COMPUTATIONAL THINKING IN HIGH SCHOOL: A TEACHING SEQUENCE ON ARTIFICIAL INTELLIGENCE AND ARTIFICIAL NEURAL NETWORKS

Abstract

This work aims to integrate mathematical modeling (MM) and artificial neural networks (ANN) into secondary education (SE) with the goal of promoting computational thinking (PCO). The teaching proposal was implemented in a third-grade class at a full-time public school, through a teaching sequence involving eleven lessons. In these lessons, students explored ANN concepts linked to mathematical content and aligned with the guidelines of the National Common Curricular Base (BNCC). A mixed methodology was used, combining qualitative and quantitative approaches. Initial questionnaires assessed students' prior knowledge and interest, while hands-on activities and interactive discussions, such as comparisons between biological and artificial neurons and the use of Google Colab programming, enriched the learning experience. The Python programming language facilitated the application of ANN concepts, such as activation functions, weight adjustment, and pattern recognition. The development of the PCO was observed, particularly in decomposition and abstraction skills, as students followed the application of simple algorithms to structure mathematical and computational problems. Although initially apprehensive about using functions in ANN composition, students eventually understood the logical simplicity of these concepts, which increased their engagement in practical activities. The conclusion is that the teaching sequence was effective in introducing students to the field of ANNs, fostering connections between mathematical and technological concepts and sparking interest in the world of programming and its applications in solving contextualized problems. With appropriate adjustments to the timing of each stage and a greater focus on practical activities, this approach has great potential to introduce students to new technologies and strengthen the development of the PCO.