WEKO3
アイテム
Can Artificial Intelligence Understand and Talk about Metaphors? : An Empirical Study on the Metaphoric Competence of ChatGPT
https://kiu.repo.nii.ac.jp/records/2000048
https://kiu.repo.nii.ac.jp/records/2000048277408f7-99ac-40b2-a6dd-0c0328524c48
| 名前 / ファイル | ライセンス | アクション |
|---|---|---|
Can Artificial Intelligence Understand and Talk about Metaphors An Empirical Study on the MetaphoricCompetence of ChatGPT.pdf (184 KB)
|
|
| Item type | 紀要論文 / Departmental Bulletin Paper(1) | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 公開日 | 2024-06-28 | |||||||||||||||
| タイトル | ||||||||||||||||
| タイトル | Can Artificial Intelligence Understand and Talk about Metaphors? : An Empirical Study on the Metaphoric Competence of ChatGPT | |||||||||||||||
| 言語 | en | |||||||||||||||
| 言語 | ||||||||||||||||
| 言語 | eng | |||||||||||||||
| キーワード | ||||||||||||||||
| 言語 | en | |||||||||||||||
| 主題Scheme | Other | |||||||||||||||
| 主題 | language intelligence; metaphorical competence; metaphor identification; metaphor comprehension; metaphor generation | |||||||||||||||
| 資源タイプ | ||||||||||||||||
| 資源タイプ識別子 | http://purl.org/coar/resource_type/c_6501 | |||||||||||||||
| 資源タイプ | departmental bulletin paper | |||||||||||||||
| 著者 |
LU, Haoyu
× LU, Haoyu
× HAN, Tao
× SUN, Yutao
× LIU, Chuan
|
|||||||||||||||
| 抄録 | ||||||||||||||||
| 内容記述タイプ | Abstract | |||||||||||||||
| 内容記述 | This article introduced a metaphorical competence test scheme tailored for Artificial Intelligence language models and conducted a comprehensive evaluation of ChatGPT’s metaphorical competence across three levels: metaphor identification, comprehension, and generation. The study revealed the following findings: (1) ChatGPT exhibits robust metaphor identification competence at the lexical level, but displays shortcomings in recognizing certain metaphors, such as container metaphors and primary metaphors. (2) While ChatGPT excels in understanding the conventionality of language expression and metaphorical inference tasks, it may face hallenges in tasks related to omprehending similarity and inducing various metaphor types. (3) Regarding metaphor generation, ChatGPT tends to produce conventional metaphors and lacks sufficient novelty. The future direction of research will explore topics such as enhancing the language proficiency of artificial intelligence and conducting comparative studies between human and machine in terms of metaphorical competence. |
|||||||||||||||
| 言語 | en | |||||||||||||||
| bibliographic_information |
ja : 九州国際大学国際・経済論集 号 13, p. 41-69, 発行日 2024-03 |
|||||||||||||||
| 出版者 | ||||||||||||||||
| 出版者 | 九州国際大学現代ビジネス学会 | |||||||||||||||
| 言語 | ja | |||||||||||||||
| item_10002_source_id_162 | ||||||||||||||||
| 収録物識別子タイプ | NCID | |||||||||||||||
| 収録物識別子 | AA12815608 | |||||||||||||||
