# Curriculum-Driven Edubot: A Framework for Developing Language Learning Chatbots Through Synthesizing Conversational Data

**Authors:** Yu Li, Shang Qu, Jili Shen, Shangchao Min, Zhou Yu  
**arXiv:** [2309.16804](https://arxiv.org/abs/2309.16804)  
**Date:** 2023-09-28  
**Summary page:** https://www.summarizepaper.com/en/paper/2309.16804v1/

## Key points

- The paper discusses the use of chatbots in education and their impact on student learning.
- The authors propose a framework called Curriculum-Driven EduBot that combines chatbot interactivity with structured textbook content.
- The approach involves extracting relevant topics from textbooks and using large language models to generate dialogues related to these topics.
- User studies show that this chatbot outperforms ChatGPT in leading curriculum-based dialogues and adapting its dialogue to match the user's English proficiency level.
- By integrating traditional textbook methodologies with conversational AI, this approach provides learners with an interactive tool that aligns with their curriculum and offers personalized conversation practice.
- This facilitates meaningful interactions between students and the chatbot, enhancing the overall learning experience within the pedagogical framework of the curriculum.
- Overall, this research presents a comprehensive framework for developing language learning chatbots that effectively assist students in improving their conversational skills by combining textbook materials with advanced AI techniques.

## Summary

The paper "Curriculum-Driven Edubot: A Framework for Developing Language Learning Chatbots Through Synthesizing Conversational Data" discusses the growing use of chatbots in education and their impact on student learning. The authors propose a framework called Curriculum-Driven EduBot that combines chatbot interactivity with structured textbook content to help students improve their conversational skills. This approach involves extracting relevant topics from textbooks and using large language models to generate dialogues related to these topics. These generated conversations are then used to fine-tune an open-source Language Model (LLM), resulting in the development of a curriculum-driven chatbot. User studies show that this chatbot outperforms ChatGPT in leading curriculum-based dialogues and adapting its dialogue to match the user's English proficiency level. By integrating traditional textbook methodologies with conversational AI, this approach provides learners with an interactive tool that aligns with their curriculum and offers personalized conversation practice. This facilitates meaningful interactions between students and the chatbot, enhancing the overall learning experience within the pedagogical framework of the curriculum. Overall, this research presents a comprehensive framework for developing language learning chatbots that effectively assist students in improving their conversational skills by combining textbook materials with advanced AI techniques.