EduRAG: Crafting Clever Educational Chatbots and Advanced QA Systems with Retrieval-Augmented Generation

A tutorial presented by Dr Noorhan Abbas (University of Leeds) and Dr Saad Ezzini (King Fahd University of Petroleum and Minerals)
at COLING 2025

Place & Time

This tutorial will be conducted 100% online via Google Colab and GitHub. No special equipment beyond a stable internet connection and a computer capable of running web-based applications is required.

Agenda for the day

Total duration: 4 hours (240 minutes)

Time	Topic	Description
30 mins	Introduction and Overview	- Welcome and tutorial agenda - Introduction to RAG - Real-world examples
20 mins	Components and Mechanism of RAG	- Understanding RAG Components - RAG architecture - Component interactions
40 mins	Retrieval Methods, Vector Databases and LLMs	- Retrieval techniques - Vector databases and embedding models - LLMs and prompt engineering
Break (30 minutes)
90 mins	Hands-on Demonstration	1- Building educational chatbots with Unstructured API (45 mins) Download Tutorial Files 2- Creating QA systems with Haystack library (45 mins)
20 mins	Challenges and Future Directions	- Current limitations - Ongoing research - Future applications
10 mins	Q&A and Wrap-up	Final discussion and questions

Prerequisites

• Familiarity with Python programming
• Basic understanding of machine learning and NLP concepts
• Knowledge of transformer-based models
• Experience with NLP APIs and libraries
• Basic understanding of prompt engineering

Target Audience

• NLP Researchers and Practitioners
• AI and Machine Learning Engineers
• Educators and Educational Technologists
• Developers and Software Engineers
• Graduate Students and Academics

Reading List

• Gao, Y., et al. (2023). Retrieval-augmented generation for large language models: A survey.
• Lewis, P., et al. (2020). Retrieval-augmented generation for knowledge-intensive nlp tasks.
• Zhao, P., et al. (2024). Retrieval-augmented generation for ai-generated content: A survey.

Presenters

Dr Noorhan Abbas (University of Leeds, UK)
Educator and researcher at the School of Computing, specializing in Data Mining and Text Analytics. Module leader for Data Mining and Text Analytics in the online Artificial Intelligence MSc programme. Contributor to the Edubots Erasmus+ Knowledge Alliance project.

Dr Saad Ezzini (King Fahd University of Petroleum and Minerals, Saudi Arabia)
Assistant Professor specializing in AI, Software Engineering, and NLP. Research focuses on chatbots, RAG, and QA systems, with publications on quality assurance chatbots and Text-to-SQL applications in banking.