Статті

Постійне посилання зібрання

https://repository.sspu.edu.ua/handle/123456789/324

Переглянути

Перегляд Статті за Ключові слова "chemistry education"

Зараз показуємо 1 - 1 з 1
  • Документ
    Advantages and Limitations of Large Language Models in Chemistry Education: A Comparative Analysis of ChatGPT, Gemini And Copilot
    (CEUR Workshop Proceedings, 2024) Kharchenko Yuliia Volodymyrivna; Babenko Olena Mykhailivna; Харченко Юлія Володимирівна; Бабенко Олена Михайлівна
    This study aims to explore the potential and limitations of large language models (LLMs) such as ChatGPT, Gemini, and Copilot, in the context of chemistry education. The primary objective of the study is to compare the effectiveness of LLMs in solving chemistry tasks and to identify the key challenges associated with their implementation in education. These LLMs were selected based on a survey of students which indicated their widespread use due to their free accessibility. To evaluate the potential of LLMs in chemistry education, we employed them to solve tasks corresponding to different levels of knowledge in different subfields of chemistry. A comparative evaluation of LLMs' performance against that of average Ukrainian students was conducted. The results indicate that while LLMs show promise mainly in tasks not demanding deep logical reasoning, they are generally inferior to students. Key challenges in using LLMs in chemistry education identified include understanding the nuances of chemistry as a complex and multifaceted science, abstract concepts used in chemistry, recognition of chemical compound formulas, chemical reaction equations, limitations in logical reasoning, language barriers, and the occurrence of AI hallucinations. Additionally, there is a need for students to develop skills in crafting effective queries and prompts to enhance the efficiency of working with LLM. While LLMs are promising, their implementation requires addressing the identified limitations.