论文翻译——Artificial Intellligence：miracle or mirage？

admin管理员组

文章数量:1584215

Artificial intelligence: miracle or mirage?

人工智能是奇迹还是幻想

section 1

• Artificial intelligence (AI) is today one of the most widely hyped technologies. Since the advent of the first computers, mathematical models have increasingly been used to support humans in an ever larger set of decision-making processes. Whether employed in the human resources area to help determine who gets hired for a job, or in the banking sector to select approved recipients for a loan, machines have been continuously making inroads into domains hitherto reserved to human judgment and adjudication. With the digitalization of many industries making large sets of data available, AI began to be the focus of renewed interest for its potential in solving an ever increasing number of problems. Machine learning techniques grew more and more powerful and sophisticated, in particular in the context of what are known as artificial neural networks (ANNs). Developed in the middle of the 20th century as a mathematical curiosity inspired by biology, neural networks have become one of the cornerstones of AI.

• 人工智能（AI）是当下被最广泛炒作的技术之一。自从第一台计算机问世以来，数学模型在更加多的领域里来辅助人类进行决策。无论是在人力资源方面，决定被雇佣的对象，还是在银行领域，决定批准的贷款对象，计算机正在不断进入那些迄今为止还由人类进行判定和裁决的领域。随着工业化的数字化正在不断加深，能够获得的数据量也越来越大，AI也因为它有潜力解决更多的问题，而成为人们重新关注的焦点。尤其是在人造神经网络成熟的背景下，机器学习技术正变的越来越强大和全面。神经网络这项技术仅仅实在20世纪中期受生物学启发，而产生的数学奇思，而今已经变成了AI领域的奠基石。

• 生词：

  • hype：议论，大肆宣传
  • ever：愈发，更加，ever larger 愈发的大
  • making inroads into ：侵袭，侵入
  • hitherto：迄今为止
  • context：背景，环境

• However, it was not until 2010 and later that dramatic improvements in machine learning, commonly referred to as deep learning, paved the way for an explosion of AI. With computing power increasing steadily, very large (“deep”) neural networks began to provide machines with novel capabilities that would have been too complex or even impossible to implement using traditional programming techniques. Since then, technologies such as computer vision and natural language processing (NLP) have been completely transformed and are being deployed on a massive scale in many different products and services. Deep learning is now being applied in a large number of industries, such as manufacturing, healthcare or finance, to uncover new patterns, make predictions and guide a wide variety of key decisions.However impressive such recent developments have been, AI remains today very much task focused and centered around well-defined pattern recognition applications. While current research is working dynamically to equip machines with human-like skills such as contextual awareness or empathy, attainment of this objective is, according to many AI scientists, still far ahead in the future.

• 然而，直到2010年之后，机器学习（通常指的是深度学习）才发生巨大的变化，为AI的进一步爆发铺平了路。随着算力的快速提升，越来越庞大的神经网络使得计算机具备之前新的能力，这些能力他过于复杂，以至于传统的机器学习并不可能实现。正是自此以后，一些诸如计算机视觉和自然语言处理的技术深度学习完全转变，并且被大规模部署在不同的产品和服务上。深度学习当前正在被应用在大量的产业中，比如说，工业，卫生行业或者金融业，用深度学习来挖掘更多的特征，做出预测，指导不同领域的关键决策。然而无论AI进来做出的发展有多让人印象深刻，今天AI仍旧是以任务为中心，和以预先设计好的模式识别应用为中心。尽管当前的研究正在动态地让机器拥有类人的能力，比如说情景意识或者同情心，但是许多人工智能的科学家说，要实现这个目标仍旧需要很遥远。

• 生词：

  • dramatic：巨大而突然的，激动人心的
  • capability：能力
  • contextual：同理心，同情心
  • attainment：达到，获得，成就，造诣

• Despite today’s limitations, AI is already profoundly impacting society, businesses and individuals and is expected to exert a growing influence on how people live, work and interact with one another. As with all major technological shifts, AI is being idolized and demonized simultaneously. All sorts of existential threats potentially posed by AI are being devised, ranging from robots increasingly appropriating jobs to AI-powered machines fighting specifically against humans. Setting such gloomy scenarios aside, it is nevertheless undeniable that new ethical and societal challenges are emerging concomitantly with innovative AI developments. Businesses, governments, regulators and society as a whole, will have to address such issues to ensure that AI truly benefits all of humanity. In this context, technical standards and conformity assessment systems could play a critical role in shaping the future of AI.

• 虽然今天，人工智能还有很多的限制，但是Ai已经深刻地影响了社会，商业和个人生活。同时他对人们的生活方式，工作方式以及交往方式所施加的影响也越来越大。和所有重大的科技转变一样，AI也逐渐正在被神化和妖魔化。人们想出各种各样AI的潜在威胁，从AI抢占人类的工作到AI赋能的专门和人类进行作战的机器人。且不论这些可怕的假设场景，不可否认的是，伴随着AI的创新发展，新的道德问题和社会挑战不在浮现。商业，政府，监管机构和社会作为一个整体，都必须去解决这个问题，确保AI切实对所有人类有利。在这个情况下下，技术标准和合格性评估系统，在塑造AI的未来中扮演者一个关键的角色。

• 生词：

  • profoundly：深刻地，极度的
  • exert：运用，施加影响
  • idolized：极度崇拜，把什么当作偶像
  • demonized：妖魔化，使成为妖魔
  • devise： 设计，发明；<法律>（通过遗嘱）遗赠（不动产）；<古>图谋
  • ethical：道德的，伦理的；合乎道德的；有益于人（类）的，环保的；（药品）凭处方出售的
  • set aside：留出；驳回，撤销；不顾
  • conformity assessment systems：合格性评估系统

1.2 From winter to rebirth of artificial intelligence

• AI does not constitute a new scientific discipline, as its origins can be traced back to the 1950s. The literature typically identifies three historical phases of AI development.In the first phase (1950s to 1980s), AI emerged from the abstract mathematical reasoning of programmable digital computers. The famous computer pioneer Alan Turing conceptualized the first test to decide whether a programme could be considered intelligent or not, the so-called Turing Test [1]. The term “artificial intelligence” was actually initially crafted by John McCarthy in 1955, who later became known as one of the fathers of AI [2], and was proposed as the subject title for the first conference on AI, which took place in Dartmouth College in 1956.

• AI并没有构成一个新的科学学科，因为他的起源可以追溯到20世纪，50年代。学术界通常将AI的发展分成三个历史性阶段。第一个阶段是1950-1980，AI起源于可编程数字电脑的抽象数学推理。著名的计算机先驱阿兰·图灵设计了第一个可以去判定一个程序是否可以被认为是人工智能的测试，也就是所谓图灵测试。“人工智能”这个词最早是由John McCarthy在1955年发明的，在这之后，他也被称为人工智能之父，同时人工智能这个词也作为第一次AI大会的主要题目，这次大会发生在1956年的达特茅斯学院。

• 生词：

  • constitute：构成，组成，设立，任命
  • conceptualize：使概念化，概念化
  • scientific discipline：科学学科
  • craft：v精心制作；n手艺，小玩具
    

• An important next step in the development of AI was the invention of an algorithm using the concept of neural networks (the “perceptron”) by Frank Rosenblatt in 1958 [3]. However, it was not until 1967, with the development of the nearest neighbour algorithm by Cover and Hart [4], that machine learning started to be used in real applications. In spite of these early achievements and the rapid development of computer-based symbolism, the reach of AI remained nevertheless limited due to the inability to formally express or represent many concepts.In the second phase (1980s to late 1990s), expert systems developed rapidly and significant breakthroughs in mathematical modelling were achieved. ANNs also started to be deployed more widely across a growing number of applications. During that period, some of the core techniques and algorithms of AI were developed and further refined: explanation-based learning (EBL) in 1981 [5], the backpropagation algorithm in 1986 [6], and the principle of the support vector machine (SVM) in 1995 [7].

• Ai发展的下一个关键阶段，是1958年Frank Rosenblatt提出使用神经网络的定义的算法的出现。然而，随着Cover和Hart的最近邻算法的发展，知道1967年，机器学习才开始在实际应用中使用。除了这些早期的成就和计算符号主义的快速发展，AI的适用范围仍旧很受限制，因为无法正式表示和表达很多概念。在第二阶段（从1980s到1990s），专家系统快速发展并且数学模型上有很多重大的突破。人工神经网络也开始应用在越来越多的应用中。在那个阶段，AI的一些核心算法和技术被提出并且进一步改良：比如说1985年的基于解释的学习（EBL）,1986年的反向传播算法和1995年的支持向量机的原理。

• 生词：

  • invention：n. 发明物；发明，创造；虚构，编造；发明才能，创造力；创意曲
  • symbolism：n. 象征，象征主义；符号论；记号

• 理解：第二阶段是以神经网络算法的出现而开始的，但是应用并不广泛。主要的作用是提出了很多新的AI技术和算法。

• One of the best-known milestones during this second phase was the Deep Blue chess programme developed by IBM in 1996, which managed to beat the world champion the following year [8]. This was the first time a computer programme was able to defeat human players in games at the world championship level. In spite of this success, limitations related to knowledge acquisition and reasoning ability, combined with the high cost of deployed AI systems, produced a certain level of disenchantment, which led some observers to speak of an “AI winter”.

• 在第二阶段，最为人所知的里程碑事件就是在1986年由IBM开发的深蓝国际象棋程序，成功打败了第二年的世界冠军。这是第一次，计算机程序在游戏中有能力打败世界冠军级的人类玩家。虽然有这个成功，但是因为AI有关知识获取和推理能力的限制，同时还伴随着部署AI系统的十分高昂的代价，使得众人对AI有了一定程度的清醒，这段时间被一些观察者称为“AI寒冬”。

• 生词：

  • milestone：里程碑，时间表（注意，不是corner stone）
  • disenchantment：n. 醒悟，清醒；不抱幻想

• 理解：推理能力和知识获取能力的限制，以及部署AI的高昂代价使得AI发展一度陷入了停滞。

• It was not until the third phase of development, which started at the beginning of the 21st century, that AI began to deliver on its initial promises. In 2006, the first powerful fast learning deep belief network was introduced in a paper by Hinton, Osindero and Teh [9]. The algorithm was used to recognize and classify numbers in a set of images. This contribution was instrumental to the development of AI and became one of the most influential works for today’s AI research. More recent developments such as IBM Watson in 2010 and AlphaGo in 2016 have since received considerable public attention.

• 直到AI发展的第三阶段，也就是21世纪的早期，AI才真正实现最初的愿景。在2006年，Hinton，Osindero和Teh介绍了第一个强力快速学习置信网络。这个算法被用来对图片集进行识别和分类。这篇出版物对于AI的发展有很大的促进作用，同时也变成了今天AI研究的最有影响力的作品。更近期的发展，比如说2010年IBM的沃森和2016年的阿尔法狗已经受到了大量的公众的关注。

• 生词：

  • deliver on：履行
  • instrumental：adj. 有帮助的，起作用的；用乐器演奏的；与乐器有关的；与（事物）的工具效用有关的；与仪器或测量器具有关的；（语法）工具格的
    

• With the explosion of collected data, sustained innovation in theoretical algorithms and the continuing rise in computing power, AI has subsequently made breakthrough progress in many application fields and now looks well prepared to take on new challenges. All of these developments have led some analysts to speak of a “rebirth of AI”.Figure 1-1 depicts some of the major milestones of AI from its early days until the present time.The success of machine learning algorithms for speech and image recognition was instrumental in attracting considerable interest from the research community, businesses and governments.

• 因为数据采集的大量爆炸，理论算法的连续性创新和计算机算力的持续性提高，AI在很多应用领域已经获得了连续性的突破，而今已经时刻准备好应对新的挑战。正式因为所有的这些发展，一些分析家也将这个时间段称为“AI的重生”。Figure1-1，就描述了AI从早期发展到当前阶段的一些主要的里程碑。而关于语言和图片识别的机器学习的成功，有助于吸引的研究领域、商业和政府的兴趣。

• 生词：

  • analyst：n. 分析家
  • collected data：采集的数据

• Additionally, parallel developments in cloud computing and big data provided the support for moving from computer-based AI simulations to more complex and intelligent systems connecting machines and people. It is now foreseen that AI will become one of the core technologies of the fourth industrial revolution, as well as a driving force for innovation in transportation, healthcare, retail, education, government services and other industries.

• 除此之外，同时发展的云计算和大数据使得以计算机为基础的AI模拟迁移到与人互联的的更加复杂和智能的系统上成为可能。可以预见，在第四次工业革命中，Ai将会变成核心技术，同时将作为交通、健康、零售、教育、政府服务和其他产业的创新驱动力。

• 生词：

  • simulations ：n. [计]模拟（simulation 的复数）；[计]仿真
  • parallel development：同时开发，平行演化，通俗的就是同时发展的
  • transportation：n. 运输，运送；交通运输系统；运输工具，交通方式；（犯人的）流放，放逐
  • retail：零售

1.3

• AI represents a huge market potential. According to a recent study from the International Data Corporation (IDC), worldwide spending on cognitive and AI systems is forecast to exceed USD 57 billion in 2021 [10]. The retail and banking sectors are expected to spend the most on AI in the coming years, followed by discrete manufacturing, healthcare and process automation. These five industries, still according to IDC, will continue to be the largest consumers of AI technology, with their combined investments representing nearly 55% of all worldwide spending on such technology by 2021.

• AI表现出巨大的市场潜力。根据国际数据通讯公司的最近的一份研究，全球在认知和AI系统上的花费在2021年有望超过五百七十亿。在未来的一年中，零售和银行板块将会是的在AI板块花费的最多的，紧接着就是零散的手工业，健康业和流程自动化。仍旧根据IDC的信息，这五个行业仍将是人工智能技术的最大消费者，到2021年，它们的投资总额将占到全球人工智能技术支出的近55%。

• 生词：

  • cognitive：adj. 认识的，认知的
  • discrete：离散的，个别的

• Taking into account the related service industry of machine intelligence, which includes programme management, education, training, hardware installation, system integration and consulting, the market size is actually much larger and AI is foreseen to become one of the fastest growing industries in the near future.

• 考虑到机器智能的相关服务产业，这主要包括项目管理、教育、训练、硬件安装、系统集成和咨询，实际上市场的规模将进一步扩大并且在未来，AI有望变成发展最快的产业之一。

• 生词：

  • programme management：项目管理
  • system integration and consulting ：系统集成和咨询

• While automated customer service and diagnostic systems will likely remain the top drivers of AI spending in the coming years, smart manufacturing is expected to take a strong position in the AI market. IDC actually sees intelligent process automation become the third largest use case of AI systems by 2021 [10]. Other use cases that will experience fast spending growth include public safety, emergency response, and shopping advisors and recommendations。

• 尽管自动化客户服务和诊断系统在未来很可能仍旧是AI未来支出的主要驱动力，但是智能制造在未来的AI市场中仍旧有可能占据强势地位。IDC实际上预计，在2021年底，智能流程自动化将会成为AI系统的第三大应用案例。别的应用，比如说公共安全，应急响应和购物顾问和推荐等，都将经历快速的投入增长。

• 生词：

  • automated customer service：自动化客户服务
  • diagnostic systems：诊断系统
  • precess automation：工程自动化
  • smart manufacturing：智能制造
  • IDC sees：IDC预计，注意这里sees值的是预计
  • emergency response：应急响应
  • shopping advisors and recommendations：购物顾问和推荐

• 未来AI发展较快的领域的，按照由大到小的顺序依次是，自动客户服务和诊断系统，智能制造，智能程序自动化，公共安全，应急响应，购物推荐和顾问

• Inevitably these exciting market prospects will also carry a certain number of risks and challenges. The impact of AI on the workforce is frequently cited as a potential threat to societies, with tensions in social relations resulting from a gradual diversification of the employment market. Increased automation and connectivity could also lead to additional or intensified wealth gaps between developed and developing economies. However uncertain such scenarios appear today, all major economies throughout the world have started to invest heavily to support AI innovations as part of their strategic technology planning activities. For instance, in 2017 China promulgated the “New Generation of AI Development Plan”, the “Three Year Action Plan for the Promotion of New Generation of AI Industry (2018-2020)” as well as various other policies to accelerate research, development and industrialization of AI technology.

• 不可避免的是，这些激动人心的市场前景也会带来一定数量的风险和挑战。随着就业市场逐渐多样化，社会关系变得越来越紧张，AI对于劳动力的影响也常被认为是社会的潜在威胁。自动化和连结度的增加也可能会加剧发达经济体和发展中经济体财富差距。（逻辑应该这种情形还没发生，故而是从当前看来）无论这种情形在今天看起来有多不确定，全球所有主要的经济体已经将加大对AI创新的投资作为国家战略科技计划活动的一部分。比如说，在2017年，中国发布了《新一代人工智能发展规划》、《推广新一代人工智能产业发展的三年行动计划（2018-2020）》，同时还有很多别的不同政策，借此来加快AI技术的究、发展和产业化。

• 生词：

  • diversification：n. 新产品的开发，多样化经营
  • a gradual diversification ：逐渐多样化
  • connectivity：n. 连通性，连接度
  • promulgate：v. 发布，公布；宣传；散布
  • market prospects：市场前景
  • employment market：就业市场
  • strategic technology planning activities：战略技术规划活动

1.4

• There are several ways of defining AI. ISO/IEC Joint Technical Committee (JTC) 1 refers to “an interdisciplinary field, usually regarded as a branch of computer science, dealing with models and systems for the performance of functions generally associated with human intelligence, such as reasoning and learning” [11]. In the IEC White Paper on edge intelligence [12], the term AI is applied when “a machine mimics cognitive functions that humans associate with other human minds, such as pattern matching, learning, and problem solving”.

• 定义AI的方式有很多。 ISO/IEC联合科技委员会（JTC）将AI定义为“一个跨学科领域，一般长指计算机学的分支，主要是和人类智能有关的功能的模型和系统，如推理和学习”。在IEC的白皮书关于边缘智能的部分中，AI一词主要是在“机器模仿人类与其他人类大脑相互动的认知功能，比如说模式匹配、学习和问题解决”时才被使用。

• 生词：

  • interdisciplinary：跨学科的
  • a machine mimics cognitive：机器模仿认知

• AI主要指的是和人类相关认知功能，比如说学习和推理。

• In other words, intelligence is demonstrated by four basic capabilities: sensing, comprehending, acting and learning. As of today, comprehending does not have the same meaning for machines as for humans. Typically, a model will be trained to “learn” how to perform better compared to more conventional methods, but AI systems cannot claim yet to “comprehend” the world around them.

• 换言之，智能常常指的是四种基本的能力：感知力，理解力，行动力和学习力。截至今天，理解对于人类和机器而言是完全不同的意义。一般来说，即使我们训练一个模型，通过学习让他比传统的方法体现的性能更加优越，AI系统也不能对外宣称已经能够理解周围的世界。

• 生词：

  • as of today：截至今天
  • conventional method ：常规方法

• 智能通常指的是：感知力，理解力，行为力和学习力。但是理解力对于AI和人来说是两种不同含义。

• Practitioners of AI often distinguish between strong AI and weak AI. Strong AI (also called general AI) refers to the more philosophical concept of a machine capable of exactly imitating human intelligence. Such a machine would be able to solve any problem in any field requiring advanced cognitive abilities. This kind of AI has not been developed yet and can only be found in various science fiction books or movies

• AI的实践者常常将AI区别为强AI和弱AI。强AI（通用人工智能）常指的是更加哲学的定义，即能够准确模仿人工智能机器能力。这种机器能够解决需要高级认知能力的领域中的任何问题。这种人工智能现今为止并没有被开发出来，而且也仅仅能够在不同的科幻小说或者电影中找到。

• 生词：

  • Practitioners：实业者，开拓者，实践者
  • cognitive：认知的，认识的

• In contrast, weak AI (also called narrow AI) supports humans in solving specific problems for a particular use case. For example, AlphaGo masters the board game Go to an almost perfect degree but is unable to solve any other problem. Speech recognition tools such as Siri represent a kind of hybrid intelligence, which combines different weak AIs. These tools have the ability to translate spoken language and connect words to their databases in order to perform different tasks. Nevertheless, such systems do not constitute any general form of intelligence [13].

• 相反的，弱人工智能（通常也称为狭义人工智能）能够帮助人类解决特定应用案例中的特定问题。比如说，阿尔法狗精通围棋甚至已经到了完美的地步，但是他却不能解决其他任何问题。语音识别工具，不如说siri，代表的就是一种混合智能，他混合了不同的弱人工智能。这些工具已经有能力翻译人类说出来的话，并且将这些话和他们的数据库联系起来，从而能够执行不同的人物。但是无奈论如何，这种人工智能并没有构成任何一种通用形式的人工智能。

• 生词：

  • the broad game Go:围棋
  • hybrid intelligence：混合智能
  • hybrid：混合的

• Other terms are tightly connected to AI, such as machine learning and deep learning. To create intelligent machines, a specific kind of knowledge is needed. In the past, such knowledge was hardcoded directly into the machines, which led to certain restrictions and limitations. The approach taken by machine learning is that the machine builds up its knowledge itself, based on a given set of data [14].

• 还有很多和AI仅仅相关的短语，比如说机器学习和深度学习。为了创建更加智能的机器，往往是需要特定的领域的知识。在以前，这类的相关知识是很难直接被编码进机器，这就导致了发展受限和瓶颈。机器学习采取的方式就是机器基于给定的数据创建自己的知识。

• 生词：

  • 机器学习的策略是自己创建和生成知识。

• As knowledge these days comes mostly from real-world data, the performance of machine learning algorithms highly correlates with the information available, also called representation. A representation consists of all the features that are available to a given machine (e.g. output of a temperature or vibration sensor in a predictive maintenance application). Selecting the right representation is a complex and time-consuming task, requiring highly specialized domain knowledge.

• 因为目前知识仅仅来自于真实世界，机器学习算法的性能，和可获得的信息 ，也就是表达是高度相关的。表达是由很多特征构成，这些特征是机器学习可以理解的（比如说在预测维修中，振动传感器和温度计的输出）。在一个复杂并且耗时的任务中选择正确的表达，需要相关领域丰富的知识。

• 生词：

  • correlate：相关联，紧密相连

• A field of machine learning called representation learning automates this task by discovering the representations needed for feature detection or classification from raw data. This set of methods is based on learning data representations, as opposed to more traditional task-specific algorithms [14].

• 机器学习其中一个领域被称为表示学习，它通过从原始数据中挖掘特征检测或者分类所需要的表示，从而使得这个任务能够自动化。这一类方法是基于学习数据表示的，他和更加传统的针对特定任务的算法是相反的。

• representation learning通过挖掘特定的数据表示来进行学习的，是基于学习的数据的。

• However, selecting the right features is usually a very difficult operation, since they are dependent on various environmental factors. For instance, colours will be perceived differently in a dark environment, which then can impact the silhouette of objects.

• 然而，选择正确的特征往往是一个十分困难的操作。因为他们依赖于不同的环境因素。比如说，在比较黑暗的环境下，被感知到的颜色是不同的，这就会影响目标物体的轮廓。

• 生词：

  • silhouette：n. （浅色背景衬托出的）暗色轮廓；剪影，（尤指人脸的）侧影；（人的）体形，（事物的）形状
  • perceive：vt. 认为，理解；察觉，注意到；意识到

• As a subcategory of representation learning, deep learning transforms features and elaborates dependencies based on inputs received. In the example of an image, the input features are the pixels. A deep learning approach will map the pixels first to the edges of the image, then to the corners and finally to the contours to identify an object [14]。

• 作为表示学习的一个子类，深度学习基于输入的数据，转换特征和产生相关依赖。以一张图片为例，输入的特征就是像素。深度学习的方法首先会将像素映射到图片的边界，然后再映射到角落，最后映射到轮廓，从而识别物体.

• 生词：

  • elaborate：
    * adj. 复杂的，详尽的；精心制作的
    * v. 详细说明，详尽阐述；精心制作
  • contours：n. 轮廓等高线

• 深度学习仅仅是表示学习的一个分支，通过映射实现对于目标的识别。

1.5

• The objective of the present White Paper is to provide an overview of where AI stands today as well as some forward thinking for the next decade, by exploring opportunities and challenges of AI for a number of application domains. Building upon several use cases, the White Paper introduces a number of recommendations, some of them targeted at the IEC and its committees for future standardization and conformity assessment work.This White Paper primarily focuses on weak AI. It is clear that today the world is running on various forms of weak AI. An email spam filter is a classic type of weak AI. It starts off loaded with a certain level of intelligence that enables it to figure out what constitutes spam and what does not, and then refines its intelligence as it acquires experience with the particular preferences of individual or collective users. Automobiles today are replete with weak AI systems, from the computer that determines when the anti-lock brakes should be activated to the processing unit that tunes the fuel injection parameters. Self-driving cars that are currently being tested include numerous robust weak AI systems allowing the vehicle to sense and react to the surrounding environment.
• 当前这个白皮书的目的就是通过探索AI很多应用领域的机遇和挑战，从而提供一个对AI现状的概述，同时还有下一个十年的进一步思考展望。通过几个使用样例，白皮书提出了一些建议，其中一些是针对IEC以及负责未来标准制定和合格评测工作相关委员会的/这个白皮书主要是针对弱AI的。可以看到的是，而今世界上正运行着很多弱AI。垃圾邮件过滤器就是经典的弱AI。一开始，他加载一定程度的智能，这样它能够弄清楚什么被视为垃圾邮件，什么不应该被视为垃圾邮件，然后当它按照个人或者集体用户的特定偏好，处理一段时间的邮件之后，他就能在一定程度上精炼自己的智能。而今的汽车都安装了弱人工系统，从电脑决定防抱死刹车开启时机，到处理单元调整燃油喷射参数。当前正在被测试的自动驾驶汽车就包含大量的稳定的弱人工智能，继而使得交通工具能够感知外界环境并对其作出反应。
• 生词
  • robust：adj，强健的，坚固的，稳固的
  • tune the fuel injection parameters:调整燃油喷射参数
  • anti-lock brake：防抱死刹车
  • automobiles：汽车，汽车相关的
  • conformity：n. 遵守，依照；随大流，循规蹈矩；符合，一致；信奉英国国教
  • spam：n 垃圾邮件
  • constitutes：v. 组成，构成；是，被视为；成立，设立；任命（注意，这里是有被动的含义）
  • collective users：集体用户
  • replete：v 充满，装满 be replete with 装满
• Creating an AI as smart as the human brain will remain an enormous challenge for quite some time. Building a computer that can multiply two ten-digit numbers in an infinitesimal amount of time is unexpectedly easy. Building one that can look at a dog and decide whether it is actually a dog or a cat is much more difficult. While creating a system that can defeat the world chess champion has already been achieved, fabricating an AI that can understand the meaning of a paragraph from a six-year old’s picture book, and not just recognize the words, is still well beyond today’s possibilities.
• 创造一个和人脑一样聪明的AI仍旧是一个需要很多时间的痛苦巨大挑战。创造一个能够在无效小的时间里处理两个十进制数电脑是难以想象地容易。建造一个能够看见狗并且决定它到底是一只狗还是一直猫却并困难许多。虽然发明一个能够打败世界国际象棋冠军的系统已经实现了，但是建造一个的能够的理解六岁小孩的图画书的段落含义，不仅仅是识别文字，仍旧是远远超出了今天所能够实现的可能性。
• 生词：
  • unexpectedly：知道什么意思，就是找不到合适的词，难以想象地
  • fabricate：v 制造，生产，捏造，组合
  • infinitesimal：无限小，无穷小
  • enormous：adj，巨大的，极大的，痛苦的
• Many such capacities that seem easy to human beings are actually extremely complicated and only seem easy because the skills involved have been optimized in humans (and most animals) across hundreds of millions of years of evolution. Since it is virtually impossible to properly define what intelligence consists of, it is very difficult to provide a clear criterion as to what would count as a success in the development of strong AI.
• 有很多看上去似乎对人类很容易的能力，仅仅是看上去容易，实际上极度的复杂，因为这个能力已经在人类经过上百万年的进化时间内被不断的优化。因为准确定义智能的组成是几乎不可能的，所以在开发强人工智能的过程中，很难去定义一个清晰的准则，去判定什么是成功。
• In order to shed light on some of the foreseen developments of AI within the next decade, the following application domains are explored in the White Paper, and for each of them a number of use cases are described：
• 为了清楚地显出未来十年AI的可能发展，白皮书主要探索了如下几个应用领域，并且每一个领域都描述了一定的使用样例
• 生词：
  • shed light on：阐明，使清楚的显出
• § Smart homes – residential environments that are equipped with connected products for controlling, automating and optimizing functions such as temperature, lighting, security, safety or entertainment.
• § 智能家庭-住宅区被配备了用于控制，自动化和优化功能的连接设备，主要是针对诸如温度、灯光、安全或者娱乐等功能。
• 生词：
  • residential ：adj. 住宅区的，居民区的 ；家庭的，住宅的；（为工作、学习或受人照料而）住宿（在某地）的
• § Smart manufacturing – a technology-driven approach using connected machinery to monitor production processes and data analytics to improve manufacturing operations.
• § 智能生产—— 这是一种使用了链接机器人的技术驱动方式，主要是为了监测生产过程和数据分析，从而实现对生产的改良。
• § Smart transportation – a scenario in which mobility-related entities (e.g. personal vehicles, public transportation, delivery vehicles, emergency services, parking) are integrated and automated in order to improve traffic performance and reduce potential negative impacts such as congestions, accidents or pollution.
• § 智能交通——将移动相关实体（个人交通工具，公共交通工具，运输工具，能源服务和停车）进行集成和自动化，从而改良交通性能并且减少潜在消极影响，比如说堵塞，事故或者污染。
• 生词：
  • mobility-related：移动相关的
  • congestion：堵塞，拥塞
• § Smart energy – a sustainable and costeffective energy system in which renewable production, consumption, and infrastructures are integrated and coordinated through energy services, active users and enabling information and communication technologies (ICTs).
• § 通过能源服务、活跃用户和信息通信技术(ict)，将可再生能源的生产、消费和基础设施整合和协调起来的可持续和具有成本效益的能源系统。
• 生词
  • cost-effective：划算的，具有成本效益的

1.6

• Section 2 describes the need for AI through identification of several key megatrends posing major challenges for societies, businesses and individuals. AI will enable and enhance a wide range of applications addressing some of these challenges, such as environmental concerns, changing demographic trends or economic disparity, to mention only a few.

• 第二部分通过指出对社会，商业和个人构成重大挑战的几个关键大趋势，来体现AI的主要需求。AI将会是的一系列广泛的应用成为可能并不断进行改良，比如环境问题的领域、不断变化的人口趋势或经济差距，这仅仅是其中的一小部分。

• 生词：

  • megatrends：n. 大趋势，向信息社会发展的趋势
  • demographic：adj. 人口的，人口统计 的 n. （具有某种特征的）群体；人口统计数据
  • disparity：n 明显差异
  • to mention only a few:这仅仅是其中一小部分

• Although AI is hardly a new discipline, it was not until 2010 and later that dramatic technological enhancements, in particular in the area of machine learning, paved the way for today’s explosion of AI. This breakthrough was enabled thanks to a number of factors explained in Section 3. Significant improvements in computational power, more sophisticated machine learning algorithms and the availability of large amounts of data to train AI systems have been the primary enablers of today’s spectacular AI developments.

• 虽然AI并不算是一个新得学科，但是它也是直到2010年以后，极大的科技提高，尤其是机器学习领域的技术进步，才为今天AI技术的爆炸铺平了道路。这次突破的实现要归功于第三节中解释的一些因素。算力的显著提高，更加精密的机器学习算法和训练人工智能系统的大量数据的可用性，一直是当今人工智能惊人发展的主要推动因素。

• 生词：

  • dramatic technological enhancement：科技的极大进步
  • thank to：正式因为，归功于。。（不要老是翻译成因为）
  • significant improvement：显著的提高
  • sophisticated：adj. 见多识广的，老练的，见过世面的；复杂巧妙的，先进的，精密的；水平高的，在行的；时尚的，精致的
  • availability of large amounts of data to train AI system：训练人工智能系统的大量数据的可用性

• A number of additional drivers that have also contributed to the flourishing field of AI research are further outlined in Section 3. These include information technology (IT) developments such as cloud and edge computing, the Internet of Things (IoT), and big data, as well as the increasing readiness of consumers and society to embrace new technologies and share data.

• AI研究的活跃领域很多的额外驱动力也将进一步被列在第三部分。其中包括信息技术发展，比如说云和边缘计算，物联网，和大数据，同时消费者和社会越来越愿意去接受新技术和共享数据。

• 生词：

  • flourishing field：繁荣领域
  • outline：v. 概述，略述 勾勒，描画。。。的轮廓
  • readiness：n. 准备就绪的状态，愿意，乐意，快

• 第三部分的主要描述人工只能领域发生技术爆炸的所有的驱动力

• Section 4 provides a high-level understanding of the most common AI systems and machine learning techniques. Without entering into deep technical detail, it also reviews the most popular AI algorithms in use today that constitute the foundation for tomorrow’s AI developments. Based on the current state of the art of AI, this technical overview is complemented by many references for readers wishing to consolidate their scientific understanding of how AI actually works from the inside.

• 第四部分主要讲述最流行的AI系统和机器学习技术的深层次的理解。在没有体积深度学习的技术细节的情况下，也对今天正在使用的AI算法进行了回顾，而这些技术也将构成未来AI发展基石。基于当前AI研究的现状，这个技术概览补充了很多的参考文献，以帮助希望巩固对人工智能内部工作原理的科学理解的读者。

• 注意：

  • for reader wishing to consolidate their scientific understanding:这里的wishing后面的补语，是修饰读者的，应该指得是什么样的读者。for readers：就是针对对应的读者，或者说帮助对应的读者

• 生词：

  • complement：v 补充，补足 n 补充物，衬托物
  • consolidate：v 使巩固，使加强，合并，统一
  • reference：n. 参考文献
  • Based on : 基于的，不要老是翻译成基于什么现状
  • the current state of the art of AI:AI的当前现状

• 第四部分主要是针对AI技术的深层次理解，在不谈论具体的技术细节的情况下，对AI当前最新的技术进行回顾

• While this White Paper cannot cover all of the possible AI application scenarios, a representation describing how today’s main AI systems map to some of the most popular application domains is provided in Section 5. This exercise furnishes a better characterization of the AI needs and requirements of several industry sectors. The rest of the section is then devoted to a more detailed description of the four application domains (smart homes, smart manufacturing, smart transportation, and smart energy), for which several AI-related use cases and scenarios are reviewed, together with some of the most pressing challenges of current and emerging AI implementations.

• 虽然这个白皮书并没有覆盖所有可能的AI应用场景，但是第五部分也讲了今天主要的AI系统是如何应用到一些最流行的应用领域。这一实践更好地描述了几个行业部门的人工智能需求。这个部分的剩余部门将会专注于四个应用领域的更加细致的描述（智能手机，智能制造，智能交通，和智能能源）。将对其中人工智能用例和场景进行综述，并结合一些当前和新兴的人工智能的实现所带来的最具有挑战性的挑战。

• 注意：

  • a representation describing … is provided in section 5:第五部分描述了，可以换一种表达方式
  • 注意几个固定的表答
    • furnish a better characterization：更好的描述
    • needs and requirements：需求

• 生词：

  • representation：n. 代理人，代表；描绘，表现；表示，象征；表现物（尤指画或模型）；<正式>抗议，陈述；（感知理论用语）表象
  • furnish：v. 为（房间或房屋）配备家具；提供，供应
  • furnish a better characterization of :更好地描述了。。。。
  • industry sector：行业部门
  • pressing challenges:紧迫的挑战
  • current and emerging AI implements:当前和新兴的AI应用

• Following this review, Section 6 consolidates the main AI challenges that can be identified in today’s implementations or foreseen in emerging AI developments. Challenges are grouped into several categories: social and economic challenges; data-related challenges, including the selection of training data and the standardization of data; algorithm-related challenges, such as algorithm robustness and interpretability; infrastructure challenges, related to hardware or platforms; trustworthiness issues, including trust, privacy and security; and challenges linked to regulations, such as liability and ethical issues.

• 再次回顾之后，第六部分整合了在当今实现中可以确定的或在新兴的AI发展中可以预见的主要AI挑战。这戏挑战被分为几个不同的类别：社会和经济挑战；数据相关的挑战，包括训练数据的选择和数据的标准化；算法相关的挑战，比如说算法的鲁棒性和可解释性；基础设施挑战，和硬件或者平台相关；诚信问题，包括信任，个人隐私和安全；和监督相关的挑战，比如说可靠性和道德问题。

• 注意：

  • the main AI challenges that can be identified … or forseen…这里是两个定语从句，并不是一个定语从句

• 生词

  • robustness：n 鲁棒性
  • interpretability：n 可解释性，解释能力
  • trustworthiness issues：诚信问题
  • ethical：adj 道德问题

• Building upon the previous section, Section 7 develops a standardization landscape for AI and identifies a number of gaps that need to be addressed in order to solve some of the AI challenges described previously. While today’s standardization activities are still at a very early stage, the White Paper clearly demonstrates that standards need to play an essential role in shaping the future of AI and mitigating the many technical and non-technical issues emerging with the deployment of AI across industries.

• 在前几个部分的基础上，第七部分将会射击一个标准的远景并且辨别出一系列需要解决的隔阂，为了解决之前提出的AI挑战，我们需要去解决这些问题。与此同时，今天的标准化行动，仍旧处于一个比较早期的水平，白皮书清晰地表明，标准需要在塑造人工智能的未来和缓解人工智能跨行业部署而出现许多技术或者非技术问题方面发挥重要的作用。

• 注意：

  • play an essential role in shaping … and mitigating … 是在这两个动名词中发挥重要的作用

• 生词：

  • mitigate：v 减轻，缓和
  • AI across industries：人工智能狂行业部署
  • play an essential role:发挥重要的作用

• Finally, Section 8 concludes the White Paper by offering a series of recommendations for industry, regulatory bodies and the IEC. While the success of AI and its acceptance within diverse societies will rely on the involvement of multiple stakeholders and communities, it is clear from this White Paper that industry, policy makers and standards development organizations such as the IEC will need to play a driving role to ensure that AI delivers on its promises.

• 最后，第八部分总结了白皮书，主要是通过给产业界，监管结构和IEC提供一些建议。虽然AI的成功以及能够被社会广泛接受主要依靠多方控股人和社区的参与，但是白皮书中清晰的指出产业界、政策制定者和标准制定结构比如说IEC，在确保AI能够实现个人愿景的基础上发挥驱动作用

• 生词

  • regulatory bodies：监管机构

• Additional, future-looking AI developments that may grow in importance over the next decade are
  discussed in Annex A.

• 除此之外，在未来十年的人工智能发展将会变得越来越重要，这将在附件 A中进行讨论

Section 2 Need for artificial intelligence

2.0

• Today’s society and business landscape are characterized by a complex and unprecedented set of challenges and opportunities. Existing markets are subject to disruption and can even disappear abruptly in a short space of time. Major global trends impacting society, the economy, business, cultures and personal lives, often called megatrends, are defining the future world of mankind and its increasing pace of change.
• 今天的社会和商业现状特点是一系列复杂和不可预测的挑战和机遇。现存的市场有可能会中断甚至会有可能在很短的时间内消失。影响着社会，经济，商业，文化和个人生活的主要全球趋势，通常被称为大趋势，正定义着人类未来的世界和日益加快的变化步伐。
• 注意
  • be characterized by ：特点是，不要翻译成被塑造成
• 生词：
  • be subject to disruption：受到破坏。be subject to受到
  • magatrends：大趋势
• Megatrends represent interconnected, global interactions that contribute to framing the impact of major technology developments such as AI. The joint effect of digitization, automation and AI is expected to significantly impact the future of work. It is anticipated that computerization will affect many low-skill jobs, with computer-guided automation becoming increasingly prevalent across numerous industries and environments, including manufacturing, planning and decision-making [15]. The growth in technological capabilities is already transforming supply chains, reshaping the workforce and redefining jobs. The challenging prospect of such change lies in the fact that the growth is not linear but rather complex and accelerating.
• 大趋势表示的是相互关联的全球联动，有助于构建人工智能等重大技术发展的影响。数码化，自动化和AI的联合效应会深远地影响工作的未来。可以预料的是，在计算机辅助自动化在包括制造，计划和决策，等大部分行业和环境中变得越来越流行，计算机化将会影响很多低技术的工作。科技能力的提高已经正在改变供应连，重塑重塑劳动力和重新定义工作。这种变化中比较有挑战性的地方在于，这个变化并不是线性的，而是极其复杂的，并且正在加速。
• 生词：
• interconnected，global interactions ：相互关联的全球联动
• frame the impact of the major technology developments：构建主要技术发展的影响
• joint effect：联合效应
• prevalent：adj 盛行的，普遍的
• At the same time, AI will enable and improve a wide range of applications that can address some of the challenges emerging from these megatrends: environmental concerns, changing demographics, or economic disparity, to mention only a few.
• 与此同时，AI将会使很多应用成为可能并不断进行改良，这些应用程序能够解决这些大趋势所带来的挑战：环境问题，人口统计的变化，或者经济差异，这里仅仅是其中一小部分。
• 生词：
  • demographic：adj. 人口的，人口统计的

2. Scarcity of natural resources 自然资源的缺乏和不足

• The planet’s natural resources are being consumed at an alarming rate and most countries are expected to double their annual global consumption of such resources by 2050 [16]. Not only are finite natural resources being depleted, humans are also using far more environmental resources than Nature can regenerate. While in the past resource conservation was often viewed as detrimental to business, today the two are by no means mutually exclusive.
• 星球的自然资源正在以一个令人警醒的速度进行消耗，而且大部分国家到2050年之前，此类资源的一年消耗都会翻倍。不尽有限的自然资源正在被大量消耗，人类也正在使用远比自然所能够生成限度的环境资源。虽然在过去自然资源保护常常被视为对商业有害，但是今天资源保护和商业已经不再是相互冲突了。
• 生词：
  • annual 和 manual：annual 一年的，manual 手工的
  • deplete：v 大量减少，耗尽
  • resource conservation：资源保护
  • detrimental：adj 有害的，不利的 n 有害的人
  • exclusive：adj. 独有的，专用的；高档的，昂贵的；不含……的；排外的；排斥的；全部的；唯一关心的；（新闻或报道）独家的；（条款）除外责任的；高级的，昂贵的
• AI is already helping countless manufacturers to optimize productions processes, thereby reducing waste and increasing output. In addition, AI will soon be used not just to optimize the processes themselves but also their inputs. By analyzing the purpose, properties and environmental impact of a production’s input materials, AI will be able to help scientists design materials that match the specifications required for more sustainable production. Ideas have even been proposed for using AI to identify a second usage for the material components of by-products created by machinery, thereby creating a near circular use of raw materials. Not only are such efficiency gains in production processes an attractive incentive for businesses, they will also have a significant impact on global resource consumption.
• AI已经帮助无尽的制造商去最优化产品制作过程，从而能够减少浪费和增加输出。除此之外，AI在最优化生产过程中的同时，也会最优化输入。通过分析产品原材料的用途，性质和环境影响，AI将能帮助科学家设计符合可持续生产所需规格的材料。甚至有人提出了使用AI去确定机械副产品的材料组件的第二用途的想法，从而创造了原材料的近似循环使用。在生产过程中生产效率的提高不仅对于企业来说也是有很大吸引力的，而且也对全球资源的消费有很大的影响。
  • 生词：
    • an attractive incentive：一个有吸引力的激励
    • production process：生产过程
    • efficiency gains：效率的提高
    • by-products created by machinery 机械副产品
    • specification：规格
      • match the specification：满足规格
    • comsumption：消费
    • a near circular use for rao materials：原材料的近似循环的使用。
• AI will also help utilities in an era of increasing urbanization, growing power consumption, scarcity of water resources, and large-scale deployment of renewable energy. This will be achieved by more intelligent management of both demand and supply. On the demand side, AI is already producing significant energy savings, for example by reducing the consumption of data centres by 15%. On the supply side, decentralized smart energy grids will be able to predict and pre-empt outages, and manage fluctuations in supply and demand to ensure the optimal level of supply while at the same time minimizing the use of fossil fuels [17] .
• 在一个越来愈城市化，能源消费逐渐增加，水资源更加短缺并且可再生能源大规模部署的年代中，AI将会助理公众事业。这是通过对于供应和需求更加智能的控制来实现的。在需求方，AI已经正在产生显著的能源节省，比如说通过减少15%的数据中心的花费，可以实现节省能源。在供应方，疏散智能电网将能够预测并且防止中断，同时还能控制供应方和需求方的波动，确保供应的最优水平，同时还能够最小化化石燃料的使用。
• 生词
  • era：年代，时代
  • utilities：n. 公用事业，实用工具，应用程序
  • urbanization：n.都市化，城市化
  • decentralize：v 使分散，疏散，使分权
  • pre-empt outages：防止中断
  • fossil：n. 化石；<非正式>思想僵化的人；僵化的事物 adj. 化石的，从地下采掘出的；不变的，古老的
• Additional examples include the optimization of renewable energy generation through solar or wind farms or the optimization of vehicle traffic flows to reduce emissions.
• 别的例子包括通过太阳能发电农场或者风能发电农场以实现可重复能源再生的最优化，还有就是交通工具率的最优化

2.2 Cimate change

• Forecasts by leading scientists have been unequivocal: without a consistent response to the climate change challenge and a more responsible use of environmental resources, unpredictable changes will threaten the planet. This raises the question of how to reconcile economic objectives with environmental sustainability. Equally important is how mankind can prepare for unexpected dramatic natural occurrences in the future.

• 顶尖科学家的预测是十分明确的：如果对于气候变化挑战没有一致的回应，同时对自然资源没有一个的合理地应用，不可预测地变化将会威胁整个星球。这就提出如何让经济目标和自然可持续性问题相互协调的问题上。同等重要的是，人类如何为未来意想不到的戏剧性自然事件做好准备。

• 注意：

  • Equally important也是倒装句，同等重要的是
  • prepare for：是为什么做好准备

• 生词

  • unequivocal：adj 明确的，不含糊的
  • leading scientist：顶尖科学家
  • reconcile：v. 调和，使协调一致；（使）和解，（使）恢复友好关系；调停，调解（争吵）；使顺从于，使接受；核对，查核（账目）

• The use of AI in a wide variety of applications is expected to play a leading role in the fight against climate change. AI can support complex decision-making processes when dealing with natural resources or when predicting unexpected incidents. The consumption and use of resources, for instance, can already be optimally coordinated during energy generation. AI makes it possible to set numerous parameters such as context-based electricity consumption and grid load in relation to weather forecasts and electricity tariffs. As a result, the behaviour of electricity consumers can be determined and addressed more efficiently.

• 在很多不同应用中使用AI，有望在应对气候变化过程中发挥重要作用。当我们在处理自然资源或者预测未曾想到的事故时，AI能够支持复杂的决策过程。比如说，在能源生成过程中，资源的消耗和使用已经被协调的最优化了。AI使得设置大量的参数成为可能，比如说考虑到的天气预报和电价将官的实际电力消耗和电网负荷。因此，更加的有效的决定和解电力消费者的行为。

• 生词：

  • context-based：基于上下文，情境
  • grid load：电网负荷
  • tariff：n 关税
  • electricity tariff：电力价格

• Building on those achievements, intelligent mobility solutions involving a more responsible se of resources can be implemented, including autonomous electromobility. Not only can vehicles and trucks be optimally and efficiently matched to one another, they can also be driven more efficiently thanks to AI.

• 在这些效果上，对于资源利用更加负责的智能移动解决方案也可以被实施，包括电力自治。不仅仅汽车和卡车可以更有效地匹配，同时因为AI他们也更方便被驾驶。

• 生词

  • intelligent mobility solutions：智能移动解决方案
  • se：相互
  • electromobility：电迁移率，电动性

• Similar developments can be devised for efficient water consumption. In agriculture, for example, AI allows to determine the optimal water demand depending on the specific needs of each individual plant, the soil situation and current weather conditions. Furthermore, supply strategies for droughts and water shortages in affected regions and countries can be developed using AI techniques.

• 在水资源得有效消耗上，发展也是相似的。比如说，在农业，AI能够根据每一株植物的特定需求，土壤的情况和当前的气候环境，决定最优的水资源需求配置。进一步说，对于受到干旱和缺水影响的区域和国家，可以使用AI技术来发展供应策略。

• 生词：

  • devise：题铭，铭文；格言，箴言，座右铭；外汇，外币；船名

• AI can also contribute to improving predictions of weather scenarios and natural disasters. Scientists are increasingly faced with the challenge of capturing and processing numerous influencing factors in order to increase the accuracy of weather forecasts. AI will effectively help people process a wide range of measurement data in order to provide early predictions for weatherrelated events and warnings of potential elements such as floods, air pollution episodes, or storms. Early-warning systems can then be set up more intelligently for diverse geographies.

• AI也能帮助预测气候现状和自然灾害。为了增加气候预测的准确度，科学家越来越需要面对采集和处理大量的有影响的数据的挑战。AI将有效帮助人处理大量的测量数据，以便为天气相关事件提供早期预测，并对洪水、空气污染或风暴等潜在因素发出警告。这样就可以为不同的地区更智能地建立预警系统。

• 生词

  • measurement data：测量数据；测定资料；计量资料
  • episode：n. 一段经历，一段时期；（电视剧或广播剧的）集，一集；疾病的发作，发病；插句，插段；（希腊悲剧中）两段合唱之间的部分

2.3 Demographic trends（人口趋势）

• The United Nations (UN) is predicting a population increase of more than one billion by 2030 due to demographic growth in emerging and developing countries (Figure 2-1). In addition, the ageing population (number of persons 65 years or older) will increase by more than 390 million due to people living longer and having fewer children.

• 联合国预测，由于新兴国家和发展中国家的人口增长，到2030年全球人口将增加超过10亿。除此之外，因为人获得更长并且有更少的后代，老龄化人也将增加三千九百亿。

• 生词：

  • a increase of more than one billion：增长10亿以上
  • emerging country 新兴国家
  • the aging population:老龄化人口
    

• As shown in Figure 2-2, the impact of an ageing population will be more immediately felt in Europe, Asia and Latin America, resulting in different regional issues. The trend in the number of working people supporting each elderly person will move from nine in 2015 to a range of a half to four in Asia, which will create a much higher dependency of the older generation on the younger one. In Europe, the availability of a suitable working age population will decrease and create an acute demand for a new generation of workers, i.e. women and the elderly. The ratio of four working age persons per elderly in 2015 will decrease by 50% in 2050. These ageing population trends will undoubtedly generate significant challenges for governments and industry [18].

• 正如t图2-2所展示的一样，老龄化人口的影响将在欧洲，亚洲和拉丁美洲的感受最为直接，同时也将导致不同的地区问题。在亚洲，供养一个老年人口的工作人口的数量从2015年的九个人下降到不到3.5个人，这将导致老一辈对年轻一辈的依赖性更大。在欧洲，适合工作年龄的人口可用度将下降，并且激发了对新一代打工人，比如说女性和老年人的迫切需求。在2015年每一个老年人对应四个适合工作年龄的人口的比率将会在2050年下降到只有一半。毫无疑问的是，老龄化人口趋势将毫无疑问地给政府和各行业造成十分重大的挑战。

• 生词：

  • be more immediately felt in：感受最为直接
  • the number of working people supporting each elderly person：供养每一个老人的劳动人口数量
  • availability：有效性，可用率，高可用性

• Another important aspect of such trends is related to large regional differences in the availability of the
  working age population, such as 1,5 working age people for each elderly person in Japan in 2050 versus 15 working age people for each elderly individual in Nigeria.

• 这种趋势的另外一个重要方面就是在工作人口获得性方面，有比较大的区域性差异，比如说对于每一个日本的老年人来说有1.5个工作人口去支持，与之相比，在尼日利亚，每一个老年人有15个工作人口去支持。

• 生词：

  • large regional differences：比较大的区域性差异
  • versus：prep. （比赛或诉讼中）以……为对手，与……竞争；与……相对，与……相比

• The ageing population trend will drive higher spending on healthcare. It is anticipated that within G7 countries, healthcare budgets will increase by about USD 200 billion every year. However, technological and scientific innovations may help reduce healthcare costs to more affordable levels.

• 老龄化人口趋势将会增加在卫生方面的开支。可以预测的是，在G7国家中，卫生预算每年都将增加两千亿。然而，技术和创新将会有助于减少卫生花销，以达到一个可以支付的水平。

• 生词：

  • budget：n. [财政]预算（budget 复数形式）v. 为……做预算（budget 的第三人称单数形式）

• Most governments in Europe are now encouraging older workers to remain within the workforce by increasing the official retirement age, and by outlawing age discrimination. In addition, industry will need to provide financial incentives as well as retraining programmes for older workers. Lifelong learning to acquire new skills during the individual’s working life, as well as mentoring of younger colleagues, will become critical for efforts to keep ageing people within the workforce.

• 在欧洲大部分政府，通过增加官方退休年龄和禁止年龄歧视，鼓励越来越老的工作者留在劳动力市场。除此之外，各行业也将需要提供经济激励和再就业培训给年龄大的工作者。在个人的工作生涯中，终身学习以获得新的技能，同时作为年轻员工的导师，对于保证老龄化人后留在劳工市场的努力而言是十分关键的。

• 生词：

  • outlaw：
    • n. 不法之徒，逃犯；<史>被剥夺法律权益和保护者
    • v. 使非法，禁止，取缔；（旧时）剥夺（某人的）法律权益
  • remain within：保持在
  • retraining programmes：再就业培训计划
  • lifelong learning：终身学习
  • mentor：导师

2.4 Ecomonic policy（经济政策）

• It has already been highlighted how automation and immense productivity gains will help countries alleviate the pressures of demographic change. While this will certainly be a welcome development for more advanced economies struggling with an ageing population, there are countless opportunities for AI to make a difference for the world’s poorest countries.
• 重点强调的是，自动化和大量提高的产能是如何帮助国家缓解人口变化的压力。对于发达国家而言，这确实会帮助改变老龄化人口问题，同时AI也会有无尽的机会使得世界上最贫穷的国家带来改变。
• 注意：
  • it is highlighted that …人们强调的是
• 生词
  • immense：adj. 极大的，巨大的；非常好，极好；无边无际的
  • demographic：adj 人口统计的
  • a welcome development:受欢迎的发展
• The UN’s Sustainable Development Goals, for example, seek to tackle these challenges by reducing poverty and hunger, and improving education. The recent AI for Good Global Summit 2017 highlighted how AI can support these efforts. Suggestions ranged from initiatives aimed at monitoring the progress of the international community toward achieving these goals and determining where resources are most needed, to predictive modelling of disease outbreaks [19]. AI is also poised to operate in conjunction with the IoT, drones and synthetic biology to power smart agriculture and provide insights concerning when and where to plant and harvest crops while optimizing nutrition, pesticides and water to increase yields and help combat world hunger [20]
• 比如说，联合国可持续发展的目标尝试通过减少贫穷人口和饥饿人口，以及该改良教育来应对这些挑战。最近的2017年的AI全球公益峰会在不断强调AI如何能对这些努力（通过减少贫困人口，饥饿人口和改良教育来缓解人口变化挑战的努力）进行支持。给了很多建议，从倡议监督相关相关国际社会的进展，搭配提出相关疾病爆发的预测模型，这些国际社会主要去实现这些目标并决定哪个地方最需要资源。AI也可以和物联网，无人机和合成生物学进行的组合，为智慧农业助力，并且在何时何地种植和收获作物提供见解，同时优化营养，农药和水的配比以提高产能帮助战胜世界饥饿。
• 生词：
  • in conjunction with：结合 同时发生 连词 相合
  • initiative:n. 措施，倡议；主动性，积极性；主动权；（美国某些州的）公民立法提案程序
  • poise：n. 沉着，镇定；（优雅的）体态，姿态；<古>平衡；（物理）泊 v. （使）平衡，（使）悬浮；准备做……（be poised）
  • drone：v. 嗡嗡叫，嗡嗡响；（人）絮絮叨叨地说，喋喋不休地说；混日子 n. 无人机驾驶
  • synthetic：adj. 合成的，人造的；（感情，行动）不诚恳的，虚假的；（语言）综合型的；（命题）综合的
  • provide insight：提供见解
  • pesticide：农药
• The interest of governments in the potential of AI is increasingly rapidly, and as it does, AI is expected to make contributions to public policy, particularly economic policy. When constructing a model, economists usually begin with a set of assumptions which they then seek to verify. AI, however, offers the power to analyze data and uncover previously unknown interactions between variables, on which a model can then be built from first principles and serve to inform public and monetary policies.
• 政府对于AI潜力的兴趣日益增长，并且与此同时，AI有望对公共政策，尤其是经济政策做出相关的帮助。当建造一个模型时，经济学家常常会以一些他们尝试验证的假设为开始。然而，AI提供数据分析的能力并且揭露参数之间之前未知的联系，在此基础上，可以跟你举第一个原则建立一个模型，并使用模型去为公共和货币政策提供信息。
• 生词：
  • inform public and money policies：为公共和货币政策提供信息
  • monetary：货币
  • offer power to：提供什么能力
  • as it does：与此同时
• AI may also support financial regulators in their monitoring activities by inspecting the balance sheets of banks for anomalies that are of concern from a prudential or conduct perspective [21]. Such explorations into how AI can be used in the public sphere should be welcomed, as they can help governments explore best practices, make more informed decisions on AI policy and build public trust.
• AI也能够支持金融监管机构的检测活动，主要是通过调查资产负债，以发现那些需要从谨慎或者行为的角度来看值得关注的异常行为。这种AI如何应用到公众领域的探索应该受欢迎的，因为他们能帮助政府探索最佳实践，在AI政策的基础上做出更加明智的决定，同时建立公众信任。
• 生词：
  • balance sheets of banks:银行的资产负债表
  • prudential：adj. 谨慎的；明辨的
  • regulator：监管机构
  • anomaly：异常
  • informed：adj. 有知识的，了解情况的；有根据的，可靠的；明智的，有见识的
  • explore beat practice：做出最佳探索

2.5 Service and product customization（服务和产品的定制）

• The integration of AI and advanced manufacturing enables mass customization that easily connects suppliers, partners and customers, and meets individualized demands with efficiency and cost close to mass production. Applying AI therefore optimizes the manufacturing value chain, so that manufacturers can track flows of materials in real-time. They can also assess more accurately engineering and quality issues, reduce excess inventory and logistics delays, increase responsiveness to customer needs and make better business decisions that reduce waste and costs. Businesses will benefit from mass customization of production by empowering internet-connected consumers to control intelligent manufacturing processes in order to develop products according to their desired specifications.
• AI和先进制造的合并使得产品定制，同时接近批量生产的成本成为可能，因为这很容易地联系起了供应方，合作者和顾客，并且十分高效地满足了个人需求。所以使用AI能够最优化制造价值链，使制造商能够实时追踪材料的流动。他们也能更加准确地评估工程和质量问题，减少过剩的库存和物流延迟，增加对客户需求的响应能力并且做出更好的，能够减少浪费和成本的决策。商业将从批量定制生产中获利，通过使互联网的消费者控制智能制造过程，能够根据他们需要的规格生产产品。
• 生词：
  • responsiveness：n 响应能力，有同情心
  • logistics：n. 后勤，组织工作，物流的
  • inventory ：n. 详细目录，清单；存货，库存；盘货，存货盘点
  • manufacturing：制造业
  • mass：n. 团，块；大量，许多；（常指混乱的）一群，一堆；大量的东西；群众，平民百姓；大多数，多数；质量；（尤指罗马天主教的）弥撒；弥撒曲
  • excess：n. 超过，过量；越轨行为，暴行；无节制，无度；（保险中的）免赔额 adj. 过多的，超额的；额外的，附加的
• AI allows the individualization of products on a completely new level. Not only product configurators that customers can use online are affected. The use of AI also opens up completely new possibilities for individualization. Products whose individualization would involve high development costs can be adapted to requirements by using AI.
• AI让产品私有化达到一个全新的水平。不仅仅是顾客能够在线使用的产品配置器能被影响。AI的使用也能完全打开私有化的新可能。个性化产品的开发成本较高，但是可以使用AI来适应需求
• Services also can be automatically tailored to customer needs. An example is the automatic translation of texts. While previously the results of rule-based techniques could often only reflect the meaning of individual words but not entire sentences that are contextualized, services supported by AI are able to perform translation based on the meaning of a text.
• 服务也能自动根据客户需求去定制。一个例子是文本的自动化翻译。以前，基于规则的技术结果往往只能反应单个单词的含义，但是不是至于上下文中的整个句子的含义，AI支持的服务能够基于文本的含义进行翻译。
• 生词：
  • tailor：n. （尤指男装）裁缝；蓝鱼（Tailorfish 的简称） v. 定做（衣服）；迎合，使适应
  • contextualize：v. 置于上下文（或背景）中

本文标签： 论文ArtificialIntellligencemiraclemirage