BLOOM

Introduction

BLOOM is a tool that facilitates the teaching and learning of a concept as complicated as Machine Learning. We want to provide both teachers and students with a more user-friendly way to delve into this concept. To achieve this, we use visual resources to simplify the understanding of abstract concepts on the subject.

BLOOM developers

• Brenda Oriolo - designer
• Zoe Perez Colman - front-end developer
• Agustina Schajris Garati - back-end developer
• Tomás "Blur" Spurio - full-stack developer

BLOOM tutors

• Ignacio Pardo
• Ignacio Vigilante
• Micaela Viegas

Problem

When it comes to learning to program, jumping straight into learning lines of code can be confusing for those who lack knowledge in the area, which is why Scratch is useful. This tool is an excellent introduction to programming logic because it allows understanding the fundamentals of programming visually, eliminating the complexity of the code.

This same situation is replicated in the area of Machine Learning. People find it difficult to understand how it works when they jump directly into lines of code, regardless of whether they know how to program or not.

We conducted some surveys with high school and university-level students and teachers, which received more than 50 responses, so we can state:

• 87.5% do not know about Machine Learning and prefer to learn with the help of an educational app.
• 83.3% think Scratch is a good tool to enter the world of programming.
• 54.5% would feel more comfortable with a graphical interface for learning Artificial Intelligence.

Objectives

• Facilitated teaching
• Friendly learning
• Simplified education

Solution

We propose creating a web application similar to Tinkercad and Scratch, but focused on Machine Learning, with a graphical editor that allows replicating the functionality of some of the most used libraries for building predictive models or neural networks, such as Scikit-Learn and TensorFlow.

Each user can create multiple projects, for which they can choose from several predetermined datasets (or their own) to base their model on. Once inside the project, the user can build their model with a kind of “blocks” that can be found in a catalog, functioning as lines of code (the block format may vary). In a tab on the side, the user will have a “dual visualization” of the “blocks” and the generated code to which these refer. Once the model is finished, the user can export the code in Jupyter notebook format to Google Colab or download it to run locally.

Additionally, each teacher can create classes for their students. There, they can write explanations and give assignments for students to complete. This section will have an auto-correction system where the teacher can limit the blocks students can use and define the correct answers.

Users

• High school and university students interested in the technology field, regardless of their previous programming knowledge.
• High school or university technology teachers who want to introduce and teach the concept of Machine Learning.