Oral Presentation Time: 1400-1500

Venue: Fanling Room, Lower Level I

Presenter(s)

– Dr Richard Wing Cheung LUI, Senior Lecturer, Department of Computing, The Hong Kong Polytechnic University

Abstract

This presentation introduces the design and implementation of GPTutor, a Generative AI (GenAI) powered Intelligent Tutoring System (ITS) developed at the Hong Kong Polytechnic University (PolyU). GPTutor aims to enhance student learning experiences through personalised tutoring and interactive exploration. It helps students gain a deeper understanding of the course materials provided by their instructors. During the first phase of our implementation, we developed features for instructors to upload and manage their course content and to create learning scenarios based on the learning content. The system includes a conversational interface for students to ask questions and explore course content to deepen their understanding. As the answers are generated based on the instructor-uploaded content, GPTutor provides more factual responses, reduces hallucinations, and aligns better with the instructors’ intended learning outcomes (ILO). We will also share findings from our pilot study, which involved approximately 200 undergraduate and postgraduate students at PolyU. Finally, we will discuss our future plans for further development and enhancement of the platform.

Theme: 1. Showcase Project Achievements

Sub-theme: Innovative Technology-in-Education

Oral Presentation Time: 1400-1500
Poster Presentation Time: 1225-1400; 1500-1600

Oral Presentation Venue: Fanling Room, Lower Level I
Poster Presentation Venue: I2, Tai Po-Shek-O Room, Lower Level I

Presenter(s)

– Dr Ting Leung Albert LEE, Lecturer, Department of Electrical and Electronic Engineering, The University of Hong Kong

– Dr Victor LEE, Lecturer, Department of Electrical and Electronic Engineering, The University of Hong Kong

– Dr Zhengyuan WEI, Research Associate, Department of Electrical and Electronic Engineering, The University of Hong Kong

– Mr Alex KIANG, The University of Hong Kong

Abstract

Retrieval-augmented generation (RAG) has been demonstrated to be highly effective in generative AI applications, resulting in substantial improvement in accuracy and reliability of large language model responses. The success of this approach is accomplished through seamless integration of AI capabilities and practical knowledge base, which fosters an interactive learning mechanism conducive to automatic question-answering augmented with references and refined prompts, leading to a more vibrant and connected learning environment. Communication portals enable effective inquiry and prompt responses while the course-specific chatbot helps reduce teachers’ workload and streamline classroom management. In this project, the RAG approach is applied to a discipline-core course named Integrated Design Project (IDP) in the second semester 2023-24. This project-based course consists of 78 EEE undergraduate students with diverse programming experience. The IDP-specific AI chatbot is developed using the Coze platform with a Discord server. To name a few, the main contents of practical knowledge base include the procedures for setting up a Raspberry Pi webcam, the installation process of Jetson Inference library on Jetson Nano, how to use YOLO model for object detection, how to install Jetson inference library on Jetson Nano, etc. The anonymous feedback survey conducted at the end of the course shows high utilization and satisfaction of the chatbot, confirming the effectiveness of this approach in facilitating students’ learning.

Theme: 1. Showcase Project Achievements

Sub-theme: Innovative Technology-in-Education