Teaching schoolchildren in the basics of artificial intelligence technologies in conditions of additional education

Abstract

The article is devoted to the description of methodological approaches to teaching schoolchildren the basics of artificial intelligence based on the technology of organizing project and educational and research activities to achieve meta-subject results of education, the formation of systemic representations and the experience of their application in real life. The experience of organizing courses within the framework of additional education for students of 10–11th grades of secondary school on the use of neural networks in solving practice-oriented problems is presented. The block structure of the content of the courses is given, which is based on the phased implementation of the idea of the future project. A detailed description of the content of each block is given with a demonstration of a presentation of complex theoretical material that is accessible to the understanding of schoolchildren. In particular, recommendations are given for the study of sequential and recurrent models of artificial neural networks, fully connected and convolutional layers, object detectors, as well as an algorithm for training neural networks by backpropagation of errors. Data acquisition tools for training artificial neural networks are described. An algorithm for writing a program for training a neural network to recognize handwritten digits of the MNIST set is characterized. Examples of projects for the use of neural networks, carried out by students together with the teacher as part of the study of the proposed courses, are given.The article is addressed to teachers of additional education, informatics teachers, teachers and students of universities — everyone who is interested in the use of artificial neural networks in solving practical problems and methods of teaching schoolchildren in project activity.