Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Using Large Language Models for education managements in Vietnamese with low resources

Duc Do Minh, Vinh Nguyen Van, Thang Dam Cong

TL;DR

This work addresses the lack of Vietnamese-specific, resource-efficient AI support for educational management by proposing VietEduFrame, a framework that applies LLMs to university administration tasks. It introduces a Vietnam-focused dataset derived from Hanoi National University regulations and demonstrates a low-resource deployment using LoRA fine-tuning on BLOOM and Vistral models, achieving notable improvements over baselines. The study combines synthetic data generation with human-in-the-loop labeling, and evaluates using Exact Match and F1 metrics, reporting favorable results and practical reductions in training time and memory. While promising, it also discusses limitations in generalizability and domain-specific robustness, outlining future work to enhance reliability in under-resourced educational settings.

Abstract

Large language models (LLMs), such as GPT-4, Gemini 1.5, Claude 3.5 Sonnet, and Llama3, have demonstrated significant advancements in various NLP tasks since the release of ChatGPT in 2022. Despite their success, fine-tuning and deploying LLMs remain computationally expensive, especially in resource-constrained environments. In this paper, we proposed VietEduFrame, a framework specifically designed to apply LLMs to educational management tasks in Vietnamese institutions. Our key contribution includes the development of a tailored dataset, derived from student education documents at Hanoi VNU, which addresses the unique challenges faced by educational systems with limited resources. Through extensive experiments, we show that our approach outperforms existing methods in terms of accuracy and efficiency, offering a promising solution for improving educational management in under-resourced environments. While our framework leverages synthetic data to supplement real-world examples, we discuss potential limitations regarding broader applicability and robustness in future implementations.

Using Large Language Models for education managements in Vietnamese with low resources

TL;DR

This work addresses the lack of Vietnamese-specific, resource-efficient AI support for educational management by proposing VietEduFrame, a framework that applies LLMs to university administration tasks. It introduces a Vietnam-focused dataset derived from Hanoi National University regulations and demonstrates a low-resource deployment using LoRA fine-tuning on BLOOM and Vistral models, achieving notable improvements over baselines. The study combines synthetic data generation with human-in-the-loop labeling, and evaluates using Exact Match and F1 metrics, reporting favorable results and practical reductions in training time and memory. While promising, it also discusses limitations in generalizability and domain-specific robustness, outlining future work to enhance reliability in under-resourced educational settings.

Abstract

Large language models (LLMs), such as GPT-4, Gemini 1.5, Claude 3.5 Sonnet, and Llama3, have demonstrated significant advancements in various NLP tasks since the release of ChatGPT in 2022. Despite their success, fine-tuning and deploying LLMs remain computationally expensive, especially in resource-constrained environments. In this paper, we proposed VietEduFrame, a framework specifically designed to apply LLMs to educational management tasks in Vietnamese institutions. Our key contribution includes the development of a tailored dataset, derived from student education documents at Hanoi VNU, which addresses the unique challenges faced by educational systems with limited resources. Through extensive experiments, we show that our approach outperforms existing methods in terms of accuracy and efficiency, offering a promising solution for improving educational management in under-resourced environments. While our framework leverages synthetic data to supplement real-world examples, we discuss potential limitations regarding broader applicability and robustness in future implementations.
Paper Structure (23 sections, 4 equations, 13 figures, 9 tables)

This paper contains 23 sections, 4 equations, 13 figures, 9 tables.

Table of Contents

  1. Introduction
  2. Related work
  3. Large language models in for study assisting
  4. Education toolkit
  5. Textbook question answering
  6. Dataset
  7. Building data
  8. Methodology
  9. Pre labeling and human labeling
  10. Training and context adaptation
  11. Training
  12. Context Adaptation
  13. Evaluate
  14. Result and Experiment
  15. Statistic of dataset
  16. ...and 8 more sections

Figures (13)

  • Figure 1: The examples of Question Answering in the education domain
  • Figure 2: Overview of our framework
  • Figure 3: Context Length Distribution
  • Figure 4: Question Length Distribution
  • Figure 5: Answer Length Distribution
  • ...and 8 more figures