Development of Teaching-Learning Contents for AI Core Principles at the Elementary School Level: With a Focus on Convolutional Neural Network

Abstract

With the rapid development and multifarious application of artificial intelligence technology, elementary learners are demanded to learn about AI. At the same time, due to the pandemic, classes without face-to-face contact have been increasing. This study aims to develop teaching-learning contents instructing core principles of artificial intelligence, respectively in face-to-face and non-face-to-face learning environments. Proposed teaching-learning contents mainly target elementary learner and provide simplified experience for convolutional neural network(CNN) algorithm. Learners accomplish convolution operation with simple numbers then extract feature map. Students experience CNN process with 5x5 virtual input image consists of numbers. 2x2 filter is given for convolution operation, and students fill 4x4 feature map with the results. These same teaching-learning activities can be also offered with online interactive worksheets, using ‘LIVEWORKSHEETS’ website. By transforming printed worksheets into online ones, learners in distance learning classes can accomplish identical coursework clicking and typing then score themselves. Teachers and learners can choose appropriate teaching-learning way or mix both of them flexibly based on their situation. Subsequent studies are expected to revise the contents after sufficient pilot classes with diverse learner groups and devise additional physical teaching-learning activity suitable for elementary learners. This study proposes the importance of continuous research on artificial intelligence education for young learners and versatile teaching-learning contents for various learning environment.