Skip to content

Prompt Engineering for Effective Personalized Learning: Unlocking the Potential of Large Language Models for Science and Mathematics Education

Abstract:

Large language models (LLMs) like ChatGPT demonstrate a remarkable ability to generate personalized and contextualized content by relating disparate ideas. This project proposes harnessing LLMs to make learning in science and mathematics more captivating and comprehensible. We will develop techniques for students to interact with LLMs to create customized conceptual descriptions and examples connecting course concepts to their personal interests and experiences. Linking abstract ideas to students’ own hobbies, passions, and backgrounds can increase motivation and retention. To complement enhanced engagement, we will also design activities focused on critical thinking and problem-solving skills in the classroom and at home. Students will learn how to practice and review for assessments by testing their own problem-solving skills through prompt engineering methods like chain of thought prompting. Our integrated approach balances engagement via personalized content with developing expertise via prompt engineering and critical analysis. Finally, our project will benefit EdUHK academic and teaching colleagues in leveraging LLMs in teaching STEM courses, as well as BEd and non-BEd students in the STEM field, in utilizing LLMs for their own learning of STEM courses and their future teaching career.

Code:

T0289

Principal Project Supervisors:

Keywords provided by authors:

Start Date:

15 Apr 2024

End Date:

14 Jun 2025

Status:

Ongoing

Financial Year:

2023-24

Type:

TDG