I'm passionate about and I was not going to tell the students in the classroom. I found the problem in how to relate this seemingly complicated idea with the didactics or methodology of a class.
To do this I started to collect and adapt different information and points of view on the subject, until I realized that many of these points led me to a single name, the mathematician Alan Turing, considered one of the fathers of computer science and precursor of modern computing. Alan Turing provided an influential formalization of the concepts of algorithm and computation, two of the fundamental pillars of Automatic learning and Artificial Intelligence.
According to an article that I read a long time ago and from which I leave the reference (http://www.eldiario.es/turing/Test-Turing-inteligencia-maquinas_0_225377744.html), Alan Turing He was a visionary and always believed that machines could evolve by achieving artificial intelligence. To expose his thesis of the thinking machines he devised the imitation game, what we know today as the Turing test
"If a machine behaves in all aspects as intelligent, then it must be smart"
When I read those words "imitation game" and "Turing test" I said ... this I have to take it somehow to the classroom. If I manage to realize some form of gamification or didactic with these concepts probably the students can understand in a simple way how a machine is capable of learning by itself.
To delve further into this idea we must know that it is exactly "the Turing Test" that according to its inventor, is born as a method to determine if a machine can think. Its development is based on the game of imitation.
"The original idea is to have three people, an interrogator, a man and a woman. The interrogator is separated from the other two, and can only communicate with them by writing in a language that everyone understands. The objective of the interrogator is to discover who is the woman and who is the man, while the other two is to convince the interrogator that they are the woman. "
This original idea encouraged me much more to take it to the classroom, I realized that it would not be difficult not only to do it in high school to understand the concept, but also to be able to try it in one of the elementary courses. Now I just had to adapt it and look for the methodological procedure to follow for this idea.
Turing test to study a topic?
Based on the original idea of three 3 people, one question, one man and one woman, and in order to guess by the questioner who the woman is, I designed my own "test" for the classroom.
Let's make a simple example ... «Let's Gamify»
Imagine that we are studying living beings in science class. Explaining this topic in the classroom will give a base of knowledge to our students, but logically we will never be able to study all the living beings in detail.
We are going to substitute the concept of man and woman for animals ... each one of our students must study an animal in particular, as a classroom task or at home. Once they collect and learn information about that animal, we will choose 3 students to do our test.
IT IS IMPORTANT THAT THEY DO NOT KNOW WHICH ANIMALS ARE EACH ONE.
- Alumno1: Giraffe
- Alumno2: Elephant
- Alumno3: Shark
We are going to study the Elephant.
Therefore we can paint the scenario as:
- Alumno1: Interrogator
- Alumno2: Elephant
- Alumno3: The Computer that learns
We decide that the interrogator must guess which of the two is the "Elephant" and who the "computer"
Logically following the idea of Turing the Alumno2 must answer correctly to the questions to convince the questioner that he is the "elephant" while the student 3 must answer the answers in the most convincing way with his small knowledge base about that animal, of which he has not studied.
The methodology of the exercise is that the interrogator should use specific questions about the elephant to take advantage of the knowledge that the student has previously studied that animal and thus be able to recognize him.
It would be a waste of time to ask such simple questions as Does he have a trunk? How many legs does he have?... they are questions that either of the two students could answer "without batting an eye".
To make this game you can raise it in different ways according to the level of the classroom, the number of students or the time spent.
I have done 2 different ways that I expose here, but as I say this Turing Test is "very adaptable" to our classroom.
- We perform the test live and both questions are heard their own answers, so you must alternate the order of answers or that the questioner chooses who wants to answer first.
- We perform the test in silence, the questioner asks the questions and the 2 and 3 students answer individually on paper. In my opinion I think that this is the best option, in addition to this scenario we can use resources like TodaysMeet to create a closed chat between the 3 students.
- Important to establish test time, it can be for real time or by number of questions.
Starting from the fact that the natural objective of the test is to guess who is really the "elephant" and who is the "computer", we also obtain that the student who studied the chosen animal, review everything learned through the questions of the interrogator.
That the interrogator learn the animal's data through his own questions, important to make the interrogator understand that asking questions without any logic can lead him to a dead end.
Finally we have our student3, in the role of "computer". He starts from a knowledge base that will improve and improve thanks to the responses of the student 2, so that in this way it will be better and better in his answers and knowledge about the animal to be able to confuse the interrogator.
A collaborative learning among students from which you can learn and review at different levels, asking the same questions, using the same time and the same procedure.
It would be great that at the end of the exercise we explained to the students as the Student3, our supposed "computer" has been able to learn the details of an elephant by itself, only using the imitation and logic of the answers and questions of the other two students .
Our little Machine Learning in the classroom !!!
To finish I pose the following question:
Will there be a moment that, as Turing said, we will not be able to distinguish whether we are talking to a human or a machine?