Abstract:

Every student deserves an individualised learning experience that is tailored to their needs in order to help them reach their full potential. The current educational system generalises students, making them vulnerable to falling behind. Despite efforts to provide more individualised attention, teachers simply have too much on their plates. This project aims to supercharge traditional education with advanced AI techniques to address some of the present issues.

The Digital Autonomous Virtual Educator project (D.A.V.E.), will aim to provide a personalised digital tutor to each student. D.A.V.E. will use AI to analyse student academic data, adapt academic material according to individual needs, and provide clear and detailed explanations in real time.

D.A.V.E. which will take the form of an AI chatbot will be centred around three core pillars: Adaptability, Accuracy, and Explain-ability. The main module in D.A.V.E. will be a Large Language Model (LLM), additionally, various AI modules will further be integrated in order to bolster the capabilities of the main module allowing for a seamless and efficient solution for both students and educators.

D.A.V.E. will create and update unique profiles for each student, allowing the system to adapt outputs according to the individual student using it. Academic data will be used to form the knowledge base of the system ensuring that all responses provided will be based on verified material. Furthermore, explainable AI techniques will be implemented in order to deconstruct the system’s chain of thought allowing for a more transparent solution.

The benefits of this project include providing students with a personalised learning experience that caters to their specific needs, improving student engagement.

Keywords:

Education, Transformers, Large Language Models, Natural Language Processing.