| 摘要: | 大多數關於英語作為外語(EFL)的研究都提到了，教師通過利用人工智能(AI)產 生學習反饋來輔助教學可以提高 EFL 學習者的學習成就和學習動機。然而，這些教育 學習的互動過程只有益於提高人類的智力和知識。然而，隨著技術的飛速發展，不僅 可以為人類提供教育，還可以為萬物(things)，例如數位或是物理的事物提供教育。因 此，在這項研究中，將利用人工智能(AI)、物聯網(IoT)和行動計算(edge computing)的結 合來教育人類，也教育萬物(all things)。因此，我們提出 XoT (Xducation of Things)框架 來教育人類、人工智能、物理對象和數位對象。
在這項研究中，我們提出了 XoT 框架作為一種新的教育視角，通過問答(Q&A)互 動促進萬物相互學習。更具體地說，XoT 框架被提出來教育人類和萬物，包括人工智 能、物理對象和數位對象。此外，我們的邊緣計算(edge computing)可以使物理對象和 數位對象具有理解人類語言的能力，這被稱為智能事物(smartthings)。因此，除了可以 在 XoT 環境中相互教授和學習包括智能事物在內的萬物外，作為人類的 EFL 學習者也 可以從 Q&A 互動中受益。因此，我們進行了兩項實證研究，包括開發 XoT 環境，並根 據我們提出的 EFL 學習 XoT 框架進行了準實驗。
在第一項研究中，我們設計並開發了一個用於 EFL 學習的 XoT 環境，並研究了 萬物如何通過 Q&A 互動來建立並提升他們的知識。此外，我們還開發了問題轉發機制 (question forwarding mechanism)，幫助萬物快速提升知識庫的建立。一項準實驗包含 22 名 EFL 學習者，隨機分為兩組，實驗組(EG)由 11 名 EFL 學習者組成，控制組(CG)也由 11 名 EFL 學習者組成，分別在有與沒有問題轉發機制的 XoT 環境中學習。本研究結果 以混和方法深入分析。結果表明，相比於 CG，使用具有問題轉發機制的智能事物的 EG 具有三倍豐富的 Q&A 互動知識基礎。因此，EG 中的 EFL 學習者獲得了更高質量的 Q&A 互動。此外，EFL 學習者與 XoT 環境中的萬物之間的交互不僅增加了智能事物的 知識基礎，而且還幫助 EG 與 CG 中的 EFL 學習者提高其 EFL 寫作表現。此外，EG 學 習者認為，在與智能事物進行 Q&A 互動後，XoT 環境還可以提高他們的 EFL 寫作能 力。
在第二項研究中，我們通過將情境化(contextualization)與智能 Q&A 互動相結合， 改進了第一項研究中的 XoT 環境，並稱之為智能 XoT 環境。詳細地說，通過識別技術 從周圍環境中識別出語境中的信息，稱為情境化(contextualization)，並用於輔助智能 XoT 環境中的學習。此外，我們還開發了智能問題轉發機制，以幫助萬物，尤其是智 能事物快速且高效地建立自己特定的知識庫。一項準實驗包含 26 名 EFL 學習者，隨機 分為兩組，EG 由 13 名 EFL 學習者組成，CG 由 13 名 EFL 學習者組成，分別在有與沒 有智能問題轉發機制的智能 XoT 環境中學習。本研究結果以混和方法深入分析。結果 表明，相較於 CG，使用具有智能問題轉發機制的智能事物的 EG 具有四倍豐富的特定 智能 Q&A 互動知識基礎。具體來說，與 CG 學習者相比，在第一階段的智能 Q&A 互動 中，EG 學習者與智能事物的智能 Q&A 互動質量更高，答案也更好。此外，EG 學習者 在第二階段的 Q&A 互動中也收到了更高質量的提問。不只如此，EFL 學習者與智能 XoT 環境中的萬物之間的互動不僅增加了智能事物的特定知識基礎，而且還幫助了兩 組中的 EFL 學習者利用情境中富有意義的資訊，在真實語境中提升 EFL 寫作。在 EFL 學習方面，我們發現徹底回顧(total revisions)是提高 EG 作文質量的重要過程。此外， EG 學習者認為智能 XoT 環境有助於促使和促進他們根據情境中的訊息撰寫出更有意義 的文章。因此，智能 XoT 環境可以幫助和促進萬物建立特定的知識基礎，幫助 EFL 學 習者提高寫作能力。
;Most studies in English as a foreign language (EFL) mentioned that teachers supported by harnessing artificial intelligence (AI) to generate learning feedback could enhance EFL learners’ learning achievement and motivation. However, the educational learning process of these interactions only educates and enhance human intelligence and knowledge. Furthermore, with the rapid development of technologies, there is a high potential to extend education to not only for human beings but also to all things, including digital or physical objects. Therefore, in this study, the combination of AI, Internet of Things (IoT), and mobile computing (formed as edge computing) will be utilized not only to educate human beings but also to all things. Hence, the XoT (Xducation of Things) framework was proposed to educate humans, AI, physical objects, and digital objects.
In this study, we proposed the XoT framework as a new education perspective, incorporating question and answering (Q&A) interactions to facilitate mutual learning among all things. To be more specific, the XoT framework was proposed to educate human beings and all things, including AI, physical objects, and digital objects. Furthermore, our edge computing approach empowers physical objects and digital objects with the ability to understand the human language, which is referred to as smartthings. Hence, besides all things including smartthings can be taught and learned from each other in the XoT environment, EFL learners as human beings also benefit from the Q&A interactions. Therefore, we conducted two empirical studies, including the development of XoT environments and quasi-experiments based on our proposed XoT framework for EFL learning.
In the first study, we designed and implemented an XoT environment specifically for EFL learning. Our investigation focused on examining how all things can effectively construct and build their knowledge through Q&A interactions. In addition, we developed a question forwarding mechanism (QFM) to help all things building their knowledge base rapidly. A quasi experiment was conducted with a total of 22 EFL learners that divided randomly into two groups, the experimental group (EG) consist of 11 EFL learners and the control groups (CG) consists of 11 EFL learners learn in the XoT environment with/without the QFM, respectively. A mix-method analysis was employed to investigate the results deeply. The results revealed that the EG, which utilized smartthings with the QFM were three times richer in the knowledge bases compared to CG. Hence, EFL learners in EG received better quality of Q&A interactions. Furthermore, the interactions between EFL learners and all things in the XoT environment not only contributed to increase the knowledge bases of smartthings but also help EFL learners enhance EFL writing in both groups. In addition, EG learners perceived that the XoT environment had potential to enhance their EFL writing skills after having the Q&A interactions with smartthing.
The results revealed that EG utilizing smartthings with the Smart QFM had achieved four times increase the specific knowledge bases for smart Q&A interactions compared to CG. In detail, EG learners received better quality of smart Q&A interactions with smartthings along with better answers compared to CG learners during in the first phase of smart Q&A interactions. In addition, the EG learners also received better quality questions during the second phase of Q&A interactions. Furthermore, the interaction between EFL learners and all things in the smart XoT environment not only increased the specific knowledge bases of smartthings but also helped EFL learners enhance their EFL writing in authentic contexts with meaningful content from contextual information. |
