- 期刊
運用生成式語言模型訓練高中生學測中英翻譯偵錯及校正能力之教學模式
A Teaching Module Utilizing a Generative Language Model to Train High School Students to Detect and Correct Errors in Chinese-English Translation Drills for SAT
摘要
本文作者開發一套以建構主義為理論基礎的「七步驟溝通式翻譯教學模組」，旨在運用生成式語言模型ChatGPT-3.5輔助高中生進行中英翻譯練習，藉由七道教學步驟以提升學生偵錯及校正能力。教案設計理念包括引導學生使用語言模型修正英譯文，逐步培養其對英文的各語言層次之掌握度。教學目標為增進學生「兩種語言文化的溝通和文本能力」與「專業和工具能力」等方面的知能。教學流程包含學生分組翻譯中文短文、使用ChatGPT-3.5輔助修正譯文、進行同儕討論及修改、自行檢視及修正譯文等。研究者觀察到學生對於使用生成式語言模型輔助中英翻譯的反應普遍正向，並透過實際教學驗證此教學模式有助於提升學生的語言能力及溝通能力。然而，研究者也指出使用語言模型的缺失，例如ChatGPT-3.5在修正學生英譯文時出現過度修改的情況等。總的來說，本研究提供了一種新穎的教學模式，結合生成式語言模型與翻譯教學，除可提升高中學生的中英翻譯能力及語言技能，亦為翻譯教學研究領域帶來許多啟發及可能性。
並列摘要
This article introduces a novel teaching module rooted in constructivist learning theory, designed to leverage the capabilities of a generative language model, ChatGPT-3.5, to enhance senior high school students' independent error detection and correction skills in Chinese-English translation. The module's seven teaching steps guide students in systematically utilizing ChatGPT-3.5 to revise their translations, fostering proficiency across various linguistic levels of English. Its pedagogical objectives aim to strengthen both communication and textual skills across language cultures. The instructional process involves student-led translation, ChatGPT-3.5-assisted revision, collaborative discussions, peer review, and self-assessment, culminating in corrected translations. Practical implementation revealed generally positive student responses to this innovative pedagogy, demonstrating effectiveness in enhancing language and communication skills. While acknowledging limitations, such as potential excessive modifications by ChatGPT-3.5, the study underscores the need for further research of such tools for optimal pedagogical value. Overall, this research presents a groundbreaking approach that seamlessly blends a generative language model with translation instruction, offering promising avenues for advancement within translation teaching research.