Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Generative AI for Enhancing Active Learning in Education: A Comparative Study of GPT-3.5 and GPT-4 in Crafting Customized Test Questions

Hamdireza Rouzegar, Masoud Makrehchi

TL;DR

The paper investigates how large language models (LLMs) can support active learning in education by generating customized test questions for Grade 9 mathematics. It introduces a novel teacher–student setup where GPT-4 acts as a question generator and GPT-3.5 as the learner, with an additional fine-tuning phase in which GPT-3.5 is trained on GPT-4 outputs. Results show GPT-4 yields higher-quality, correctly formed questions across topics, while GPT-3.5 demonstrates measurable learning gains after targeted teaching, albeit with limited benefit from explanatory content. The study highlights the potential of LLMs to enable personalized, adaptive assessment and learning pathways, while acknowledging scope, evaluation, and generalizability limitations that warrant further research across more subjects, contexts, and longer time horizons.

Abstract

This study investigates how LLMs, specifically GPT-3.5 and GPT-4, can develop tailored questions for Grade 9 math, aligning with active learning principles. By utilizing an iterative method, these models adjust questions based on difficulty and content, responding to feedback from a simulated 'student' model. A novel aspect of the research involved using GPT-4 as a 'teacher' to create complex questions, with GPT-3.5 as the 'student' responding to these challenges. This setup mirrors active learning, promoting deeper engagement. The findings demonstrate GPT-4's superior ability to generate precise, challenging questions and notable improvements in GPT-3.5's ability to handle more complex problems after receiving instruction from GPT-4. These results underscore the potential of LLMs to mimic and enhance active learning scenarios, offering a promising path for AI in customized education. This research contributes to understanding how AI can support personalized learning experiences, highlighting the need for further exploration in various educational contexts

Generative AI for Enhancing Active Learning in Education: A Comparative Study of GPT-3.5 and GPT-4 in Crafting Customized Test Questions

TL;DR

The paper investigates how large language models (LLMs) can support active learning in education by generating customized test questions for Grade 9 mathematics. It introduces a novel teacher–student setup where GPT-4 acts as a question generator and GPT-3.5 as the learner, with an additional fine-tuning phase in which GPT-3.5 is trained on GPT-4 outputs. Results show GPT-4 yields higher-quality, correctly formed questions across topics, while GPT-3.5 demonstrates measurable learning gains after targeted teaching, albeit with limited benefit from explanatory content. The study highlights the potential of LLMs to enable personalized, adaptive assessment and learning pathways, while acknowledging scope, evaluation, and generalizability limitations that warrant further research across more subjects, contexts, and longer time horizons.

Abstract

This study investigates how LLMs, specifically GPT-3.5 and GPT-4, can develop tailored questions for Grade 9 math, aligning with active learning principles. By utilizing an iterative method, these models adjust questions based on difficulty and content, responding to feedback from a simulated 'student' model. A novel aspect of the research involved using GPT-4 as a 'teacher' to create complex questions, with GPT-3.5 as the 'student' responding to these challenges. This setup mirrors active learning, promoting deeper engagement. The findings demonstrate GPT-4's superior ability to generate precise, challenging questions and notable improvements in GPT-3.5's ability to handle more complex problems after receiving instruction from GPT-4. These results underscore the potential of LLMs to mimic and enhance active learning scenarios, offering a promising path for AI in customized education. This research contributes to understanding how AI can support personalized learning experiences, highlighting the need for further exploration in various educational contexts
Paper Structure (15 sections, 1 figure, 2 tables, 1 algorithm)

This paper contains 15 sections, 1 figure, 2 tables, 1 algorithm.

Table of Contents

  1. Literature Review
  2. Personalized Education and Technology Integration
  3. AI in Educational Settings
  4. LLMs in Educational Contexts:
  5. Methodology
  6. Iterative Question-Answering Process Using LLM
  7. LLMs as Teacher and Student
  8. Experimental Setup
  9. Question-Answering Process
  10. Teaching the smaller model
  11. Evaluation and Results
  12. Comparative Results of LLMs as Teacher
  13. Results of LLMs as Teacher and Student
  14. Limitations and Future Work
  15. Conclusion

Figures (1)

  • Figure 1: Example of JSON input and output for Fine-Tuning GPT-3.5