AI in Education: Change at the Speed of Learning（Ⅵ） 譯題 《教育中的人工智能：學(xué)習(xí)速度的變化》（六）

Product by UNESCO IITE （The UNESCO Institute for Information Technologies in Education）聯(lián)合國教科文組織教育信息技術(shù)研究所

該報告為UNESCO IITE新推出的“教育的數(shù)字化轉(zhuǎn)型”（Digital Transformation of Education）系列出版物的第一期。據(jù)悉，“教育的數(shù)字化轉(zhuǎn)型”系列將包括政策簡報、分析報告和反思論文，以探討由于技術(shù)的使用及其對教育和人類生活其他領(lǐng)域的影響，教育領(lǐng)域正在發(fā)生和正在出現(xiàn)的根本性變化。

5. Whats Next？

5.2 New Ways of Seeing

5.2.1 Cognitive Services

Cognitive Services are a set of emergent technologies which help developers to create applications that can see， hear， speak， understand， and even begin to reason. They allow developers to add features such as emotion and sentiment detection， vision and speech recognition， and natural language understanding to their applications without having direct AI or data science skills or knowledge themselves. Because these services are made available to developers by large companies such as Microsoft — as sets of AI algorithms and application programming interfaces （APIs） — they can easily be added as enhancements to existing solutions.

Such technologies are already becoming ubiquitous in the commercial world. Computer vision and speech recognition services are now widely used as a means of unlocking a smartphone. National Language Processing （NLP） allows people to perform web searches simply by asking a question in their own voice， using everyday words and phrases. Text to speech and speech to text services can be used to both create and narrate documents and webpages.

The implications for education are immense. For learners with a disability in particular they can help to remove barriers of access and opportunity， and serve to democratize education. For learners with foundational learning challenges， such as literacy and numeracy deficiencies， they can enable active participation in higher-order learning at the same time as the students develop those skills， rather than excluding them until those capabilities are mastered. When added to existing AI solutions and services they can accelerate the generation of vital insights — an educator can simply ask the AI a question， rather than having to input a detailed query using their keyboard or mouse.

5.2.2 Virtual， Mixed and Augmented Reality

VR， AR and MR learning content can revolutionize education， making learning immersive and more engaging. With VR， students can travel anywhere in the world from their classroom and experience a recreation of that world， at any time in history， within a virtual environment. With AR， students can view holographic instructions or information overlaid upon a physical object， providing them with additional information about the object they are viewing. MR can be used， for example， in medical training or engineering， allowing students to ‘dissect limbs or ‘repair machinery which they view as holograms within the real-world environment， enabling learners to focus on practice rather than theory. Although these new technologies are still evolving， the speed of change is exponential and some education technologists are already predicting that， in the future， low-cost versions of such technologies will supplant the textbook and free learners from the confines of their classroom.

AI will make it possible not just to deploy these exciting new technologies — but to analyze their effectiveness and to optimize the benefits they can provide to students and educators alike.

5.2.3 IoT and Edge Computing

The Internet of Things （IoT） describes the network of physical objects （things） that are embedded with sensors， software and other technologies， which allow them to connect and exchange data with other devices and systems over the internet， without involving a human being. An IoT device can be as simple as a lightbulb which can be turned on and off remotely， or as complex as a driverless car. There are already more connected ‘things than there are people， and researchers have predicted that there will be more than 40 billion such devices by 2025.

The IoT is already being used by many education institutions to help manage their infrastructure more efficiently and cheaper — from using sensors to control heating and lighting based on the actual occupancy of classrooms or buildings， leading to increased energy efficiency and reduced operating costs; to the use of remote cameras and biometrics to help make campuses more secure.

Edge computing is an information technology （IT） architecture in which data is processed at the periphery （edge） of the network， as close to the user or the source of the data as possible. This means that the data does not need to be sent to a centralized （cloud） location to be processed or stored， reducing time and bandwidth requirements. The key benefit of edge computing is the potential reduction in latency — simply put， things can happen faster.

5. 未來展望

5.2 視覺方式變革

5.2.1 認(rèn)知服務(wù)

認(rèn)知服務(wù)是一種新興技術(shù)，認(rèn)知服務(wù)技術(shù)的發(fā)展可以幫助開發(fā)人員創(chuàng)建能夠看、聽、說、理解甚至開始推理的應(yīng)用程序。它們允許開發(fā)人員在應(yīng)用程序中添加諸如情感和情感檢測、視覺和語音識別以及自然語言理解等功能，而無需擁有直接的人工智能或數(shù)據(jù)科學(xué)技能或知識。這些服務(wù)是由大型公司（如微軟）向開發(fā)人員提供的，作為人工智能算法和應(yīng)用程序編程接口的集合，可以被輕松地添加為現(xiàn)有解決方案的增強功能。

此類技術(shù)在商業(yè)領(lǐng)域中早已無處不在。計算機視覺和語音識別服務(wù)現(xiàn)在被廣泛用作解鎖智能手機的手段。自然語言處理技術(shù)允許人們通過自己的聲音，使用日常詞匯和短語，簡單地進(jìn)行網(wǎng)絡(luò)搜索。無論是文本到語音的服務(wù)，還是語音到文本的服務(wù)，都可用于創(chuàng)建和敘述文檔和網(wǎng)頁。

認(rèn)知服務(wù)技術(shù)對教育的影響是巨大的，特別是對殘疾學(xué)生，這一技術(shù)可以幫助消除入學(xué)難和機會不平等問題，促進(jìn)教育民主化。對有基礎(chǔ)學(xué)習(xí)挑戰(zhàn)的學(xué)習(xí)者，例如讀寫和計算能力不足的人，他們可以在發(fā)展這些技能的同時，積極參與更高階的學(xué)習(xí)，而不是在掌握這些能力之前將其排除在外。將認(rèn)知服務(wù)技術(shù)添加到現(xiàn)有的人工智能解決方案和服務(wù)中，可以加速重要見解的生成。教育者可以簡單地向人工智能提出一個問題，而不必使用鍵盤或鼠標(biāo)輸入詳細(xì)的查詢。

5.2.2 虛擬現(xiàn)實、混合現(xiàn)實與增強現(xiàn)實

虛擬現(xiàn)實、增強現(xiàn)實和混合現(xiàn)實學(xué)習(xí)內(nèi)容可以徹底改變教育，使學(xué)習(xí)更具沉浸感和吸引力。有了虛擬現(xiàn)實，學(xué)生們可以在虛擬環(huán)境中，在歷史上的任何時候，從教室到世界的任何地方旅行，體驗對這個世界的再創(chuàng)造。通過增強現(xiàn)實，學(xué)生可以查看覆蓋在物理對象上的全息指令或信息，為他們提供有關(guān)所查看對象的附加信息。而混合現(xiàn)實可用于醫(yī)學(xué)訓(xùn)練或工程，允許學(xué)生在真實環(huán)境中“解剖”四肢或“修理”他們視為全息圖的機械，使學(xué)習(xí)者能夠?qū)Ｗ⒂趯嵺`而不是理論。盡管這些新技術(shù)仍在不斷發(fā)展，但變化的速度是指數(shù)級的，一些教育技術(shù)專家已經(jīng)預(yù)測，在未來，這些技術(shù)的低成本版本將取代教科書，讓學(xué)習(xí)者徹底擺脫課堂的束縛。

人工智能不僅使得部署這些令人興奮的新技術(shù)成為可能，而且能夠分析新技術(shù)的有效性并優(yōu)化它們，為學(xué)生和教育工作者帶來好處。

5.2.3 物聯(lián)網(wǎng)與邊緣計算

物聯(lián)網(wǎng)是指嵌入傳感器、軟件和其他技術(shù)的物理對象（物）所形成的網(wǎng)絡(luò)，允許它們通過互聯(lián)網(wǎng)與其他設(shè)備和系統(tǒng)連接并交換數(shù)據(jù)，而不涉及人。物聯(lián)網(wǎng)設(shè)備既可以像遠(yuǎn)程打開和關(guān)閉電燈一樣簡單，也可以像無人駕駛汽車一樣復(fù)雜。聯(lián)網(wǎng)的“物品”已經(jīng)比人多，研究人員預(yù)測，到2025年，此類設(shè)備將超過400億臺。

許多教育機構(gòu)已經(jīng)在使用物聯(lián)網(wǎng)來幫助更高效、更廉價地管理其基礎(chǔ)設(shè)施——例如使用傳感器根據(jù)教室或建筑物的實際占用情況來控制加熱和照明，從而提高能源利用效率并降低運營成本;使用遠(yuǎn)程攝像頭和生物識別技術(shù)幫助校園更加安全。

邊緣計算是一種信息技術(shù)架構(gòu)，其中客戶端數(shù)據(jù)在網(wǎng)絡(luò)外圍（邊緣）進(jìn)行處理，盡可能靠近用戶或數(shù)據(jù)源。這意味著數(shù)據(jù)不需要發(fā)送到集中的（云端）位置進(jìn)行處理或存儲，從而減少了時間和帶寬需求。邊緣計算的主要好處是潛在的延遲減少——簡單地說，就是事情可以處理得更快。