Note: this is an automatic translation of this post performed by https://www.deepl.com
In December 2018, Jesús Moreno León, then head of the Experimentation Department at the INTEF Classroom, proposed me to develop a teaching resource on Artificial Intelligence based on a fantastic tool I had discovered: Machine Learning for Kids (ML4K), by Dave Lane, an IBM worker who has developed this tool within the IBM volunteer program. The experience was fascinating and I was really excited to discover that something as complex as Machine Learning could be effectively explained in school.
So I decided to continue exploring this topic on my own and build the prototype of an application similar to ML4K but with an important condition: that, unlike ML4K, it was not necessary to create any kind of account to have a complete experience with the application. It may seem silly, but it is something that makes it quite difficult to use the tool in a school environment.
The first problem I had in avoiding the need to create an account was that the Machine Learning algorithms had to be executed locally in the user’s own web browser, since using one of the artificial intelligence platforms offered by large companies like IBM, or more modest ones like ClarifAI, necessarily implies the creation of an account.
Then I started to investigate the possibility of running “heavy” Machine Learning algorithms in the web browser… and the result was encouraging; I discovered the simple Brain.js library, with which I made the first prototype for text recognition and the very powerful, professional and not so simple Tensorflow.js library, with which I implemented image recognition. The conclusion was clear: it was possible to execute Machine Learning algorithms without the need to depend on third party services, since the execution of such algorithms in the web browser is possible, even more so considering that the tool to be developed has a pedagogical purpose and we are not going to use massive amounts of data as can happen in specific Machine Learning applications.
Shortly afterwards, at the hand of Jesús Moreno, I met Gregorio Robles and Marcos Román, two researchers who, together with the first, have been making interesting contributions to the world of Computing Education Research for several years. They saw, on the one hand, the possibility of building a useful tool to work on computational thinking, and on the other hand, the seed for a thesis in this field. So they encouraged me to start this intense but rewarding task. I did not hesitate for a moment to take the opportunity to launch myself again into the world of research. And I say again because it has been more than 20 years since I left it due to circumstances that are not relevant now.
A little over a year later, we have a first operational version of LearningML with which you can build Machine Learning models and program Scratch applications capable of recognizing text and images. Now it’s time to show it to teachers, students and researchers to test and see if what we are doing is a good resource to promote computer thinking and serves to learn content on artificial intelligence in a practical way. And with the desire to do so, we will continue working on improving LearningML.