Application of Large Language Models in Foreign Language Vocabulary Teaching

Traditional foreign language vocabulary teaching struggles with static materials and decontextualized learning, limiting learners' practical application. This study evaluates GPT4 and DeepSeek in teaching English near-synonyms ("fare" vs. "fee" and "promise" vs. &q...

Full description

Saved in:
Bibliographic Details
Published in2025 13th International Conference on Information and Education Technology (ICIET) pp. 204 - 208
Main Authors Li, Jingyi, Cao, Li, Ling, Shuzhen, Han, Fangting
Format Conference Proceeding
LanguageEnglish
Published IEEE 18.04.2025
Subjects
Accuracy
Adaptation models
AI-assisted education
Automation
Context modeling
Education
foreign language vocabulary teaching
Large language models
Pragmatics
semantic modeling
Semantics
Training
Vocabulary
Online AccessGet full text
DOI10.1109/ICIET66371.2025.11046265

Cover

More Information
Summary:Traditional foreign language vocabulary teaching struggles with static materials and decontextualized learning, limiting learners' practical application. This study evaluates GPT4 and DeepSeek in teaching English near-synonyms ("fare" vs. "fee" and "promise" vs. "assure"). Through the qualitative and quantitative analysis of the generated exercises, results reveal GPT-4's strength in semantic expansion and contextual adaptability, while DeepSeek reinforces grammatical accuracy through structured rules. A hybrid approach combining both models bridges foundational knowledge and pragmatic fluency, enabling learners to shift from memorization to contextual mastery. However, challenges include technological gaps, teacher training, and AI biases. This research highlights AI's role as a supplementary tool, advocating human-AI collaboration to enhance cross-cultural competence and effective learning.
DOI:10.1109/ICIET66371.2025.11046265