Personalized Learning in Automation: A 3D AI-Based Approach

In today's technology-driven world, the need for interdisciplinary skills is increasing. This has become challenging in tertiary education to provide students with applicable knowledge of various fields. Anderson's Adaptive Control of Thought (ACT) theory suggests that universities have traditionally focused on imparting declarative knowledge, which involves memorization of facts and concepts. However, imparting the ability to apply such knowledge on individual students and create procedural knowledge is the challenge. This includes teachers dealing with a diverse range of student abilities, particularly at university-level where they teach the same course content to students with different levels of prior knowledge and, given the structure of modern education systems, the resources required to monitor and provide feedback for a number of decisions and attempts independently performed by the students. Intelligent Tutoring Systems (ITS) have proven to be effective in addressing the aforementioned challenges by creating personalized learning environments that provide instant feedback, adapt to individual student needs, and promote the development of procedural knowledge. In the field of automation education at the university level, we are creating a 3D artificial intelligence (AI)-based ITS software named KIAAA (An AI Assistant for teaching in the field of automation), specifically designed to teach computer programming to students. KIAAA aims to assist students in transitioning from their abilities to procedural aptitude by providing personalized learning scenarios that allow them to apply their knowledge and receive immediate feedback. Our approach is based mainly on the pedagogical model of ITS, which focuses on creating a supportive and inclusive learning environment that promotes success for all students, regardless of their initial level of knowledge. One of the key aspects of our approach is the utilization of personalized learning. We propose a scheme that, subsequent to evaluate student's initial levels of procedural knowledge, creates 3D learning environments tailored to each individual student. By analyzing the solutions proposed by the students, we select the difficulty level of subsequent tasks. This approach takes into consideration student's discrete competence throughout the learning process, enabling them to progress on their prior knowledge. Additionally, the software provides customized feedback to each student on their performance, helping students identify areas that require improvement. Concepts for and implementations of ITS for a variety of fields, including introductory programming classes, have evolved for a long time. Our main contribution lies in presenting an end to end solution for ITS focused on teaching programming for automation students with realistically 3D simulated factory environments. While we strongly believe to have created a pedagogically sound, integrated intelligent teaching system for assisting programming classes in tertiary automation education, a robust user study for methodically evaluating our concept and implementation is still to be performed. Thus, we limit ourselves to presenting the underlying didactic concepts of KIAAA as a work in progress paper with a comprehensive evaluation to follow at a later date