Towards a comprehensive method for evaluating and utilizing AI-generated bilingual lexicographic data in language learning using the example of Chinese as a foreign language

Rink C, Ganslmayer C, Evert S (2024)

Publication Language: English

Publication Type: Conference contribution

Publication year: 2024

Publisher: 東洋大学 (Toyo University)

Series: AsiaLex 2024 Proceedings

City/Town: Tokyo

Pages Range: 133–142

Conference Proceedings Title: Asian Lexicography - Merging cutting-edge and established approaches

Event location: Toyo University, Tokyo JP

ISBN: 978-4-9901771-2-6

URI: https://drive.google.com/file/d/11eiezuzidbVhlr-0Cr_k2NxgTg5CEUcP/view

Abstract

In this paper, we take an analytical and a practical approach to the topic of AI-generated lexicographical data. In the context of a language-learning environment, we want to take a closer look at the quality of bilingual lexicographical data (Chinese-German) produced by generative large language models (LLMs). Our aim is to create a scientifically supported and comprehensive method that includes effective prompting strategies, an objective evaluation of the generated lexicographical data, and learner strategies for the use of conversational AI. The focus of our paper will be on the evaluation method, for which we use a multi-level analysis based on lexicographical standards, in comparison with existing lexicographical resources made by experts, the consideration of the findings of a didactical experiment, as well as a discussion about AI specific questions, and on the proposal of a didactical concept, which introduces a reasonable use of conversational AI chatbots in a language-learning environment. The results of this study are expected to give further insight into the functionality of generative LLMs in lexicographical settings, to provide a critical framework for the evaluation of AI-generated lexicographical data, and offer perspectives on how it can be used as a valuable lexicographical tool in language-learning.

Authors with CRIS profile

Christian Rink Professur für Germanistische Linguistik mit dem Schwerpunkt Lexikographie Christine Ganslmayer Lehrstuhl für Germanistische Sprachwissenschaft Stephanie Evert Lehrstuhl für Korpus- und Computerlinguistik

How to cite

APA:

Rink, C., Ganslmayer, C., & Evert, S. (2024). Towards a comprehensive method for evaluating and utilizing AI-generated bilingual lexicographic data in language learning using the example of Chinese as a foreign language. In Ai Inoue, Naho Kawamoto, Makoto Sumiyoshi (Eds.), Asian Lexicography - Merging cutting-edge and established approaches (pp. 133–142). Toyo University, Tokyo, JP: Tokyo: 東洋大学 (Toyo University).

MLA:

Rink, Christian, Christine Ganslmayer, and Stephanie Evert. "Towards a comprehensive method for evaluating and utilizing AI-generated bilingual lexicographic data in language learning using the example of Chinese as a foreign language." Proceedings of the AsiaLex 2024, Toyo University, Tokyo Ed. Ai Inoue, Naho Kawamoto, Makoto Sumiyoshi, Tokyo: 東洋大学 (Toyo University), 2024. 133–142.

BibTeX: Download