透過您的圖書館登入
IP:3.146.37.183
  • 期刊
  • OpenAccess

結合AI語音評測的聊天機器人於華語語音教學成效之探究—以《學華語向前走》K1入門冊為對象

A Study Combining AI Chat Robots in Chinese Phonetic Teaching and Learning

摘要


華語為全球第二大外語,為了提供優質的華語文教學資源,行政院僑務委員會多年前即啟動《學華語向前走》新編教材計畫,邀集國內外華語教學研究者、學科專家及資深華語教師共同參與編訂,以因應華語做為第二語言(Chinese as a Second Language)的教學趨勢及學習需求。並且,該教材已全面線上數位化,以提供更為便捷的使用模式,且能符應當代科技數位教學的趨勢。目前,華語教學已經開展出紙本、數位等教材形式,以利教學的即時應用。就《學華語向前走》為例,僑委會於網頁上說明其提供的教材形式,計有:聲音、影像、文字、圖片、簡報等。藉由上述資訊我們得知,科技的協助將教學相關資訊及素材加以整合及深化,使教學系統的建構趨於完善。舉教材為例:從傳統實體書籍、圖卡的書面模式,逐步來到影音、APP應用程式、AR擴增實境、VR虚擬實境、線上平台……等數位型態,當傳統紙本教材與當代數位科技結合後,華語教學逐步突破真人實境之傳統教學模式,渐渐轉型成可以不受時空限制的「線上華語教學」。第二語言學習者的華語發音普遍存在著「洋腔洋調」的問題,且這種「外國口音」(accent)的中文,不會隨著學習者華語水平的提高而遞減;在高階的華語學習者的中文發音裡,我們發現「洋腔洋調」的現象依然存在,這個部分與華語是一種具有「聲調」的語言相關。目前華語課程裡的語音教學,以糾正學生的發音偏誤(errors)為主,並且,以語音產出的正確與否為主要訓練目標。實體華語課堂教學及線上華語教學中,我們發現,教師有其教學的局限性。它的意思係指,實際教學現場常常是一對多的授課情況,也就是,教師一人,同時面對不同程度及需求的學生。如此,難免會有不能全面性顧及學生學習狀況的現象,亦即,教師無法對所有學生進行個別的差異化指導,或提供適性的評斷。是故,差異化教學及評量的理想,在一對多的華語教學現場,有其根本性成因的限制。本研究試圖以聊天機器人為輔助介面來建構華語教材,藉由自然語言處理(Natural Language Processing)作為引導學生學習的基礎,並且,結合語音辨識進行發音學習與成效評測,給予學生反覆練習的機會與即時性之回饋,以強化學生學習成效、提升學習動機與態度。更期待據此以解決實體課堂或是線上課程時,一位教學者與多位學習者共構課程時之局限。此外,本研究更進階的研究目標在收集學習者於歷程中的文字或語音回饋之資料與數據,以進行分析,俾能提供學習者更為優化的學習路徑。

並列摘要


The Chinese pronunciation of second language learners generally has the problem of "foreign accents", and the Chinese with this "foreign accent" will not decrease with the improvement of the learners' Chinese proficiency. In the Chinese pronunciation of learners with higher language proficiency, we found that the phenomenon of "foreign accents" still exists. This part is related to the fact that Chinese is a language with "tones". In general Chinese classroom teaching and online Chinese teaching, we found that teachers have their teaching limitations. It means that the teaching site is often one-to-many. Therefore, the ideal of differentiated teaching and suitability assessment has its fundamental limitations in the one-to-many Chinese teaching scene. This research attempts to construct Chinese textbooks using chat robots as the interface, natural language understanding as the basis for guiding students' learning, combining with speech recognition to learn and evaluate learners' pronunciation and pronunciation, and to give students opportunities for repeated practice and instant feedback to strengthen students' learning effectiveness, enhancing students' learning motivation and attitudes. In addition, the further goal of this research is to collect the text or voice feedback data and data of learners in the course for collection and analysis, so as to provide learners with a more optimized learning path.

參考文獻


Chen, J. C., Whittinghill, D. C., & Kadlowec, J. A. (2010). Classes that click: fast, rich feedback to enhance student learning and satisfaction. Journal of Engineering Education, 99(2), 159–168.
Dodson, S. L. (2000). FAQs: Learning Languages through Drama. Texas papers in foreign language education, 5(1), 129-141.
Flege, J. E. (1995). Second language speech learning: Theory, findings, and problems. Speech perception and linguistic experience: Issues in cross-language research, 92, 233-277.
Frost, R., Katz, L., & Bentin, S. (1987). Strategies for visual word recognition and orthographical depth: A multilingual comparison. Journal of Experimental Psychology: Human Perception and Performance, 13, 104-115.
Lively, S. E., Logan, J. S., & Pisoni, D. B. (1993). Training Japanese listeners to identify English/r/and/l/. II: The role of phonetic environment and talker variability in learning new perceptual categories. The Journal of the acoustical society of America, 94(3), 1242-1255.

延伸閱讀