更改

人工智能 (查看源代码)

2021年7月25日 (日) 00:00的版本

删除75,626字节、 2021年7月25日 (日) 00:00

无编辑摘要

第80行：第80行：

Intelligent agents must be able to set goals and achieve them.<ref name="Planning"/> They need a way to visualize the future—a representation of the state of the world and be able to make predictions about how their actions will change it—and be able to make choices that maximize the [[utility]] (or "value") of available choices.<ref name="Information value theory"/>

−

Intelligent agents must be able to set goals and achieve them. They need a way to visualize the future—a representation of the state of the world and be able to make predictions about how their actions will change it—and be able to make choices that maximize the utility (or "value") of available choices.

智能体必须能够设定并实现目标。他们需要能够有设想未来的办法——这是一种对其所处环境状况的表述，并能够预测他们的行动将如何改变环境——依此能够选择使效用（或者“价值”）最大化的选项。

第169行：第167行：

There is no established unifying theory or [[paradigm]] that guides AI research. Researchers disagree about many issues.<ref>[[Nils Nilsson (researcher)|Nils Nilsson]] writes: "Simply put, there is wide disagreement in the field about what AI is all about" {{Harv|Nilsson|1983|p=10}}.</ref> A few of the most long standing questions that have remained unanswered are these: should artificial intelligence simulate natural intelligence by studying [[psychology]] or [[Neuroscience|neurobiology]]? Or is [[human biology]] as irrelevant to AI research as bird biology is to [[aeronautical engineering]]?<ref name="Biological intelligence vs. intelligence in general"/>

−

There is no established unifying theory or paradigm that guides AI research. Researchers disagree about many issues. A few of the most long standing questions that have remained unanswered are these: should artificial intelligence simulate natural intelligence by studying psychology or neurobiology? Or is human biology as irrelevant to AI research as bird biology is to aeronautical engineering?

目前还没有统一的理论或范式来指导AI的研究。研究人员在许多问题上存在分歧。<ref>[[Nils Nilsson (researcher)|Nils Nilsson]] writes: "Simply put, there is wide disagreement in the field about what AI is all about" {{Harv|Nilsson|1983|p=10}}.</ref>一些长期悬而未决的问题是: AI是否应该通过研究心理学或神经生物学来模拟天然智能？人类生物学和AI研究的关系和鸟类生物学和航空工程学的关系一样吗？

Can intelligent behavior be described using simple, elegant principles (such as [[logic]] or [[optimization (mathematics)|optimization]])? Or does it necessarily require solving a large number of completely unrelated problems?<ref name="Neats vs. scruffies"/>

−

~~Can intelligent behavior be described using simple, elegant principles (such as logic or optimization)? Or does it necessarily require solving a large number of completely unrelated problems?~~

智能行为可以用简单、优雅的原则(如逻辑或优化)来描述吗？还是需要去解决大量完全不相关的问题？

第186行：第181行：

In the 1940s and 1950s, a number of researchers explored the connection between [[neurobiology]], [[information theory]], and [[cybernetics]]. Some of them built machines that used electronic networks to exhibit rudimentary intelligence, such as [[W. Grey Walter]]'s [[turtle (robot)|turtles]] and the [[Johns Hopkins Beast]]. Many of these researchers gathered for meetings of the Teleological Society at [[Princeton University]] and the [[Ratio Club]] in England.<ref name="AI's immediate precursors"/> By 1960, this approach was largely abandoned, although elements of it would be revived in the 1980s.

−

In the 1940s and 1950s, a number of researchers explored the connection between neurobiology, information theory, and cybernetics. Some of them built machines that used electronic networks to exhibit rudimentary intelligence, such as W. Grey Walter's turtles and the Johns Hopkins Beast. Many of these researchers gathered for meetings of the Teleological Society at Princeton University and the Ratio Club in England. By 1960, this approach was largely abandoned, although elements of it would be revived in the 1980s.

在20世纪四五十年代，许多研究人员探索了神经生物学、信息论和控制论之间的联系。他们中的一些人利用电子网络制造机器来表现基本的智能，比如 '''W·格雷·沃尔特 W. Grey Walter'''的乌龟和'''约翰·霍普金斯 Johns Hopkins'''的野兽。这些研究人员中的许多人参加了在普林斯顿大学的'''目的论学社'''和英格兰的'''比率俱乐部'''举办的集会。到了1960年，这种方法基本上被放弃了，直到二十世纪八十年代一些部分又被重新使用。

第196行：第190行：

When access to digital computers became possible in the mid-1950s, AI research began to explore the possibility that human intelligence could be reduced to symbol manipulation. The research was centered in three institutions: [[Carnegie Mellon University]], [[Stanford]] and [[MIT]], and as described below, each one developed its own style of research. [[John Haugeland]] named these symbolic approaches to AI "good old fashioned AI" or "[[GOFAI]]".<ref name="GOFAI"/> During the 1960s, symbolic approaches had achieved great success at simulating high-level "thinking" in small demonstration programs. Approaches based on [[cybernetics]] or [[artificial neural network]]s were abandoned or pushed into the background.<ref>The most dramatic case of sub-symbolic AI being pushed into the background was the devastating critique of [[perceptron]]s by [[Marvin Minsky]] and [[Seymour Papert]] in 1969. See [[History of AI]], [[AI winter]], or [[Frank Rosenblatt]].</ref>

−

When access to digital computers became possible in the mid-1950s, AI research began to explore the possibility that human intelligence could be reduced to symbol manipulation. The research was centered in three institutions: Carnegie Mellon University, Stanford and MIT, and as described below, each one developed its own style of research. John Haugeland named these symbolic approaches to AI "good old fashioned AI" or "GOFAI".

20世纪50年代中期，当数字计算机成为可能时，AI研究开始探索把人类智能化归为符号操纵的可能性。这项研究集中在3个机构: 卡内基梅隆大学，斯坦福和麻省理工学院，正如下面所描述的，每个机构都有自己的研究风格。约翰 · 豪格兰德将这些具有象征意义的AI方法命名为“好的老式人工智能 Good Old Fashioned AI”或“ GOFAI”<ref name="GOFAI"/>。20世纪60年代的时候，符号化方法在模拟高层次“思考”的小型程序中取得了巨大的成就。而基于[[控制论]]和[[人工神经网络]]的方法则被抛弃，或者只作为背景出现。<ref>The most dramatic case of sub-symbolic AI being pushed into the background was the devastating critique of [[perceptron]]s by [[Marvin Minsky]] and [[Seymour Papert]] in 1969. See [[History of AI]], [[AI winter]], or [[Frank Rosenblatt]].</ref>

−

Researchers in the 1960s and the 1970s were convinced that symbolic approaches would eventually succeed in creating a machine with [[artificial general intelligence]] and considered this the goal of their field.

−

Researchers in the 1960s and the 1970s were convinced that symbolic approaches would eventually succeed in creating a machine with artificial general intelligence and considered this the goal of their field.

20世纪六七十年代的研究人员相信，符号方法最终会成功地创造出一台具有[[通用人工智能]]的机器，并以此作为他们研究领域的目标。

第214行：第203行：

Economist [[Herbert A. Simon|Herbert Simon]] and [[Allen Newell]] studied human problem-solving skills and attempted to formalize them, and their work laid the foundations of the field of artificial intelligence, as well as [[cognitive science]], [[operations research]] and [[management science]]. Their research team used the results of [[psychology|psychological]] experiments to develop programs that simulated the techniques that people used to solve problems. This tradition, centered at [[Carnegie Mellon University]] would eventually culminate in the development of the [[Soar (cognitive architecture)|Soar]] architecture in the middle 1980s.<ref name="AI at CMU in the 60s"/><ref name="Soar"/>

−

Economist Herbert Simon and Allen Newell studied human problem-solving skills and attempted to formalize them, and their work laid the foundations of the field of artificial intelligence, as well as cognitive science, operations research and management science. Their research team used the results of psychological experiments to develop programs that simulated the techniques that people used to solve problems. This tradition, centered at Carnegie Mellon University would eventually culminate in the development of the Soar architecture in the middle 1980s.

经济学家[[赫伯特·西蒙]]和[[艾伦·纽厄尔]]研究了人类解决问题的技能，并试图将其形式化。他们的工作为AI、认知科学、运筹学和管理科学奠定了基础。他们的研究团队利用心理学实验的结果来开发程序，模拟人们用来解决问题的方法。以卡内基梅隆大学为中心，这种研究传统最终在20世纪80年代中期的SOAR架构开发过程中达到顶峰。

第223行：第211行：

Unlike Simon and Newell, [[John McCarthy (computer scientist)|John McCarthy]] felt that machines did not need to simulate human thought, but should instead try to find the essence of abstract reasoning and problem-solving, regardless whether people used the same algorithms.<ref name="Biological intelligence vs. intelligence in general"/> His laboratory at [[Stanford University|Stanford]] ([[Stanford Artificial Intelligence Laboratory|SAIL]]) focused on using formal [[logic]] to solve a wide variety of problems, including [[knowledge representation]], [[automated planning and scheduling|planning]] and [[machine learning|learning]].<ref name="AI at Stanford in the 60s"/> Logic was also the focus of the work at the [[University of Edinburgh]] and elsewhere in Europe which led to the development of the programming language [[Prolog]] and the science of [[logic programming]].<ref name="AI at Edinburgh and France in the 60s"/>

−

Unlike Simon and Newell, John McCarthy felt that machines did not need to simulate human thought, but should instead try to find the essence of abstract reasoning and problem-solving, regardless whether people used the same algorithms. His laboratory at Stanford (SAIL) focused on using formal logic to solve a wide variety of problems, including knowledge representation, planning and learning. Logic was also the focus of the work at the University of Edinburgh and elsewhere in Europe which led to the development of the programming language Prolog and the science of logic programming.

与西蒙和纽厄尔不同，约翰·麦卡锡认为机器不需要模拟人类的思维，而是应该尝试寻找抽象推理和解决问题的本质，不管人们是否使用相同的算法。<ref name="Biological intelligence vs. intelligence in general"/> 他在斯坦福大学的实验室(SAIL)致力于使用形式逻辑来解决各种各样的问题，包括知识表示、规划和学习。逻辑也是爱丁堡大学和欧洲其他地方工作的重点，这促进了编程语言 Prolog 和逻辑编程科学的发展。

第232行：第219行：

Researchers at [[MIT]] (such as [[Marvin Minsky]] and [[Seymour Papert]])<ref name="AI at MIT in the 60s"/> found that solving difficult problems in [[computer vision|vision]] and [[natural language processing]] required ad-hoc solutions—they argued that there was no simple and general principle (like [[logic]]) that would capture all the aspects of intelligent behavior. [[Roger Schank]] described their "anti-logic" approaches as "[[Neats vs. scruffies|scruffy]]" (as opposed to the "[[neats vs. scruffies|neat]]" paradigms at [[Carnegie Mellon University|CMU]] and Stanford).<ref name="Neats vs. scruffies"/> [[Commonsense knowledge bases]] (such as [[Doug Lenat]]'s [[Cyc]]) are an example of "scruffy" AI, since they must be built by hand, one complicated concept at a time.<ref name="Cyc"/>

−

Researchers at MIT (such as Marvin Minsky and Seymour Papert) found that solving difficult problems in vision and natural language processing required ad-hoc solutions—they argued that there was no simple and general principle (like logic) that would capture all the aspects of intelligent behavior. Roger Schank described their "anti-logic" approaches as "scruffy" (as opposed to the "neat" paradigms at CMU and Stanford). Commonsense knowledge bases (such as Doug Lenat's Cyc) are an example of "scruffy" AI, since they must be built by hand, one complicated concept at a time.

麻省理工学院(MIT)的研究人员马文•明斯基和西摩•派珀特等发现<ref name="AI at MIT in the 60s"/>，视觉和自然语言处理中的难题需要特定的解决方案——他们认为，没有简单而普遍的原则(如逻辑)可以涵盖智能行为。罗杰•尚克将他们的“反逻辑”方法形容为“邋遢的”(相对于卡内基梅隆大学和斯坦福大学的“整洁”范式)。常识库(如常识知识库的 Cyc)是“邋遢”AI的一个例子，因为它们必须人工一个一个地构建复杂概念。

第239行：第225行：

When computers with large memories became available around 1970, researchers from all three traditions began to build [[knowledge representation|knowledge]] into AI applications.<ref name="Knowledge revolution"/> This "knowledge revolution" led to the development and deployment of [[expert system]]s (introduced by [[Edward Feigenbaum]]), the first truly successful form of AI software.<ref name="Expert systems"/> A key component of the system architecture for all expert systems is the knowledge base, which stores facts and rules that illustrate AI.<ref>{{Cite journal |last=Frederick |first=Hayes-Roth |last2=William |first2=Murray |last3=Leonard |first3=Adelman |title=Expert systems|journal=AccessScience |language=en |doi=10.1036/1097-8542.248550}}</ref> The knowledge revolution was also driven by the realization that enormous amounts of knowledge would be required by many simple AI applications.

−

When computers with large memories became available around 1970, researchers from all three traditions began to build knowledge into AI applications. The knowledge revolution was also driven by the realization that enormous amounts of knowledge would be required by many simple AI applications.

1970年左右，当拥有大容量存储器的计算机出现时，来自这三个研究方向的研究人员开始将知识应用于AI领域<ref name="Knowledge revolution"/> 。这一轮“知识革命”的一大成果是开发和部署专家系统，第一个真正成功的AI软件<ref name="Expert systems"/>。所有专家系统的一个关键部件是存储着事实和规则的知识库<ref>{{Cite journal |last=Frederick |first=Hayes-Roth |last2=William |first2=Murray |last3=Leonard |first3=Adelman |title=Expert systems|journal=AccessScience |language=en |doi=10.1036/1097-8542.248550}}</ref>。推动知识革命的另一个原因是人们认识到，许多简单的AI应用程序也需要大量的知识。

第250行：第234行：

By the 1980s, progress in symbolic AI seemed to stall and many believed that symbolic systems would never be able to imitate all the processes of human cognition, especially [[machine perception|perception]], [[robotics]], [[machine learning|learning]] and [[pattern recognition]]. A number of researchers began to look into "sub-symbolic" approaches to specific AI problems.<ref name="Symbolic vs. sub-symbolic"/> Sub-symbolic methods manage to approach intelligence without specific representations of knowledge.

−

By the 1980s, progress in symbolic AI seemed to stall and many believed that symbolic systems would never be able to imitate all the processes of human cognition, especially perception, robotics, learning and pattern recognition. A number of researchers began to look into "sub-symbolic" approaches to specific AI problems. Sub-symbolic methods manage to approach intelligence without specific representations of knowledge.

到了20世纪80年代，符号AI的进步似乎停滞不前，许多人认为符号系统永远无法模仿人类认知的所有过程，尤其在感知、机器人学、学习和模式识别等方面。许多研究人员开始研究针对特定AI问题的“亚符号”方法<ref name="Symbolic vs. sub-symbolic"/>。亚符号方法能在没有特定知识表示的情况下，做到接近智能。

第259行：第242行：

This includes [[embodied agent|embodied]], [[situated]], [[behavior-based AI|behavior-based]], and [[nouvelle AI]]. Researchers from the related field of [[robotics]], such as [[Rodney Brooks]], rejected symbolic AI and focused on the basic engineering problems that would allow robots to move and survive.<ref name="Embodied AI"/> Their work revived the non-symbolic point of view of the early [[cybernetic]]s researchers of the 1950s and reintroduced the use of [[control theory]] in AI. This coincided with the development of the [[embodied mind thesis]] in the related field of [[cognitive science]]: the idea that aspects of the body (such as movement, perception and visualization) are required for higher intelligence.

−

This includes embodied, situated, behavior-based, and nouvelle AI. Researchers from the related field of robotics, such as Rodney Brooks, rejected symbolic AI and focused on the basic engineering problems that would allow robots to move and survive. Their work revived the non-symbolic point of view of the early cybernetics researchers of the 1950s and reintroduced the use of control theory in AI. This coincided with the development of the embodied mind thesis in the related field of cognitive science: the idea that aspects of the body (such as movement, perception and visualization) are required for higher intelligence.

'''具身智慧 Embodied Intelligence'''包括具体化的、情境化的、基于行为的新式 AI。来自机器人相关领域的研究人员，如罗德尼·布鲁克斯，放弃了符号化AI的方法，而专注于使机器人能够移动和生存的基本工程问题<ref name="Embodied AI"/>。他们的工作重启了20世纪50年代早期控制论研究者的非符号观点，并将控制论重新引入到AI的应用中。这与认知科学相关领域的具身理论的发展相吻合: 认为如运动、感知和视觉等身体的各个功能是高智能所必需的。

第266行：第248行：

Within [[developmental robotics]], developmental learning approaches are elaborated upon to allow robots to accumulate repertoires of novel skills through autonomous self-exploration, social interaction with human teachers, and the use of guidance mechanisms (active learning, maturation, motor synergies, etc.).{{sfn|Weng|McClelland|Pentland|Sporns|2001}}{{sfn|Lungarella|Metta|Pfeifer|Sandini|2003}}{{sfn|Asada|Hosoda|Kuniyoshi|Ishiguro|2009}}{{sfn|Oudeyer|2010}}

−

Within developmental robotics, developmental learning approaches are elaborated upon to allow robots to accumulate repertoires of novel skills through autonomous self-exploration, social interaction with human teachers, and the use of guidance mechanisms (active learning, maturation, motor synergies, etc.).

在发展型机器人中，人们开发了发展型学习方法，通过自主的自我探索、与人类教师的社会互动，以及使用主动学习、成熟、协同运动等指导机制，使机器人积累新技能的能力。{{sfn|Weng|McClelland|Pentland|Sporns|2001}}{{sfn|Lungarella|Metta|Pfeifer|Sandini|2003}}{{sfn|Asada|Hosoda|Kuniyoshi|Ishiguro|2009}}{{sfn|Oudeyer|2010}}

第276行：第256行：

Interest in [[Artificial neural network|neural networks]] and "[[connectionism]]" was revived by [[David Rumelhart]] and others in the middle of the 1980s.<ref name="Revival of connectionism"/> [[Artificial neural network]]s are an example of [[soft computing]]—they are solutions to problems which cannot be solved with complete logical certainty, and where an approximate solution is often sufficient. Other [[soft computing]] approaches to AI include [[fuzzy system]]s, [[Grey system theory]], [[evolutionary computation]] and many statistical tools. The application of soft computing to AI is studied collectively by the emerging discipline of [[computational intelligence]].<ref name="Computational intelligence"/>

−

Interest in neural networks and "connectionism" was revived by David Rumelhart and others in the middle of the 1980s. Artificial neural networks are an example of soft computing—they are solutions to problems which cannot be solved with complete logical certainty, and where an approximate solution is often sufficient. Other soft computing approaches to AI include fuzzy systems, Grey system theory, evolutionary computation and many statistical tools. The application of soft computing to AI is studied collectively by the emerging discipline of computational intelligence.

上世纪80年代中期，大卫•鲁梅尔哈特等人重新激发了人们对神经网络和“'''连接主义 Connectionism'''”的兴趣。人工神经网络是软计算的一个例子ーー它们解决不能完全用逻辑确定性地解决，但常常只需要近似解的问题。AI的其他软计算方法包括'''模糊系统 Fuzzy Systems '''、'''灰度系统理论 Grey System Theory'''、'''演化计算 Evolutionary Computation '''和许多统计工具。软计算在AI中的应用是计算智能这一新兴学科的集中研究领域。

第286行：第265行：

Much of traditional [[Symbolic artificial intelligence|GOFAI]] got bogged down on ''ad hoc'' patches to [[symbolic computation]] that worked on their own toy models but failed to generalize to real-world results. However, around the 1990s, AI researchers adopted sophisticated mathematical tools, such as [[hidden Markov model]]s (HMM), [[information theory]], and normative Bayesian [[decision theory]] to compare or to unify competing architectures. The shared mathematical language permitted a high level of collaboration with more established fields (like [[mathematics]], economics or [[operations research]]).{{efn|While such a "victory of the neats" may be a consequence of the field becoming more mature, [[Artificial Intelligence: A Modern Approach|AIMA]] states that in practice both [[neats and scruffies|neat and scruffy]] approaches continue to be necessary in AI research.}} Compared with GOFAI, new "statistical learning" techniques such as HMM and neural networks were gaining higher levels of accuracy in many practical domains such as [[data mining]], without necessarily acquiring a semantic understanding of the datasets. The increased successes with real-world data led to increasing emphasis on comparing different approaches against shared test data to see which approach performed best in a broader context than that provided by idiosyncratic toy models; AI research was becoming more [[scientific method|scientific]]. Nowadays results of experiments are often rigorously measurable, and are sometimes (with difficulty) reproducible.<ref name="Formal methods in AI"/><ref>{{cite news|last1=Hutson|first1=Matthew|title=Artificial intelligence faces reproducibility crisis|url=http://science.sciencemag.org/content/359/6377/725|accessdate=28 April 2018|work=[[Science Magazine|Science]]|date=16 February 2018|pages=725–726|language=en|doi=10.1126/science.359.6377.725|bibcode=2018Sci...359..725H}}</ref> Different statistical learning techniques have different limitations; for example, basic HMM cannot model the infinite possible combinations of natural language.{{sfn|Norvig|2012}} Critics note that the shift from GOFAI to statistical learning is often also a shift away from [[explainable AI]]. In AGI research, some scholars caution against over-reliance on statistical learning, and argue that continuing research into GOFAI will still be necessary to attain general intelligence.{{sfn|Langley|2011}}{{sfn|Katz|2012}}

−

Much of traditional GOFAI got bogged down on ad hoc patches to symbolic computation that worked on their own toy models but failed to generalize to real-world results. However, around the 1990s, AI researchers adopted sophisticated mathematical tools, such as hidden Markov models (HMM), information theory, and normative Bayesian decision theory to compare or to unify competing architectures. The shared mathematical language permitted a high level of collaboration with more established fields (like mathematics, economics or operations research). Compared with GOFAI, new "statistical learning" techniques such as HMM and neural networks were gaining higher levels of accuracy in many practical domains such as data mining, without necessarily acquiring a semantic understanding of the datasets. The increased successes with real-world data led to increasing emphasis on comparing different approaches against shared test data to see which approach performed best in a broader context than that provided by idiosyncratic toy models; AI research was becoming more scientific. Nowadays results of experiments are often rigorously measurable, and are sometimes (with difficulty) reproducible. Different statistical learning techniques have different limitations; for example, basic HMM cannot model the infinite possible combinations of natural language. Critics note that the shift from GOFAI to statistical learning is often also a shift away from explainable AI. In AGI research, some scholars caution against over-reliance on statistical learning, and argue that continuing research into GOFAI will still be necessary to attain general intelligence.

许多传统的 GOFAI 在实验模型中行之有效，但不能推广到现实世界，陷入了需要不断给符号计算修补漏洞的困境中。然而，在20世纪90年代前后，AI研究人员采用了复杂的数学工具，如'''隐马尔可夫模型 Hidden Markov Model,HMM'''、信息论和'''标准贝叶斯决策理论 Normative Bayesian Decision Theory'''来比较或统一各种互相竞争的架构。共通的数学语言允许其与数学、经济学或运筹学等更成熟的领域进行高层次的融合。与 GOFAI 相比，隐马尔可夫模型和神经网络等新的“统计学习”技术在数据挖掘等许多实际领域中不必理解数据集的语义，却能得到更高的精度，随着现实世界数据的日益增加，人们越来越注重用不同的方法测试相同的数据，并进行比较，看哪种方法在比特殊实验室环境更广泛的背景下表现得更好; AI研究正变得更加科学。如今，实验结果一般是严格可测的，有时可以重现（但有难度）。不同的统计学习技术有不同的局限性，例如，基本的 HMM 不能为自然语言的无限可能的组合建模。评论者们指出，从 GOFAI 到统计学习的转变也经常是可解释AI的转变。在 [[通用人工智能]] 的研究中，一些学者警告不要过度依赖统计学习，并认为继续研究 GOFAI 仍然是实现通用智能的必要条件。

第295行：第273行：

;Intelligent agent paradigm: An [[intelligent agent]] is a system that perceives its environment and takes actions that maximize its chances of success. The simplest intelligent agents are programs that solve specific problems. More complicated agents include human beings and organizations of human beings (such as [[firm]]s). The paradigm allows researchers to directly compare or even combine different approaches to isolated problems, by asking which agent is best at maximizing a given "goal function". An agent that solves a specific problem can use any approach that works—some agents are symbolic and logical, some are sub-symbolic [[artificial neural network]]s and others may use new approaches. The paradigm also gives researchers a common language to communicate with other fields—such as [[decision theory]] and economics—that also use concepts of abstract agents. Building a complete agent requires researchers to address realistic problems of integration; for example, because sensory systems give uncertain information about the environment, planning systems must be able to function in the presence of uncertainty. The intelligent agent paradigm became widely accepted during the 1990s.<ref name="Intelligent agents"/>

−

Intelligent agent paradigm: An intelligent agent is a system that perceives its environment and takes actions that maximize its chances of success. The simplest intelligent agents are programs that solve specific problems. More complicated agents include human beings and organizations of human beings (such as firms). The paradigm allows researchers to directly compare or even combine different approaches to isolated problems, by asking which agent is best at maximizing a given "goal function". An agent that solves a specific problem can use any approach that works—some agents are symbolic and logical, some are sub-symbolic artificial neural networks and others may use new approaches. The paradigm also gives researchers a common language to communicate with other fields—such as decision theory and economics—that also use concepts of abstract agents. Building a complete agent requires researchers to address realistic problems of integration; for example, because sensory systems give uncertain information about the environment, planning systems must be able to function in the presence of uncertainty. The intelligent agent paradigm became widely accepted during the 1990s.

;智能主体范式: 智能主体是一个感知其环境并采取行动，最大限度地提高其成功机会的系统。最简单的智能主体是解决特定问题的程序。更复杂的智能主体包括人类和人类组织(如公司)。这种范式使得研究人员能通过观察哪一个智能主体能最大化给定的“目标函数”，直接比较甚至结合不同的方法来解决孤立的问题。解决特定问题的智能主体可以使用任何有效的方法——可以是是符号化和逻辑化的，也可以是亚符号化的人工神经网络，还可以是新的方法。这种范式还为研究人员提供了一种与其他领域(如决策理论和经济学)进行交流的共同语言，因为这些领域也使用了抽象智能主体的概念。建立一个完整的智能主体需要研究人员解决现实的整合协调问题; 例如，由于传感系统提供关于环境的信息不确定，决策系统就必须在不确定性的条件下运作。智能体范式在20世纪90年代被广泛接受。

第302行：第279行：

;[[Agent architecture]]s and [[cognitive architecture]]s:Researchers have designed systems to build intelligent systems out of interacting [[intelligent agent]]s in a [[multi-agent system]].<ref name="Agent architectures"/> A [[hierarchical control system]] provides a bridge between sub-symbolic AI at its lowest, reactive levels and traditional symbolic AI at its highest levels, where relaxed time constraints permit planning and world modeling.<ref name="Hierarchical control system"/> Some cognitive architectures are custom-built to solve a narrow problem; others, such as [[Soar (cognitive architecture)|Soar]], are designed to mimic human cognition and to provide insight into general intelligence. Modern extensions of Soar are [[hybrid intelligent system]]s that include both symbolic and sub-symbolic components.<ref>{{cite journal|last1=Laird|first1=John|title=Extending the Soar cognitive architecture|journal=Frontiers in Artificial Intelligence and Applications|date=2008|volume=171|page=224|citeseerx=10.1.1.77.2473}}</ref><ref>{{cite journal|last1=Lieto|first1=Antonio|last2=Lebiere|first2=Christian|last3=Oltramari|first3=Alessandro|title=The knowledge level in cognitive architectures: Current limitations and possibile developments|journal=Cognitive Systems Research|date=May 2018|volume=48|pages=39–55|doi=10.1016/j.cogsys.2017.05.001|hdl=2318/1665207|hdl-access=free}}</ref><ref>{{cite journal|last1=Lieto|first1=Antonio|last2=Bhatt|first2=Mehul|last3=Oltramari|first3=Alessandro|last4=Vernon|first4=David|title=The role of cognitive architectures in general artificial intelligence|journal=Cognitive Systems Research|date=May 2018|volume=48|pages=1–3|doi=10.1016/j.cogsys.2017.08.003|hdl=2318/1665249|hdl-access=free}}</ref>

−

~~Agent architectures and cognitive architectures:Researchers have designed systems to build intelligent systems out of interacting intelligent agents in a multi-agent system.~~

;智能主体体系结构和认知体系结构: 研究人员已经设计了一些在多智能体系统中利用相互作用的智能体构建智能系统的系统<ref name="Agent architectures"/>。分层控制系统为亚符号AI、反应层和符号AI提供了一座桥梁，亚符号AI在底层、反应层和符号AI在顶层<ref name="Hierarchical control system"/>。

第309行：第284行：

==工具 ==

−

~~AI has developed many tools to solve the most difficult problems in [[computer science]]. A few of the most general of these methods are discussed below.~~

−

~~AI has developed many tools to solve the most difficult problems in computer science. A few of the most general of these methods are discussed below.~~

AI领域已经开发出许多工具来解决计算机科学中最困难的问题。下面将讨论其中一些最常用的方法。

第323行：第293行：

Many problems in AI can be solved in theory by intelligently searching through many possible solutions:<ref name="Search"/> [[#Deduction, reasoning, problem solving|Reasoning]] can be reduced to performing a search. For example, logical proof can be viewed as searching for a path that leads from [[premise]]s to [[Logical consequence|conclusions]], where each step is the application of an [[inference rule]].<ref name="Logic as search"/> [[Automated planning and scheduling|Planning]] algorithms search through trees of goals and subgoals, attempting to find a path to a target goal, a process called [[means-ends analysis]].<ref name="Planning as search"/> [[Robotics]] algorithms for moving limbs and grasping objects use [[local search (optimization)|local searches]] in [[Configuration space (physics)|configuration space]].<ref name="Configuration space"/> Many [[machine learning|learning]] algorithms use search algorithms based on [[optimization (mathematics)|optimization]].

−

Many problems in AI can be solved in theory by intelligently searching through many possible solutions: Reasoning can be reduced to performing a search. For example, logical proof can be viewed as searching for a path that leads from premises to conclusions, where each step is the application of an inference rule. Planning algorithms search through trees of goals and subgoals, attempting to find a path to a target goal, a process called means-ends analysis. Robotics algorithms for moving limbs and grasping objects use local searches in configuration space. Many learning algorithms use search algorithms based on optimization.

AI中的许多问题可以通过智能地搜索许多可能的解决方案而在理论上得到解决<ref name="Search"/>: 推理可以简化为执行一次搜索。例如，逻辑证明可以看作是寻找一条从前提到结论的路径，其中每一步都用到了推理规则。规划算法通过搜索目标和子目标的树，试图找到一条通往目标的路径，这个过程称为“目的-手段”分析。机器人学中移动肢体和抓取物体的算法使用的是位形空间的局部搜索。许多学习算法也使用到了基于优化的搜索算法。

第330行：第298行：

Simple exhaustive searches<ref name="Uninformed search"/> are rarely sufficient for most real-world problems: the [[search algorithm|search space]] (the number of places to search) quickly grows to [[Astronomically large|astronomical numbers]]. The result is a search that is [[Computation time|too slow]] or never completes. The solution, for many problems, is to use "[[heuristics]]" or "rules of thumb" that prioritize choices in favor of those that are more likely to reach a goal and to do so in a shorter number of steps. In some search methodologies heuristics can also serve to entirely eliminate some choices that are unlikely to lead to a goal (called "[[pruning (algorithm)|pruning]] the [[search tree]]"). [[Heuristics]] supply the program with a "best guess" for the path on which the solution lies.<ref name="Informed search"/> Heuristics limit the search for solutions into a smaller sample size.{{sfn|Tecuci|2012}}

−

Simple exhaustive searches are rarely sufficient for most real-world problems: the search space (the number of places to search) quickly grows to astronomical numbers. The result is a search that is too slow or never completes. The solution, for many problems, is to use "heuristics" or "rules of thumb" that prioritize choices in favor of those that are more likely to reach a goal and to do so in a shorter number of steps. In some search methodologies heuristics can also serve to entirely eliminate some choices that are unlikely to lead to a goal (called "pruning the search tree"). Heuristics supply the program with a "best guess" for the path on which the solution lies. Heuristics limit the search for solutions into a smaller sample size.

对于大多数真实世界的问题，简单的穷举搜索<ref name="Uninformed search"/>很难满足要求: 搜索空间(要搜索的位置数)很快就会增加到天文数字。结果就是搜索速度太慢或者永远不能完成。对于许多问题，解决方法是使用“'''启发式 Heuristics''' ”或“'''经验法则 Rules of Thumb''' ” ，优先考虑那些更有可能达到目标的选择，并且在较短的步骤内完成。在一些搜索方法中，启发式方法还可以完全移去一些不可能通向目标的选择(称为“修剪搜索树”)。启发式为程序提供了解决方案所在路径的“最佳猜测”。启发式把搜索限制在了更小的样本规模里。。

第337行：第303行：

A very different kind of search came to prominence in the 1990s, based on the mathematical theory of [[optimization (mathematics)|optimization]]. For many problems, it is possible to begin the search with some form of a guess and then refine the guess incrementally until no more refinements can be made. These algorithms can be visualized as blind [[hill climbing]]: we begin the search at a random point on the landscape, and then, by jumps or steps, we keep moving our guess uphill, until we reach the top. Other optimization algorithms are [[simulated annealing]], [[beam search]] and [[random optimization]].<ref name="Optimization search"/>

−

A very different kind of search came to prominence in the 1990s, based on the mathematical theory of optimization. For many problems, it is possible to begin the search with some form of a guess and then refine the guess incrementally until no more refinements can be made. These algorithms can be visualized as blind hill climbing: we begin the search at a random point on the landscape, and then, by jumps or steps, we keep moving our guess uphill, until we reach the top. Other optimization algorithms are simulated annealing, beam search and random optimization.

在20世纪90年代，一种非常不同的基于数学最优化理论的搜索引起了人们的注意。对于许多问题，可以从某种形式的猜测开始搜索，然后逐步细化猜测，直到无法进行更多的细化。这些算法可以喻为盲目地爬山: 我们从地形上的一个随机点开始搜索，然后，通过跳跃或登爬，我们把猜测点继续向山上移动，直到我们到达山顶。其他的优化算法有 '''模拟退火算法''' 、'''定向搜索''' 和'''随机优化''' 。<ref name="Optimization search"/>

第352行：第316行：

[[Evolutionary computation]] uses a form of optimization search. For example, they may begin with a population of organisms (the guesses) and then allow them to mutate and recombine, [[artificial selection|selecting]] only the fittest to survive each generation (refining the guesses). Classic [[evolutionary algorithms]] include [[genetic algorithms]], [[gene expression programming]], and [[genetic programming]].<ref name="Genetic programming"/> Alternatively, distributed search processes can coordinate via [[swarm intelligence]] algorithms. Two popular swarm algorithms used in search are [[particle swarm optimization]] (inspired by bird [[flocking (behavior)|flocking]]) and [[ant colony optimization]] (inspired by [[ant trail]]s).<ref name="Society based learning"/><ref>{{cite book|author1=Daniel Merkle|author2=Martin Middendorf|editor1-last=Burke|editor1-first=Edmund K.|editor2-last=Kendall|editor2-first=Graham|title=Search Methodologies: Introductory Tutorials in Optimization and Decision Support Techniques|date=2013|publisher=Springer Science & Business Media|isbn=978-1-4614-6940-7|language=en|chapter=Swarm Intelligence}}</ref>

−

Evolutionary computation uses a form of optimization search. For example, they may begin with a population of organisms (the guesses) and then allow them to mutate and recombine, selecting only the fittest to survive each generation (refining the guesses). Classic evolutionary algorithms include genetic algorithms, gene expression programming, and genetic programming.

[[演化计算]]用到了优化搜索的形式。例如，他们可能从一群有机体(猜测)开始，然后让它们变异和重组，选择适者继续生存 (改进猜测)。经典的演化算法包括遗传算法、基因表达编程和遗传编程。<ref name="Genetic programming"/> Alternatively, distributed search processes can coordinate via [[swarm intelligence]] algorithms. Two popular swarm algorithms used in search are [[particle swarm optimization]] (inspired by bird [[flocking (behavior)|flocking]]) and [[ant colony optimization]] (inspired by [[ant trail]]s).<ref name="Society based learning"/><ref>{{cite book|author1=Daniel Merkle|author2=Martin Middendorf|editor1-last=Burke|editor1-first=Edmund K.|editor2-last=Kendall|editor2-first=Graham|title=Search Methodologies: Introductory Tutorials in Optimization and Decision Support Techniques|date=2013|publisher=Springer Science & Business Media|isbn=978-1-4614-6940-7|language=en|chapter=Swarm Intelligence}}</ref>

第358行：第321行：

===逻辑===

−

~~{{Main|Logic programming|Automated reasoning}}~~

[[Logic]]<ref name="Logic"/> is used for knowledge representation and problem solving, but it can be applied to other problems as well. For example, the [[satplan]] algorithm uses logic for [[automated planning and scheduling|planning]]<ref name="Satplan"/> and [[inductive logic programming]] is a method for [[machine learning|learning]].<ref name="Symbolic learning techniques"/>

−

Logic is used for knowledge representation and problem solving, but it can be applied to other problems as well. For example, the satplan algorithm uses logic for planning and inductive logic programming is a method for learning.

逻辑<ref name="Logic"/>被用来表示知识和解决问题，还可以应用到其他问题上。例如，satplan 算法就使用逻辑进行规划<ref name="Satplan"/>。另外，归纳逻辑编程是一种学习方法。

Several different forms of logic are used in AI research. [[Propositional logic]]<ref name="Propositional logic"/> involves [[truth function]]s such as "or" and "not". [[First-order logic]]<ref name="First-order logic"/> adds [[quantifier (logic)|quantifiers]] and [[predicate (mathematical logic)|predicates]], and can express facts about objects, their properties, and their relations with each other. [[Fuzzy set theory]] assigns a "degree of truth" (between 0 and 1) to vague statements such as "Alice is old" (or rich, or tall, or hungry) that are too linguistically imprecise to be completely true or false. [[Fuzzy logic]] is successfully used in [[control system]]s to allow experts to contribute vague rules such as "if you are close to the destination station and moving fast, increase the train's brake pressure"; these vague rules can then be numerically refined within the system. Fuzzy logic fails to scale well in knowledge bases; many AI researchers question the validity of chaining fuzzy-logic inferences.{{efn|"There exist many different types of uncertainty, vagueness, and ignorance... [We] independently confirm the inadequacy of systems for reasoning about uncertainty that propagates numerical factors according to only to which connectives appear in assertions."<ref>{{cite journal|last1=Elkan|first1=Charles|title=The paradoxical success of fuzzy logic|journal=IEEE Expert|date=1994|volume=9|issue=4|pages=3–49|doi=10.1109/64.336150|citeseerx=10.1.1.100.8402}}</ref>}}<ref name="Fuzzy logic"/><ref>{{cite news|title=What is 'fuzzy logic'? Are there computers that are inherently fuzzy and do not apply the usual binary logic?|url=https://www.scientificamerican.com/article/what-is-fuzzy-logic-are-t/|accessdate=5 May 2018|work=Scientific American|language=en}}</ref>

−

~~Several different forms of logic are used in AI research. Propositional logic}}~~

AI研究中使用了多种不同形式的逻辑。命题逻辑<ref name="Propositional logic"/>包含诸如“或”和“否”这样的真值函数。一阶逻辑<ref name="First-order logic"/>增加了量词和谓词，可以表达关于对象、对象属性和对象之间的关系。模糊集合论给诸如“爱丽丝老了”(或是富有的、高的、饥饿的)这样模糊的表述赋予了一个“真实程度”(介于0到1之间)，这些表述在语言上很模糊，不能完全判定为正确或错误。模糊逻辑在控制系统中得到了成功应用，使专家能够制定模糊规则，比如“如果你正以较快的速度接近终点站，那么就增加列车的制动压力”;这些模糊的规则可以在系统内用数值细化。但是，模糊逻辑无助于扩展知识库，许多AI研究者质疑把模糊逻辑和推理结合起来的有效性。{{efn|"There exist many different types of uncertainty, vagueness, and ignorance... [We] independently confirm the inadequacy of systems for reasoning about uncertainty that propagates numerical factors according to only to which connectives appear in assertions."<ref>{{cite journal|last1=Elkan|first1=Charles|title=The paradoxical success of fuzzy logic|journal=IEEE Expert|date=1994|volume=9|issue=4|pages=3–49|doi=10.1109/64.336150|citeseerx=10.1.1.100.8402}}</ref>}}<ref name="Fuzzy logic"/><ref>{{cite news|title=What is 'fuzzy logic'? Are there computers that are inherently fuzzy and do not apply the usual binary logic?|url=https://www.scientificamerican.com/article/what-is-fuzzy-logic-are-t/|accessdate=5 May 2018|work=Scientific American|language=en}}</ref>

第375行：第332行：

[[Default logic]]s, [[non-monotonic logic]]s and [[circumscription (logic)|circumscription]]<ref name="Default reasoning and non-monotonic logic"/> are forms of logic designed to help with default reasoning and the [[qualification problem]]. Several extensions of logic have been designed to handle specific domains of [[knowledge representation|knowledge]], such as: [[description logic]]s;<ref name="Representing categories and relations"/> [[situation calculus]], [[event calculus]] and [[fluent calculus]] (for representing events and time);<ref name="Representing time"/> [[Causality#Causal calculus|causal calculus]];<ref name="Representing causation"/> [[Belief revision|belief calculus (belief revision)]];<ref>"The Belief Calculus and Uncertain Reasoning", Yen-Teh Hsia</ref> and [[modal logic]]s.<ref name="Representing knowledge about knowledge"/> Logics to model contradictory or inconsistent statements arising in multi-agent systems have also been designed, such as [[paraconsistent logic]]s.

−

Default logics, non-monotonic logics and circumscription and modal logics. Logics to model contradictory or inconsistent statements arising in multi-agent systems have also been designed, such as paraconsistent logics.

第383行：第338行：

===不确定推理的概率方法===

−

~~{{Main|Bayesian network|Hidden Markov model|Kalman filter|Particle filter|Decision theory|Utility theory}}~~

−

[[File:EM Clustering of Old Faithful data.gif|right|frame|[[Expectation-maximization]] clustering of [[Old Faithful]] eruption data starts from a random guess but then successfully converges on an accurate clustering of the two physically distinct modes of eruption.[[期望-最大化老实泉喷发数据的聚类从一个随机的猜测开始，然后成功地收敛到两个物理上截然不同的喷发模式的精确聚类]]]]

第393行：第345行：

Many problems in AI (in reasoning, planning, learning, perception, and robotics) require the agent to operate with incomplete or uncertain information. AI researchers have devised a number of powerful tools to solve these problems using methods from [[probability]] theory and economics.<ref name="Stochastic methods for uncertain reasoning"/>

−

Many problems in AI (in reasoning, planning, learning, perception, and robotics) require the agent to operate with incomplete or uncertain information. AI researchers have devised a number of powerful tools to solve these problems using methods from probability theory and economics.

AI中的许多问题(在推理、规划、学习、感知和机器人技术方面)要求主体在信息不完整或不确定的情况下进行操作。AI研究人员从概率论和经济学的角度设计了许多强大的工具来解决这些问题。

第400行：第350行：

[[Bayesian network]]s<ref name="Bayesian networks"/> are a very general tool that can be used for various problems: reasoning (using the [[Bayesian inference]] algorithm),<ref name="Bayesian inference"/> [[Machine learning|learning]] (using the [[expectation-maximization algorithm]]),{{efn|Expectation-maximization, one of the most popular algorithms in machine learning, allows clustering in the presence of unknown [[latent variables]]{{sfn|Domingos|2015|p=210}}}}<ref name="Bayesian learning"/> [[Automated planning and scheduling|planning]] (using [[decision network]]s)<ref name="Bayesian decision networks"/> and [[machine perception|perception]] (using [[dynamic Bayesian network]]s).<ref name="Stochastic temporal models"/> Probabilistic algorithms can also be used for filtering, prediction, smoothing and finding explanations for streams of data, helping [[machine perception|perception]] systems to analyze processes that occur over time (e.g., [[hidden Markov model]]s or [[Kalman filter]]s).<ref name="Stochastic temporal models"/> Compared with symbolic logic, formal Bayesian inference is computationally expensive. For inference to be tractable, most observations must be [[conditionally independent]] of one another. Complicated graphs with diamonds or other "loops" (undirected [[cycle (graph theory)|cycles]]) can require a sophisticated method such as [[Markov chain Monte Carlo]], which spreads an ensemble of [[random walk]]ers throughout the Bayesian network and attempts to converge to an assessment of the conditional probabilities. Bayesian networks are used on [[Xbox Live]] to rate and match players; wins and losses are "evidence" of how good a player is{{citation needed|date=July 2019}}. [[Google AdSense|AdSense]] uses a Bayesian network with over 300 million edges to learn which ads to serve.{{sfn|Domingos|2015|loc=chapter 6}}

−

Bayesian networks are a very general tool that can be used for various problems: reasoning (using the Bayesian inference algorithm), learning (using the expectation-maximization algorithm),}} planning (using decision networks) and perception (using dynamic Bayesian networks). Probabilistic algorithms can also be used for filtering, prediction, smoothing and finding explanations for streams of data, helping perception systems to analyze processes that occur over time (e.g., hidden Markov models or Kalman filters). Compared with symbolic logic, formal Bayesian inference is computationally expensive. For inference to be tractable, most observations must be conditionally independent of one another. Complicated graphs with diamonds or other "loops" (undirected cycles) can require a sophisticated method such as Markov chain Monte Carlo, which spreads an ensemble of random walkers throughout the Bayesian network and attempts to converge to an assessment of the conditional probabilities. Bayesian networks are used on Xbox Live to rate and match players; wins and losses are "evidence" of how good a player is. AdSense uses a Bayesian network with over 300 million edges to learn which ads to serve.

'''[[贝叶斯网络]] Bayesian Networks ''' 是一个非常通用的工具，可用于各种问题: 推理(使用贝叶斯推断算法) ，学习(使用期望最大化算法) ，规划(使用决策网络)和感知(使用动态贝叶斯网络)。概率算法也可以用于滤波、预测、平滑和解释数据流，帮助传感系统分析随时间发生的过程(例如，隐马尔可夫模型或'''卡尔曼滤波器 Kalman Filters''')。与符号逻辑相比，形式化的贝叶斯推断逻辑运算量很大。为了使推理易于处理，大多数观察值必须彼此条件独立。含有菱形或其他“圈”(无向循环)的复杂图形可能需要比如马尔科夫-蒙特卡罗图的复杂方法，这种方法将一组随机行走遍布整个贝叶斯网络，并试图收敛到对条件概率的评估。贝叶斯网络在 Xbox Live 上被用来评估和匹配玩家：胜率是证明一个玩家有多有优秀的“证据”。AdSense使用一个有超过3亿条边的贝叶斯网络来学习如何推送广告的。

第407行：第355行：

A key concept from the science of economics is "[[utility]]": a measure of how valuable something is to an intelligent agent. Precise mathematical tools have been developed that analyze how an agent can make choices and plan, using [[decision theory]], [[decision analysis]],<ref name="Decisions theory and analysis"/> and [[applied information economics|information value theory]].<ref name="Information value theory"/> These tools include models such as [[Markov decision process]]es,<ref name="Markov decision process"/> dynamic [[decision network]]s,<ref name="Stochastic temporal models"/> [[game theory]] and [[mechanism design]].<ref name="Game theory and mechanism design"/>

−

A key concept from the science of economics is "utility": a measure of how valuable something is to an intelligent agent. Precise mathematical tools have been developed that analyze how an agent can make choices and plan, using decision theory, decision analysis, and information value theory. These tools include models such as Markov decision processes, dynamic decision networks, game theory and mechanism design.

经济学中的一个关键概念是“效用” :这是一种衡量某物对于一个智能主体的价值的方法。人们运用决策理论、决策分析和信息价值理论开发出了精确的数学工具来分析智能主体应该如何选择和计划。这些工具包括马尔可夫决策过程、动态决策网络、博弈论和机制设计等模型。

第419行：第365行：

The simplest AI applications can be divided into two types: classifiers ("if shiny then diamond") and controllers ("if shiny then pick up"). Controllers do, however, also classify conditions before inferring actions, and therefore classification forms a central part of many AI systems. [[Classifier (mathematics)|Classifiers]] are functions that use [[pattern matching]] to determine a closest match. They can be tuned according to examples, making them very attractive for use in AI. These examples are known as observations or patterns. In supervised learning, each pattern belongs to a certain predefined class. A class can be seen as a decision that has to be made. All the observations combined with their class labels are known as a data set. When a new observation is received, that observation is classified based on previous experience.<ref name="Classifiers"/>

−

The simplest AI applications can be divided into two types: classifiers ("if shiny then diamond") and controllers ("if shiny then pick up"). Controllers do, however, also classify conditions before inferring actions, and therefore classification forms a central part of many AI systems. Classifiers are functions that use pattern matching to determine a closest match. They can be tuned according to examples, making them very attractive for use in AI. These examples are known as observations or patterns. In supervised learning, each pattern belongs to a certain predefined class. A class can be seen as a decision that has to be made. All the observations combined with their class labels are known as a data set. When a new observation is received, that observation is classified based on previous experience.

最简单的AI应用程序可以分为两类: '''分类器 Classifiers''' (“若闪光，则为钻石”)和'''控制器 Controllers''' (“若闪光，则捡起来”)。然而，控制器在推断前也对条件进行分类，因此分类构成了许多AI系统的核心部分。分类器一组是使用匹配模式来判断最接近的匹配的函数。它们可以根据样例进行性能调优，使它们在AI应用中更有效。这些样例被称为“观察”或“模式”。在监督学习中，每个模式都属于某个预定义的类别。可以把一个类看作是一个必须做出的决定。所有的样例和它们的对应的类别标签被称为数据集。当接收一个新样例时，它会被分类器根据以前的经验进行分类。

第446行：第391行：

Neural networks were inspired by the architecture of neurons in the human brain. A simple "neuron" ''N'' accepts input from other neurons, each of which, when activated (or "fired"), cast a weighted "vote" for or against whether neuron ''N'' should itself activate. Learning requires an algorithm to adjust these weights based on the training data; one simple algorithm (dubbed "[[Hebbian learning|fire together, wire together]]") is to increase the weight between two connected neurons when the activation of one triggers the successful activation of another. The neural network forms "concepts" that are distributed among a subnetwork of shared{{efn|Each individual neuron is likely to participate in more than one concept.}} neurons that tend to fire together; a concept meaning "leg" might be coupled with a subnetwork meaning "foot" that includes the sound for "foot". Neurons have a continuous spectrum of activation; in addition, neurons can process inputs in a nonlinear way rather than weighing straightforward votes. Modern neural networks can learn both continuous functions and, surprisingly, digital logical operations. Neural networks' early successes included predicting the stock market and (in 1995) a mostly self-driving car.{{efn|Steering for the 1995 "[[History of autonomous cars#1990s|No Hands Across America]]" required "only a few human assists".}}{{sfn|Domingos|2015|loc=Chapter 4}} In the 2010s, advances in neural networks using deep learning thrust AI into widespread public consciousness and contributed to an enormous upshift in corporate AI spending; for example, AI-related [[mergers and acquisitions|M&A]] in 2017 was over 25 times as large as in 2015.<ref>{{cite news|title=Why Deep Learning Is Suddenly Changing Your Life|url=http://fortune.com/ai-artificial-intelligence-deep-machine-learning/|accessdate=12 March 2018|work=Fortune|date=2016}}</ref><ref>{{cite news|title=Google leads in the race to dominate artificial intelligence|url=https://www.economist.com/news/business/21732125-tech-giants-are-investing-billions-transformative-technology-google-leads-race|accessdate=12 March 2018|work=The Economist|date=2017|language=en}}</ref>

−

Neural networks were inspired by the architecture of neurons in the human brain. A simple "neuron" N accepts input from other neurons, each of which, when activated (or "fired"), cast a weighted "vote" for or against whether neuron N should itself activate. Learning requires an algorithm to adjust these weights based on the training data; one simple algorithm (dubbed "fire together, wire together") is to increase the weight between two connected neurons when the activation of one triggers the successful activation of another. The neural network forms "concepts" that are distributed among a subnetwork of shared neurons that tend to fire together; a concept meaning "leg" might be coupled with a subnetwork meaning "foot" that includes the sound for "foot". Neurons have a continuous spectrum of activation; in addition, neurons can process inputs in a nonlinear way rather than weighing straightforward votes. Modern neural networks can learn both continuous functions and, surprisingly, digital logical operations. Neural networks' early successes included predicting the stock market and (in 1995) a mostly self-driving car. In the 2010s, advances in neural networks using deep learning thrust AI into widespread public consciousness and contributed to an enormous upshift in corporate AI spending; for example, AI-related M&A in 2017 was over 25 times as large as in 2015.

神经网络的诞生受到人脑神经元结构的启发。一个简单的“神经元”''N'' 接受来自其他神经元的输入，每个神经元在被激活(或者说“放电”)时，都会对''N''是否应该被激活按一定的权重赋上值。学习的过程需要一个根据训练数据调整这些权重的算法：一个被称为“相互放电，彼此联系”简单的算法在一个神经元激活触发另一个神经元的激活时增加两个连接神经元之间的权重。神经网络中形成一种分布在一个共享的神经元子网络中的“概念”，这些神经元往往一起放电。“腿”的概念可能和“脚”概念的子网络相结合，后者包括”脚”的发音。神经元有一个连续的激活频谱; 此外，神经元还可以用非线性的方式处理输入，而不是简单地加权求和。现代神经网络可以学习连续函数甚至的数字逻辑运算。神经网络早期的成功包括预测股票市场和自动驾驶汽车(1995年)。{{efn|Steering for the 1995 "[[History of autonomous cars#1990s|No Hands Across America]]" required "only a few human assists".}}2010年代，神经网络使用深度学习取得巨大进步，也因此将AI推向了公众视野里，并促使企业对AI投资急速增加; 例如2017年与AI相关的并购交易规模是2015年的25倍多。<ref>{{cite news|title=Why Deep Learning Is Suddenly Changing Your Life|url=http://fortune.com/ai-artificial-intelligence-deep-machine-learning/|accessdate=12 March 2018|work=Fortune|date=2016}}</ref><ref>{{cite news|title=Google leads in the race to dominate artificial intelligence|url=https://www.economist.com/news/business/21732125-tech-giants-are-investing-billions-transformative-technology-google-leads-race|accessdate=12 March 2018|work=The Economist|date=2017|language=en}}</ref>

第452行：第396行：

The study of non-learning [[artificial neural network]]s<ref name="Neural networks"/> began in the decade before the field of AI research was founded, in the work of [[Walter Pitts]] and [[Warren McCullouch]]. [[Frank Rosenblatt]] invented the [[perceptron]], a learning network with a single layer, similar to the old concept of [[linear regression]]. Early pioneers also include [[Alexey Grigorevich Ivakhnenko]], [[Teuvo Kohonen]], [[Stephen Grossberg]], [[Kunihiko Fukushima]], [[Christoph von der Malsburg]], David Willshaw, [[Shun-Ichi Amari]], [[Bernard Widrow]], [[John Hopfield]], [[Eduardo R. Caianiello]], and others{{citation needed|date=July 2019}}.

−

The study of non-learning artificial neural networks began in the decade before the field of AI research was founded, in the work of Walter Pitts and Warren McCullouch. Frank Rosenblatt invented the perceptron, a learning network with a single layer, similar to the old concept of linear regression. Early pioneers also include Alexey Grigorevich Ivakhnenko, Teuvo Kohonen, Stephen Grossberg, Kunihiko Fukushima, Christoph von der Malsburg, David Willshaw, Shun-Ichi Amari, Bernard Widrow, John Hopfield, Eduardo R. Caianiello, and others.

沃尔特·皮茨和沃伦·麦克卢奇共同完成的非学习型人工神经网络<ref name="Neural networks"/>的研究比AI研究领域成立早十年。他们发明了'''感知机 Perceptron'''，这是一个单层的学习网络，类似于线性回归的概念。早期的先驱者还包括 Alexey Grigorevich Ivakhnenko，Teuvo Kohonen，Stephen Grossberg，Kunihiko Fukushima，Christoph von der Malsburg，David Willshaw，Shun-Ichi Amari，Bernard Widrow，John Hopfield，Eduardo r. Caianiello 等人。

第460行：第402行：

The main categories of networks are acyclic or [[feedforward neural network]]s (where the signal passes in only one direction) and [[recurrent neural network]]s (which allow feedback and short-term memories of previous input events). Among the most popular feedforward networks are [[perceptron]]s, [[multi-layer perceptron]]s and [[radial basis network]]s.<ref name="Feedforward neural networks"/> Neural networks can be applied to the problem of [[intelligent control]] (for robotics) or [[machine learning|learning]], using such techniques as [[Hebbian learning]] ("fire together, wire together"), [[GMDH]] or [[competitive learning]].<ref name="Learning in neural networks"/>

−

The main categories of networks are acyclic or feedforward neural networks (where the signal passes in only one direction) and recurrent neural networks (which allow feedback and short-term memories of previous input events). Among the most popular feedforward networks are perceptrons, multi-layer perceptrons and radial basis networks. Neural networks can be applied to the problem of intelligent control (for robotics) or learning, using such techniques as Hebbian learning ("fire together, wire together"), GMDH or competitive learning.

网络主要分为''' 非循环或前馈神经网络 Acyclic or Feedforward Neural Networks'''(信号只向一个方向传递)和'''循环神经网络 Recurrent Neural Network''' (允许反馈和对以前的输入事件进行短期记忆)。其中最常用的前馈网络.<ref name="Feedforward neural networks"/>有感知机、'''''' 和''' 径向基网络 Radial Basis Networks'''。使用'''赫布型学习 Hebbian Learning ''' (“相互放电，共同链接”) ，GMDH 或竞争学习等技术的神经网络可以被应用于智能控制(机器人)或学习问题。

第467行：第408行：

Today, neural networks are often trained by the [[backpropagation]] algorithm, which had been around since 1970 as the reverse mode of [[automatic differentiation]] published by [[Seppo Linnainmaa]],<ref name="lin1970">[[Seppo Linnainmaa]] (1970). The representation of the cumulative rounding error of an algorithm as a Taylor expansion of the local rounding errors. Master's Thesis (in Finnish), Univ. Helsinki, 6–7.</ref><ref name="grie2012">Griewank, Andreas (2012). Who Invented the Reverse Mode of Differentiation?. Optimization Stories, Documenta Matematica, Extra Volume ISMP (2012), 389–400.</ref> and was introduced to neural networks by [[Paul Werbos]].<ref name="WERBOS1974">[[Paul Werbos]], "Beyond Regression: New Tools for Prediction and Analysis in the Behavioral Sciences", ''PhD thesis, Harvard University'', 1974.</ref><ref name="werbos1982">[[Paul Werbos]] (1982). Applications of advances in nonlinear sensitivity analysis. In System modeling and optimization (pp. 762–770). Springer Berlin Heidelberg. [http://werbos.com/Neural/SensitivityIFIPSeptember1981.pdf Online] {{webarchive|url=https://web.archive.org/web/20160414055503/http://werbos.com/Neural/SensitivityIFIPSeptember1981.pdf |date=14 April 2016 }}</ref><ref name="Backpropagation"/>

−

Today, neural networks are often trained by the backpropagation algorithm, which had been around since 1970 as the reverse mode of automatic differentiation published by Seppo Linnainmaa, and was introduced to neural networks by Paul Werbos.

当下神经网络常用'''反向传播算法''' 来训练，1970年反向传播算法出现，被认为是 Seppo Linnainmaa提出的自动微分的反向模式出现<ref name="lin1970">[[Seppo Linnainmaa]] (1970). The representation of the cumulative rounding error of an algorithm as a Taylor expansion of the local rounding errors. Master's Thesis (in Finnish), Univ. Helsinki, 6–7.</ref><ref name="grie2012">Griewank, Andreas (2012). Who Invented the Reverse Mode of Differentiation?. Optimization Stories, Documenta Matematica, Extra Volume ISMP (2012), 389–400.</ref>，被保罗·韦伯引入神经网络。<ref name="WERBOS1974">[[Paul Werbos]], "Beyond Regression: New Tools for Prediction and Analysis in the Behavioral Sciences", ''PhD thesis, Harvard University'', 1974.</ref><ref name="werbos1982">[[Paul Werbos]] (1982). Applications of advances in nonlinear sensitivity analysis. In System modeling and optimization (pp. 762–770). Springer Berlin Heidelberg. [http://werbos.com/Neural/SensitivityIFIPSeptember1981.pdf Online] {{webarchive|url=https://web.archive.org/web/20160414055503/http://werbos.com/Neural/SensitivityIFIPSeptember1981.pdf |date=14 April 2016 }}</ref><ref name="Backpropagation"/>

第483行：第422行：

To summarize, most neural networks use some form of [[gradient descent]] on a hand-created neural topology. However, some research groups, such as [[Uber]], argue that simple [[neuroevolution]] to mutate new neural network topologies and weights may be competitive with sophisticated gradient descent approaches{{citation needed|date=July 2019}}. One advantage of neuroevolution is that it may be less prone to get caught in "dead ends".<ref>{{cite news|title=Artificial intelligence can 'evolve' to solve problems|url=http://www.sciencemag.org/news/2018/01/artificial-intelligence-can-evolve-solve-problems|accessdate=7 February 2018|work=Science {{!}} AAAS|date=10 January 2018|language=en}}</ref>

−

To summarize, most neural networks use some form of gradient descent on a hand-created neural topology. However, some research groups, such as Uber, argue that simple neuroevolution to mutate new neural network topologies and weights may be competitive with sophisticated gradient descent approaches. One advantage of neuroevolution is that it may be less prone to get caught in "dead ends".

总之，大多数神经网络都会在人工神经拓扑结构上使用某种形式的'''梯度下降法 Gradient Descent'''。然而，一些研究组，比如 Uber的，认为通过简单的神经进化改变新神经网络拓扑结构和神经元间的权重可能比复杂的梯度下降法更适用{{citation needed|date=July 2019}}。神经进化的一个优势是，它不容易陷入“死胡同”。<ref>{{cite news|title=Artificial intelligence can 'evolve' to solve problems|url=http://www.sciencemag.org/news/2018/01/artificial-intelligence-can-evolve-solve-problems|accessdate=7 February 2018|work=Science {{!}} AAAS|date=10 January 2018|language=en}}</ref>

第496行：第433行：

[[Deep learning]] is any [[artificial neural network]] that can learn a long chain of causal links{{dubious|date=July 2019}}. For example, a feedforward network with six hidden layers can learn a seven-link causal chain (six hidden layers + output layer) and has a [[deep learning#Credit assignment|"credit assignment path"]] (CAP) depth of seven{{citation needed|date=July 2019}}. Many deep learning systems need to be able to learn chains ten or more causal links in length.<ref name="schmidhuber2015"/> Deep learning has transformed many important subfields of artificial intelligence{{why|date=July 2019}}, including [[computer vision]], [[speech recognition]], [[natural language processing]] and others.<ref name="goodfellow2016">Ian Goodfellow, Yoshua Bengio, and Aaron Courville (2016). Deep Learning. MIT Press. [http://www.deeplearningbook.org Online] {{webarchive|url=https://web.archive.org/web/20160416111010/http://www.deeplearningbook.org/ |date=16 April 2016 }}</ref><ref name="HintonDengYu2012">{{cite journal | last1 = Hinton | first1 = G. | last2 = Deng | first2 = L. | last3 = Yu | first3 = D. | last4 = Dahl | first4 = G. | last5 = Mohamed | first5 = A. | last6 = Jaitly | first6 = N. | last7 = Senior | first7 = A. | last8 = Vanhoucke | first8 = V. | last9 = Nguyen | first9 = P. | last10 = Sainath | first10 = T. | last11 = Kingsbury | first11 = B. | year = 2012 | title = Deep Neural Networks for Acoustic Modeling in Speech Recognition – The shared views of four research groups | url = | journal = IEEE Signal Processing Magazine | volume = 29 | issue = 6| pages = 82–97 | doi=10.1109/msp.2012.2205597}}</ref><ref name="schmidhuber2015">{{cite journal |last=Schmidhuber |first=J. |year=2015 |title=Deep Learning in Neural Networks: An Overview |journal=Neural Networks |volume=61 |pages=85–117 |arxiv=1404.7828 |doi=10.1016/j.neunet.2014.09.003|pmid=25462637 }}</ref>

−

Deep learning is any artificial neural network that can learn a long chain of causal links. For example, a feedforward network with six hidden layers can learn a seven-link causal chain (six hidden layers + output layer) and has a "credit assignment path" (CAP) depth of seven. Many deep learning systems need to be able to learn chains ten or more causal links in length.

深度学习是任何可以学习长因果链的人工神经网络。例如，一个具有六个隐藏层的前馈网络可以学习有七个链接的因果链(六个隐藏层 + 一个输出层) ，并且具深度为7的“'''信用分配路径 Credit Assignment Path，CAP''' ”。许多深度学习系统需要学习长度在十及以上的因果链。<ref name="goodfellow2016">Ian Goodfellow, Yoshua Bengio, and Aaron Courville (2016). Deep Learning. MIT Press. [http://www.deeplearningbook.org Online] {{webarchive|url=https://web.archive.org/web/20160416111010/http://www.deeplearningbook.org/ |date=16 April 2016 }}</ref><ref name="HintonDengYu2012">{{cite journal | last1 = Hinton | first1 = G. | last2 = Deng | first2 = L. | last3 = Yu | first3 = D. | last4 = Dahl | first4 = G. | last5 = Mohamed | first5 = A. | last6 = Jaitly | first6 = N. | last7 = Senior | first7 = A. | last8 = Vanhoucke | first8 = V. | last9 = Nguyen | first9 = P. | last10 = Sainath | first10 = T. | last11 = Kingsbury | first11 = B. | year = 2012 | title = Deep Neural Networks for Acoustic Modeling in Speech Recognition – The shared views of four research groups | url = | journal = IEEE Signal Processing Magazine | volume = 29 | issue = 6| pages = 82–97 | doi=10.1109/msp.2012.2205597}}</ref><ref name="schmidhuber2015">{{cite journal |last=Schmidhuber |first=J. |year=2015 |title=Deep Learning in Neural Networks: An Overview |journal=Neural Networks |volume=61 |pages=85–117 |arxiv=1404.7828 |doi=10.1016/j.neunet.2014.09.003|pmid=25462637 }}</ref>

第506行：第442行： −

According to one overview,<ref name="scholarpedia">{{cite journal | last1 = Schmidhuber | first1 = Jürgen | authorlink = Jürgen Schmidhuber | year = 2015 | title = Deep Learning | journal = Scholarpedia | volume = 10 | issue = 11 | page = 32832 | doi = 10.4249/scholarpedia.32832 | df = dmy-all | bibcode = 2015SchpJ..1032832S | doi-access = free }}</ref> the expression "Deep Learning" was introduced to the [[machine learning]] community by [[Rina Dechter]] in 1986<ref name="dechter1986">[[Rina Dechter]] (1986). Learning while searching in constraint-satisfaction problems. University of California, Computer Science Department, Cognitive Systems Laboratory.[https://www.researchgate.net/publication/221605378_Learning_While_Searching_in_Constraint-Satisfaction-Problems Online] {{webarchive|url=https://web.archive.org/web/20160419054654/https://www.researchgate.net/publication/221605378_Learning_While_Searching_in_Constraint-Satisfaction-Problems |date=19 April 2016 }}</ref> and gained traction after Igor Aizenberg and colleagues introduced it to [[artificial neural network]]s in 2000.

−

According to one overview, the expression "Deep Learning" was introduced to the machine learning community by Rina Dechter in 1986 and gained traction after Igor Aizenberg and colleagues introduced it to artificial neural networks in 2000.<ref name="aizenberg2000">Igor Aizenberg, Naum N. Aizenberg, Joos P.L. Vandewalle (2000). Multi-Valued and Universal Binary Neurons: Theory, Learning and Applications. Springer Science & Business Media.</ref> The first functional Deep Learning networks were published by [[Alexey Grigorevich Ivakhnenko]] and V. G. Lapa in 1965.<ref>{{Cite book|title=Cybernetic Predicting Devices|last=Ivakhnenko|first=Alexey|publisher=Naukova Dumka|year=1965|isbn=|location=Kiev|pages=}}</ref>

−

根据一篇综述<ref name="scholarpedia">{{cite journal | last1 = Schmidhuber | first1 = Jürgen | authorlink = Jürgen Schmidhuber | year = 2015 | title = Deep Learning | journal = Scholarpedia | volume = 10 | issue = 11 | page = 32832 | doi = 10.4249/scholarpedia.32832 | df = dmy-all | bibcode = 2015SchpJ..1032832S | doi-access = free }}</ref>，“深度学习”这种表述是在1986年<ref name="dechter1986">[[Rina Dechter]] (1986). Learning while searching in constraint-satisfaction problems. University of California, Computer Science Department, Cognitive Systems Laboratory.[https://www.researchgate.net/publication/221605378_Learning_While_Searching_in_Constraint-Satisfaction-Problems Online] {{webarchive|url=https://web.archive.org/web/20160419054654/https://www.researchgate.net/publication/221605378_Learning_While_Searching_in_Constraint-Satisfaction-Problems |date=19 April 2016 }}</ref>被里纳·德克特引入到机器学习领域的，并在2000年伊克尔·艾森贝格和他的同事将其引入人工神经网络后获得了关注。<ref name="aizenberg2000">Igor Aizenberg, Naum N. Aizenberg, Joos P.L. Vandewalle (2000). Multi-Valued and Universal Binary Neurons: Theory, Learning and Applications. Springer Science & Business Media.</ref> The first functional Deep Learning networks were published by [[Alexey Grigorevich Ivakhnenko]] and V. G. Lapa in 1965.<ref>{{Cite book|title=Cybernetic Predicting Devices|last=Ivakhnenko|first=Alexey|publisher=Naukova Dumka|year=1965|isbn=|location=Kiev|pages=}}</ref>

+

根据一篇综述<ref name="scholarpedia">{{cite journal | last1 = Schmidhuber | first1 = Jürgen | authorlink = Jürgen Schmidhuber | year = 2015 | title = Deep Learning | journal = Scholarpedia | volume = 10 | issue = 11 | page = 32832 | doi = 10.4249/scholarpedia.32832 | df = dmy-all | bibcode = 2015SchpJ..1032832S | doi-access = free }}</ref>，“深度学习”这种表述是在1986年<ref name="dechter1986">[[Rina Dechter]] (1986). Learning while searching in constraint-satisfaction problems. University of California, Computer Science Department, Cognitive Systems Laboratory.[https://www.researchgate.net/publication/221605378_Learning_While_Searching_in_Constraint-Satisfaction-Problems Online]</ref>被里纳·德克特引入到机器学习领域的，并在2000年伊克尔·艾森贝格和他的同事将其引入人工神经网络后获得了关注。<ref name="aizenberg2000">Igor Aizenberg, Naum N. Aizenberg, Joos P.L. Vandewalle (2000). Multi-Valued and Universal Binary Neurons: Theory, Learning and Applications. Springer Science & Business Media.</ref> The first functional Deep Learning networks were published by [[Alexey Grigorevich Ivakhnenko]] and V. G. Lapa in 1965.<ref>{{Cite book|title=Cybernetic Predicting Devices|last=Ivakhnenko|first=Alexey|publisher=Naukova Dumka|year=1965|isbn=|location=Kiev|pages=}}</ref>

−

These networks are trained one layer at a time. Ivakhnenko's 1971 paper<ref name="ivak1971">{{Cite journal |doi = 10.1109/TSMC.1971.4308320|title = Polynomial Theory of Complex Systems|journal = IEEE Transactions on Systems, Man, and Cybernetics|issue = 4|pages = 364–378|year = 1971|last1 = Ivakhnenko|first1 = A. G.|url = https://semanticscholar.org/paper/b7efb6b6f7e9ffa017e970a098665f76d4dfeca2}}</ref> describes the learning of a deep feedforward multilayer perceptron with eight layers, already much deeper than many later networks. In 2006, a publication by [[Geoffrey Hinton]] and Ruslan Salakhutdinov introduced another way of pre-training many-layered [[feedforward neural network]]s (FNNs) one layer at a time, treating each layer in turn as an [[unsupervised learning|unsupervised]] [[restricted Boltzmann machine]], then using [[supervised learning|supervised]] [[backpropagation]] for fine-tuning.{{sfn|Hinton|2007}} Similar to shallow artificial neural networks, deep neural networks can model complex non-linear relationships. Over the last few years, advances in both machine learning algorithms and computer hardware have led to more efficient methods for training deep neural networks that contain many layers of non-linear hidden units and a very large output layer.<ref>{{cite web|last1=Research|first1=AI|title=Deep Neural Networks for Acoustic Modeling in Speech Recognition|url=http://airesearch.com/ai-research-papers/deep-neural-networks-for-acoustic-modeling-in-speech-recognition/|website=airesearch.com|accessdate=23 October 2015|date=23 October 2015}}</ref>

−

The first functional Deep Learning networks were published by Alexey Grigorevich Ivakhnenko and V. G. Lapa in 1965. These networks are trained one layer at a time. Ivakhnenko's 1971 paper describes the learning of a deep feedforward multilayer perceptron with eight layers, already much deeper than many later networks. In 2006, a publication by Geoffrey Hinton and Ruslan Salakhutdinov introduced another way of pre-training many-layered feedforward neural networks (FNNs) one layer at a time, treating each layer in turn as an unsupervised restricted Boltzmann machine, then using supervised backpropagation for fine-tuning. Similar to shallow artificial neural networks, deep neural networks can model complex non-linear relationships. Over the last few years, advances in both machine learning algorithms and computer hardware have led to more efficient methods for training deep neural networks that contain many layers of non-linear hidden units and a very large output layer.

第一个可以用的深度学习网络是由A. G.伊瓦赫年科和V.G.拉帕在1965年发表的。这些网络每次只训练一层。1971年伊瓦赫年科的论文描述了一个8层的深度前馈多层感知机网络的学习过程，这个网络已经比许多后来的网络要深得多了<ref name="ivak1971">{{Cite journal |doi = 10.1109/TSMC.1971.4308320|title = Polynomial Theory of Complex Systems|journal = IEEE Transactions on Systems, Man, and Cybernetics|issue = 4|pages = 364–378|year = 1971|last1 = Ivakhnenko|first1 = A. G.|url = https://semanticscholar.org/paper/b7efb6b6f7e9ffa017e970a098665f76d4dfeca2}}</ref>。2006年，杰弗里•辛顿和特迪诺夫的文章介绍了另一种预训练'''多层前馈神经网络 Many-layered Feedforward Neural Networks, FNNs''' 的方法，一次训练一层，将每一层都视为无监督的[[受限玻尔兹曼机]]，然后使用监督式反向传播进行微调。与浅层人工神经网络类似，深层神经网络可以模拟复杂的非线性关系。在过去的几年里，机器学习算法和计算机硬件的进步催生了更有效的方法，训练包含许多层非线性隐藏单元和一个非常大的输出层的深层神经网络。<ref>{{cite web|last1=Research|first1=AI|title=Deep Neural Networks for Acoustic Modeling in Speech Recognition|url=http://airesearch.com/ai-research-papers/deep-neural-networks-for-acoustic-modeling-in-speech-recognition/|website=airesearch.com|accessdate=23 October 2015|date=23 October 2015}}</ref>

第522行：第452行：

Deep learning often uses [[convolutional neural network]]s (CNNs), whose origins can be traced back to the [[Neocognitron]] introduced by [[Kunihiko Fukushima]] in 1980.<ref name="FUKU1980">{{cite journal | last1 = Fukushima | first1 = K. | year = 1980 | title = Neocognitron: A self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position | url = | journal = Biological Cybernetics | volume = 36 | issue = 4| pages = 193–202 | doi=10.1007/bf00344251 | pmid=7370364}}</ref> In 1989, [[Yann LeCun]] and colleagues applied [[backpropagation]] to such an architecture. In the early 2000s, in an industrial application, CNNs already processed an estimated 10% to 20% of all the checks written in the US.<ref name="lecun2016slides">[[Yann LeCun]] (2016). Slides on Deep Learning [https://indico.cern.ch/event/510372/ Online] {{webarchive|url=https://web.archive.org/web/20160423021403/https://indico.cern.ch/event/510372/ |date=23 April 2016 }}</ref>

−

Deep learning often uses convolutional neural networks (CNNs), whose origins can be traced back to the Neocognitron introduced by Kunihiko Fukushima in 1980. In 1989, Yann LeCun and colleagues applied backpropagation to such an architecture. In the early 2000s, in an industrial application, CNNs already processed an estimated 10% to 20% of all the checks written in the US.

深度学习通常使用'卷积神经网络 ConvolutionalNeural Networks CNNs''' ，其起源可以追溯到1980年由福岛邦彦引进的新认知机。1989年扬·勒丘恩(Yann LeCun)和他的同事将反向传播算法应用于这样的架构。在21世纪初，在一项工业应用中，CNNs已经处理了美国大约10% 到20%的签发支票。

−

~~Since 2011, fast implementations of CNNs on GPUs have won many visual pattern recognition competitions.<ref name="schmidhuber2015"/>~~

−

~~Since 2011, fast implementations of CNNs on GPUs have won many visual pattern recognition competitions.~~

自2011年以来，在 GPUs上快速实现的 CNN 赢得了许多视觉模式识别比赛。<ref name="schmidhuber2015"/>

−

CNNs with 12 convolutional layers were used in conjunction with [[reinforcement learning]] by Deepmind's "[[AlphaGo]] Lee", the program that beat a top [[Go (game)|Go]] champion in 2016.<ref name="Nature2017">{{cite journal |first1=David |last1=Silver|author-link1=David Silver (programmer)|first2= Julian|last2= Schrittwieser|first3= Karen|last3= Simonyan|first4= Ioannis|last4= Antonoglou|first5= Aja|last5= Huang|author-link5=Aja Huang|first6=Arthur|last6= Guez|first7= Thomas|last7= Hubert|first8= Lucas|last8= Baker|first9= Matthew|last9= Lai|first10= Adrian|last10= Bolton|first11= Yutian|last11= Chen|author-link11=Chen Yutian|first12= Timothy|last12= Lillicrap|first13=Hui|last13= Fan|author-link13=Fan Hui|first14= Laurent|last14= Sifre|first15= George van den|last15= Driessche|first16= Thore|last16= Graepel|first17= Demis|last17= Hassabis |author-link17=Demis Hassabis|title=Mastering the game of Go without human knowledge|journal=[[Nature (journal)|Nature]]|issn= 0028-0836|pages=354–359|volume =550|issue =7676|doi =10.1038/nature24270|pmid=29052630|date=19 October 2017|quote=AlphaGo Lee... 12 convolutional layers|bibcode=2017Natur.550..354S|url=http://discovery.ucl.ac.uk/10045895/1/agz_unformatted_nature.pdf}}{{closed access}}</ref>

−

~~CNNs with 12 convolutional layers were used in conjunction with reinforcement learning by Deepmind's "AlphaGo Lee", the program that beat a top Go champion in 2016.~~

2016年Deepmind 的“AlphaGo Lee”使用了有12个卷积层的 CNNs 和强化学习，击败了一个顶级围棋冠军。<ref name="Nature2017">{{cite journal |first1=David |last1=Silver|author-link1=David Silver (programmer)|first2= Julian|last2= Schrittwieser|first3= Karen|last3= Simonyan|first4= Ioannis|last4= Antonoglou|first5= Aja|last5= Huang|author-link5=Aja Huang|first6=Arthur|last6= Guez|first7= Thomas|last7= Hubert|first8= Lucas|last8= Baker|first9= Matthew|last9= Lai|first10= Adrian|last10= Bolton|first11= Yutian|last11= Chen|author-link11=Chen Yutian|first12= Timothy|last12= Lillicrap|first13=Hui|last13= Fan|author-link13=Fan Hui|first14= Laurent|last14= Sifre|first15= George van den|last15= Driessche|first16= Thore|last16= Graepel|first17= Demis|last17= Hassabis |author-link17=Demis Hassabis|title=Mastering the game of Go without human knowledge|journal=[[Nature (journal)|Nature]]|issn= 0028-0836|pages=354–359|volume =550|issue =7676|doi =10.1038/nature24270|pmid=29052630|date=19 October 2017|quote=AlphaGo Lee... 12 convolutional layers|bibcode=2017Natur.550..354S|url=http://discovery.ucl.ac.uk/10045895/1/agz_unformatted_nature.pdf}}{{closed access}}</ref>

第547行：第468行：

Early on, deep learning was also applied to sequence learning with [[recurrent neural network]]s (RNNs)<ref name="Recurrent neural networks"/> which are in theory Turing complete<ref>{{cite journal|last1=Hyötyniemi|first1=Heikki|title=Turing machines are recurrent neural networks|journal=Proceedings of STeP '96/Publications of the Finnish Artificial Intelligence Society|pages=13–24|date=1996}}</ref> and can run arbitrary programs to process arbitrary sequences of inputs. The depth of an RNN is unlimited and depends on the length of its input sequence; thus, an RNN is an example of deep learning.<ref name="schmidhuber2015"/> RNNs can be trained by [[gradient descent]]<ref>P. J. Werbos. Generalization of backpropagation with application to a recurrent gas market model" ''Neural Networks'' 1, 1988.</ref><ref>A. J. Robinson and F. Fallside. The utility driven dynamic error propagation network. Technical Report CUED/F-INFENG/TR.1, Cambridge University Engineering Department, 1987.</ref><ref>R. J. Williams and D. Zipser. Gradient-based learning algorithms for recurrent networks and their computational complexity. In Back-propagation: Theory, Architectures and Applications. Hillsdale, NJ: Erlbaum, 1994.</ref> but suffer from the [[vanishing gradient problem]].<ref name="goodfellow2016"/><ref name="hochreiter1991">[[Sepp Hochreiter]] (1991), [http://people.idsia.ch/~juergen/SeppHochreiter1991ThesisAdvisorSchmidhuber.pdf Untersuchungen zu dynamischen neuronalen Netzen] {{webarchive|url=https://web.archive.org/web/20150306075401/http://people.idsia.ch/~juergen/SeppHochreiter1991ThesisAdvisorSchmidhuber.pdf |date=6 March 2015 }}, Diploma thesis. Institut f. Informatik, Technische Univ. Munich. Advisor: J. Schmidhuber.</ref> In 1992, it was shown that unsupervised pre-training of a stack of [[recurrent neural network]]s can speed up subsequent supervised learning of deep sequential problems.<ref name="SCHMID1992">{{cite journal | last1 = Schmidhuber | first1 = J. | year = 1992 | title = Learning complex, extended sequences using the principle of history compression | url = | journal = Neural Computation | volume = 4 | issue = 2| pages = 234–242 | doi=10.1162/neco.1992.4.2.234| citeseerx = 10.1.1.49.3934}}</ref>

−

Early on, deep learning was also applied to sequence learning with recurrent neural networks (RNNs)<ref name="Recurrent neural networks"/> and can run arbitrary programs to process arbitrary sequences of inputs. The depth of an RNN is unlimited and depends on the length of its input sequence; thus, an RNN is an example of deep learning. but suffer from the vanishing gradient problem. In 1992, it was shown that unsupervised pre-training of a stack of recurrent neural networks can speed up subsequent supervised learning of deep sequential problems.

早期，深度学习也被用于'''循环神经网络 Recurrent Neural Networks,RNNs''' 的序列学习<ref>{{cite journal|last1=Hyötyniemi|first1=Heikki|title=Turing machines are recurrent neural networks|journal=Proceedings of STeP '96/Publications of the Finnish Artificial Intelligence Society|pages=13–24|date=1996}}</ref>，可以运行任意程序来处理任意的输入序列。一个循环神经网络的深度是无限制的，取决于其输入序列的长度; 因此，循环神经网络是一个深度学习的例子<ref name="schmidhuber2015"/>，但却存在梯度消失问题。1992年的一项研究表明无监督的预训练循环神经网络可以加速后续的深度序列问题的监督式学习。]<ref>P. J. Werbos. Generalization of backpropagation with application to a recurrent gas market model" ''Neural Networks'' 1, 1988.</ref><ref>A. J. Robinson and F. Fallside. The utility driven dynamic error propagation network. Technical Report CUED/F-INFENG/TR.1, Cambridge University Engineering Department, 1987.</ref><ref>R. J. Williams and D. Zipser. Gradient-based learning algorithms for recurrent networks and their computational complexity. In Back-propagation: Theory, Architectures and Applications. Hillsdale, NJ: Erlbaum, 1994.</ref> but suffer from the [[vanishing gradient problem]].<ref name="goodfellow2016"/><ref name="hochreiter1991">[[Sepp Hochreiter]] (1991), [http://people.idsia.ch/~juergen/SeppHochreiter1991ThesisAdvisorSchmidhuber.pdf Untersuchungen zu dynamischen neuronalen Netzen] {{webarchive|url=https://web.archive.org/web/20150306075401/http://people.idsia.ch/~juergen/SeppHochreiter1991ThesisAdvisorSchmidhuber.pdf |date=6 March 2015 }}, Diploma thesis. Institut f. Informatik, Technische Univ. Munich. Advisor: J. Schmidhuber.</ref> In 1992, it was shown that unsupervised pre-training of a stack of [[recurrent neural network]]s can speed up subsequent supervised learning of deep sequential problems.<ref name="SCHMID1992">{{cite journal | last1 = Schmidhuber | first1 = J. | year = 1992 | title = Learning complex, extended sequences using the principle of history compression | url = | journal = Neural Computation | volume = 4 | issue = 2| pages = 234–242 | doi=10.1162/neco.1992.4.2.234| citeseerx = 10.1.1.49.3934}}</ref>

第556行：第475行：

Numerous researchers now use variants of a deep learning recurrent NN called the [[long short-term memory]] (LSTM) network published by Hochreiter & Schmidhuber in 1997.<ref name=lstm>[[Sepp Hochreiter|Hochreiter, Sepp]]; and [[Jürgen Schmidhuber|Schmidhuber, Jürgen]]; ''Long Short-Term Memory'', Neural Computation, 9(8):1735–1780, 1997</ref> LSTM is often trained by [[Connectionist temporal classification|Connectionist Temporal Classification]] (CTC).<ref name="graves2006">Alex Graves, Santiago Fernandez, Faustino Gomez, and [[Jürgen Schmidhuber]] (2006). Connectionist temporal classification: Labelling unsegmented sequence data with recurrent neural nets. Proceedings of ICML'06, pp. 369–376.</ref> At Google, Microsoft and Baidu this approach has revolutionized [[speech recognition]].<ref name="hannun2014">{{cite arXiv

−

Numerous researchers now use variants of a deep learning recurrent NN called the long short-term memory (LSTM) network published by Hochreiter & Schmidhuber in 1997. LSTM is often trained by Connectionist Temporal Classification (CTC). At Google, Microsoft and Baidu this approach has revolutionized speech recognition.<ref name="hannun2014">{{cite arXiv

许多研究人员现在使用着一种被称为 '''长短期记忆 Long Short-term Memory, LSTM '''的网络——一种深度学习循环神经网络的变体，由霍克赖特和施米德胡贝在1997年提出。人们通常使用'''连接时序分类 Connectionist Temporal Classification, CTC'''训练LSTM<ref name="graves2006">Alex Graves, Santiago Fernandez, Faustino Gomez, and [[Jürgen Schmidhuber]] (2006). Connectionist temporal classification: Labelling unsegmented sequence data with recurrent neural nets. Proceedings of ICML'06, pp. 369–376.</ref>。谷歌，微软和百度用CTC彻底改变了语音识别。例如,2015年谷歌的语音识别性能大幅提升了49%,现在数十亿智能手机用户都可以通过谷歌声音使用这项技术。谷歌也使用LSTM来改进机器翻译，例如2015年，通过训练的LSTM，谷歌的语音识别性能大幅提升了49%，现在通过谷歌语音可以被数十亿的智能手机用户使用。谷歌还使用LSTM来改进机器翻译、语言建模和多语言语言处理。LSTM与CNNs一起使用改进了自动图像字幕的功能等众多应用。

第562行：第480行：

===评估进度===

−

~~{{Further|Progress in artificial intelligence|Competitions and prizes in artificial intelligence}}~~

AI, like electricity or the steam engine, is a general purpose technology. There is no consensus on how to characterize which tasks AI tends to excel at.<ref>{{cite news|last1=Brynjolfsson|first1=Erik|last2=Mitchell|first2=Tom|title=What can machine learning do? Workforce implications|url=http://science.sciencemag.org/content/358/6370/1530|accessdate=7 May 2018|work=Science|date=22 December 2017|pages=1530–1534|language=en|doi=10.1126/science.aap8062|bibcode=2017Sci...358.1530B}}</ref> While projects such as [[AlphaZero]] have succeeded in generating their own knowledge from scratch, many other machine learning projects require large training datasets.<ref>{{cite news|last1=Sample|first1=Ian|title='It's able to create knowledge itself': Google unveils AI that learns on its own|url=https://www.theguardian.com/science/2017/oct/18/its-able-to-create-knowledge-itself-google-unveils-ai-learns-all-on-its-own|accessdate=7 May 2018|work=the Guardian|date=18 October 2017|language=en}}</ref><ref>{{cite news|title=The AI revolution in science|url=http://www.sciencemag.org/news/2017/07/ai-revolution-science|accessdate=7 May 2018|work=Science {{!}} AAAS|date=5 July 2017|language=en}}</ref> Researcher [[Andrew Ng]] has suggested, as a "highly imperfect rule of thumb", that "almost anything a typical human can do with less than one second of mental thought, we can probably now or in the near future automate using AI."<ref>{{cite news|title=Will your job still exist in 10 years when the robots arrive?|url=http://www.scmp.com/tech/innovation/article/2098164/robots-are-coming-here-are-some-jobs-wont-exist-10-years|accessdate=7 May 2018|work=[[South China Morning Post]]|date=2017|language=en}}</ref> [[Moravec's paradox]] suggests that AI lags humans at many tasks that the human brain has specifically evolved to perform well.<ref name="The Economist"/>

−

AI, like electricity or the steam engine, is a general purpose technology. There is no consensus on how to characterize which tasks AI tends to excel at. While projects such as AlphaZero have succeeded in generating their own knowledge from scratch, many other machine learning projects require large training datasets. Researcher Andrew Ng has suggested, as a "highly imperfect rule of thumb", that "almost anything a typical human can do with less than one second of mental thought, we can probably now or in the near future automate using AI." Moravec's paradox suggests that AI lags humans at many tasks that the human brain has specifically evolved to perform well.

AI和电或蒸汽机一样，是一种通用技术。AI 擅长什么样的任务，这个问题尚未达成共识<ref>{{cite news|last1=Brynjolfsson|first1=Erik|last2=Mitchell|first2=Tom|title=What can machine learning do? Workforce implications|url=http://science.sciencemag.org/content/358/6370/1530|accessdate=7 May 2018|work=Science|date=22 December 2017|pages=1530–1534|language=en|doi=10.1126/science.aap8062|bibcode=2017Sci...358.1530B}}</ref>。虽然像 AlphaZero 这样的项目已经能做到从零开始产生知识，但是许多其他的机器学习项目仍需要大量的训练数据集<ref>{{cite news|last1=Sample|first1=Ian|title='It's able to create knowledge itself': Google unveils AI that learns on its own|url=https://www.theguardian.com/science/2017/oct/18/its-able-to-create-knowledge-itself-google-unveils-ai-learns-all-on-its-own|accessdate=7 May 2018|work=the Guardian|date=18 October 2017|language=en}}</ref><ref>{{cite news|title=The AI revolution in science|url=http://www.sciencemag.org/news/2017/07/ai-revolution-science|accessdate=7 May 2018|work=Science {{!}} AAAS|date=5 July 2017|language=en}}</ref>。研究人员吴恩达认为，作为一个“极不完美的经验法则”，“几乎任何普通人只需要不到一秒钟的思考就能做到的事情，我们现在或者在不久的将来都可以使用AI做到。”莫拉维克悖论表明，AI在执行许多人类大脑专门进化出来的、能够很好完成的任务时表现不如人类。<ref>{{cite news|title=Will your job still exist in 10 years when the robots arrive?|url=http://www.scmp.com/tech/innovation/article/2098164/robots-are-coming-here-are-some-jobs-wont-exist-10-years|accessdate=7 May 2018|work=[[South China Morning Post]]|date=2017|language=en}}</ref> [[Moravec's paradox]] suggests that AI lags humans at many tasks that the human brain has specifically evolved to perform well.<ref name="The Economist"/>

第574行：第489行：

Games provide a well-publicized benchmark for assessing rates of progress. [[AlphaGo]] around 2016 brought the era of classical board-game benchmarks to a close. Games of imperfect knowledge provide new challenges to AI in the area of [[game theory]].<ref>{{cite news|last1=Borowiec|first1=Tracey Lien, Steven|title=AlphaGo beats human Go champ in milestone for artificial intelligence|url=https://www.latimes.com/world/asia/la-fg-korea-alphago-20160312-story.html|accessdate=7 May 2018|work=latimes.com|date=2016}}</ref><ref>{{cite news|last1=Brown|first1=Noam|last2=Sandholm|first2=Tuomas|title=Superhuman AI for heads-up no-limit poker: Libratus beats top professionals|url=http://science.sciencemag.org/content/359/6374/418|accessdate=7 May 2018|work=Science|date=26 January 2018|pages=418–424|language=en|doi=10.1126/science.aao1733}}</ref> [[Esports|E-sports]] such as [[StarCraft]] continue to provide additional public benchmarks.<ref>{{cite journal|last1=Ontanon|first1=Santiago|last2=Synnaeve|first2=Gabriel|last3=Uriarte|first3=Alberto|last4=Richoux|first4=Florian|last5=Churchill|first5=David|last6=Preuss|first6=Mike|title=A Survey of Real-Time Strategy Game AI Research and Competition in StarCraft|journal=IEEE Transactions on Computational Intelligence and AI in Games|date=December 2013|volume=5|issue=4|pages=293–311|doi=10.1109/TCIAIG.2013.2286295|citeseerx=10.1.1.406.2524}}</ref><ref>{{cite news|title=Facebook Quietly Enters StarCraft War for AI Bots, and Loses|url=https://www.wired.com/story/facebook-quietly-enters-starcraft-war-for-ai-bots-and-loses/|accessdate=7 May 2018|work=WIRED|date=2017}}</ref> There are many competitions and prizes, such as the [[ImageNet|Imagenet Challenge]], to promote research in artificial intelligence. The most common areas of competition include general machine intelligence, conversational behavior, data-mining, [[autonomous car|robotic cars]], and robot soccer as well as conventional games.<ref>{{Cite web|url=http://image-net.org/challenges/LSVRC/2017/|title=ILSVRC2017|website=image-net.org|language=en|access-date=2018-11-06}}</ref>

−

Games provide a well-publicized benchmark for assessing rates of progress. AlphaGo around 2016 brought the era of classical board-game benchmarks to a close. Games of imperfect knowledge provide new challenges to AI in the area of game theory. E-sports such as StarCraft continue to provide additional public benchmarks. There are many competitions and prizes, such as the Imagenet Challenge, to promote research in artificial intelligence. The most common areas of competition include general machine intelligence, conversational behavior, data-mining, robotic cars, and robot soccer as well as conventional games.

+

游戏是评估进步率用的一个广泛认可的基准。2016年前后，AlphaGo 为传统棋类基准的时代的拉下终幕。不过，不完全知识的游戏给AI在博弈论领域提出了新的挑战。星际争霸等电子竞技现在仍然是一项的公众基准。现在出现了设立了有许多如 Imagenet 挑战赛的比赛和奖项以促进AI研究。最常见的比赛内容包括通用机器智能、对话行为、数据挖掘、机器人汽车、机器人足球以及传统游戏。

第581行：第496行：

The "imitation game" (an interpretation of the 1950 [[Turing test]] that assesses whether a computer can imitate a human) is nowadays considered too exploitable to be a meaningful benchmark.<ref>{{cite journal|last1=Schoenick|first1=Carissa|last2=Clark|first2=Peter|last3=Tafjord|first3=Oyvind|last4=Turney|first4=Peter|last5=Etzioni|first5=Oren|title=Moving beyond the Turing Test with the Allen AI Science Challenge|journal=Communications of the ACM|date=23 August 2017|volume=60|issue=9|pages=60–64|doi=10.1145/3122814|arxiv=1604.04315}}</ref> A derivative of the Turing test is the Completely Automated Public Turing test to tell Computers and Humans Apart ([[CAPTCHA]]). As the name implies, this helps to determine that a user is an actual person and not a computer posing as a human. In contrast to the standard Turing test, CAPTCHA is administered by a machine and targeted to a human as opposed to being administered by a human and targeted to a machine. A computer asks a user to complete a simple test then generates a grade for that test. Computers are unable to solve the problem, so correct solutions are deemed to be the result of a person taking the test. A common type of CAPTCHA is the test that requires the typing of distorted letters, numbers or symbols that appear in an image undecipherable by a computer.{{sfn|O'Brien|Marakas|2011}}

−

The "imitation game" (an interpretation of the 1950 Turing test that assesses whether a computer can imitate a human) is nowadays considered too exploitable to be a meaningful benchmark. A derivative of the Turing test is the Completely Automated Public Turing test to tell Computers and Humans Apart (CAPTCHA). As the name implies, this helps to determine that a user is an actual person and not a computer posing as a human. In contrast to the standard Turing test, CAPTCHA is administered by a machine and targeted to a human as opposed to being administered by a human and targeted to a machine. A computer asks a user to complete a simple test then generates a grade for that test. Computers are unable to solve the problem, so correct solutions are deemed to be the result of a person taking the test. A common type of CAPTCHA is the test that requires the typing of distorted letters, numbers or symbols that appear in an image undecipherable by a computer.

+

“模仿游戏”(对1950年图灵测试的一种解释，用来评估计算机是否可以模仿人类)如今被认为过于灵活，所以不能成为有一项意义的基准。图灵测试衍生出了'''验证码 Completely Automated Public Turing test to tell Computers and Humans Apart,CAPTCHA'''(即全自动区分计算机和人类的图灵测试)，顾名思义，这有助于确定用户是一个真实的人，而不是一台伪装成人的计算机。与标准的图灵测试不同，CAPTCHA 是由机器控制，面向人测试，而不是由人控制的，面向机器测试的。计算机要求用户完成一个简单的测试，然后给测试评出一个等级。计算机无法解决这个问题，所以一般认为只有人参加测试才能得出正确答案。验证码的一个常见类型是要求输入一幅计算机无法破译的图中扭曲的字母，数字或符号测试。

−

Proposed "universal intelligence" tests aim to compare how well machines, humans, and even non-human animals perform on problem sets that are generic as possible. At an extreme, the test suite can contain every possible problem, weighted by [[Kolmogorov complexity]]; unfortunately, these problem sets tend to be dominated by impoverished pattern-matching exercises where a tuned AI can easily exceed human performance levels.<ref name="Mathematical definitions of intelligence"/><ref>{{cite journal|last1=Hernández-Orallo|first1=José|last2=Dowe|first2=David L.|last3=Hernández-Lloreda|first3=M.Victoria|title=Universal psychometrics: Measuring cognitive abilities in the machine kingdom|journal=Cognitive Systems Research|date=March 2014|volume=27|pages=50–74|doi=10.1016/j.cogsys.2013.06.001|hdl=10251/50244|hdl-access=free}}</ref>

−

Proposed "universal intelligence" tests aim to compare how well machines, humans, and even non-human animals perform on problem sets that are generic as possible. At an extreme, the test suite can contain every possible problem, weighted by Kolmogorov complexity; unfortunately, these problem sets tend to be dominated by impoverished pattern-matching exercises where a tuned AI can easily exceed human performance levels.

“通用智能”测试旨在比较机器、人类甚至非人类动物在尽可能通用的问题集上的表现。在极端情况下，测试集可以包含所有可能出现的问题，再通过柯尔莫哥洛夫复杂度赋予权重；可是这些问题集里大多数问题都是不怎么难的模式匹配练习，在这些练习中，优化过的AI可以轻易地超过人类。<ref name="Mathematical definitions of intelligence"/><ref>{{cite journal|last1=Hernández-Orallo|first1=José|last2=Dowe|first2=David L.|last3=Hernández-Lloreda|first3=M.Victoria|title=Universal psychometrics: Measuring cognitive abilities in the machine kingdom|journal=Cognitive Systems Research|date=March 2014|volume=27|pages=50–74|doi=10.1016/j.cogsys.2013.06.001|hdl=10251/50244|hdl-access=free}}</ref>

第596行：第507行：

[[File:Automated online assistant.png|thumb|An [[automated online assistant]] providing customer service on a web page – one of many very primitive applications of artificial intelligence]AI的初级应用之一：提供客户服务的网页自动化助理] ]

−

~~An [[automated online assistant providing customer service on a web page – one of many very primitive applications of artificial intelligence]]~~

−

~~{{Main|Applications of artificial intelligence}}~~

−

AI is relevant to any intellectual task.{{sfn|Russell|Norvig|2009|p=1}} Modern artificial intelligence techniques are pervasive<ref name=":1">{{Cite book|last=|first=|url=https://ec.europa.eu/info/sites/info/files/commission-white-paper-artificial-intelligence-feb2020_en.pdf|title=White Paper: On Artificial Intelligence - A European approach to excellence and trust|publisher=European Commission|year=2020|isbn=|location=Brussels|pages=1}}</ref> and are too numerous to list here. Frequently, when a technique reaches mainstream use, it is no longer considered artificial intelligence; this phenomenon is described as the [[AI effect]].{{sfn|''CNN''|2006}}

−

AI is relevant to any intellectual task. Modern artificial intelligence techniques are pervasive and are too numerous to list here. Frequently, when a technique reaches mainstream use, it is no longer considered artificial intelligence; this phenomenon is described as the AI effect.

AI与任何智力任务都息息相关{{sfn|Russell|Norvig|2009|p=1}}。现代AI技术无处不在<ref name=":1">{{Cite book|last=|first=|url=https://ec.europa.eu/info/sites/info/files/commission-white-paper-artificial-intelligence-feb2020_en.pdf|title=White Paper: On Artificial Intelligence - A European approach to excellence and trust|publisher=European Commission|year=2020|isbn=|location=Brussels|pages=1}}</ref> ，数量众多，无法在此列举。通常，当一种技术变成主流应用时，它就不再被认为是AI; 这种现象被称为AI效应。{{sfn|''CNN''|2006}}

−

High-profile examples of AI include autonomous vehicles (such as [[Unmanned aerial vehicle|drones]] and [[self-driving cars]]), medical diagnosis, creating art (such as poetry), proving mathematical theorems, playing games (such as Chess or Go), search engines (such as [[Google search]]), online assistants (such as [[Siri]]), image recognition in photographs, spam filtering, predicting flight delays,<ref>[https://ishti.org/2018/11/19/using-artificial-intelligence-to-predict-flight-delays/ Using AI to predict flight delays], Ishti.org.</ref> prediction of judicial decisions,<ref name="ecthr2016">{{cite journal |author1=N. Aletras |author2=D. Tsarapatsanis |author3=D. Preotiuc-Pietro |author4=V. Lampos |title=Predicting judicial decisions of the European Court of Human Rights: a Natural Language Processing perspective |journal=PeerJ Computer Science |volume=2 |pages=e93 |year=2016 |df=dmy-all |doi=10.7717/peerj-cs.93 |doi-access=free }}</ref> targeting online advertisements, {{sfn|Russell|Norvig|2009|p=1}}<ref>{{cite news|title=The Economist Explains: Why firms are piling into artificial intelligence|url=https://www.economist.com/blogs/economist-explains/2016/04/economist-explains|accessdate=19 May 2016|work=[[The Economist]]|date=31 March 2016|url-status=live|archiveurl=https://web.archive.org/web/20160508010311/http://www.economist.com/blogs/economist-explains/2016/04/economist-explains|archivedate=8 May 2016|df=dmy-all}}</ref><ref>{{cite news|url=https://www.nytimes.com/2016/02/29/technology/the-promise-of-artificial-intelligence-unfolds-in-small-steps.html|title=The Promise of Artificial Intelligence Unfolds in Small Steps|last=Lohr|first=Steve|work=[[The New York Times]]|date=28 February 2016|accessdate=29 February 2016|url-status=live|archiveurl=https://web.archive.org/web/20160229171843/http://www.nytimes.com/2016/02/29/technology/the-promise-of-artificial-intelligence-unfolds-in-small-steps.html|archivedate=29 February 2016|df=dmy-all}}</ref> and [[energy storage]]<ref>{{Cite web|url=https://www.cnbc.com/2019/06/14/the-business-using-ai-to-change-how-we-think-about-energy-storage.html|title=A Californian business is using A.I. to change the way we think about energy storage|last=Frangoul|first=Anmar|date=2019-06-14|website=CNBC|language=en|access-date=2019-11-05}}</ref>

−

High-profile examples of AI include autonomous vehicles (such as drones and self-driving cars), medical diagnosis, creating art (such as poetry), proving mathematical theorems, playing games (such as Chess or Go), search engines (such as Google search), online assistants (such as Siri), image recognition in photographs, spam filtering, predicting flight delays, prediction of judicial decisions, targeting online advertisements, and energy storage

大众常见的AI包括自动驾驶(如无人机和自动驾驶汽车)、医疗诊断、艺术创作(如诗歌)、证明数学定理、玩游戏(如国际象棋或围棋)、搜索引擎(如谷歌搜索)、在线助手(如 Siri)、图像识别、垃圾邮件过滤、航班延误预测<ref>[https://ishti.org/2018/11/19/using-artificial-intelligence-to-predict-flight-delays/ Using AI to predict flight delays], Ishti.org.</ref> 、司法判决预测<ref name="ecthr2016">{{cite journal |author1=N. Aletras |author2=D. Tsarapatsanis |author3=D. Preotiuc-Pietro |author4=V. Lampos |title=Predicting judicial decisions of the European Court of Human Rights: a Natural Language Processing perspective |journal=PeerJ Computer Science |volume=2 |pages=e93 |year=2016 |df=dmy-all |doi=10.7717/peerj-cs.93 |doi-access=free }}</ref> 、投放在线广告{{sfn|Russell|Norvig|2009|p=1}}<ref>{{cite news|title=The Economist Explains: Why firms are piling into artificial intelligence|url=https://www.economist.com/blogs/economist-explains/2016/04/economist-explains|accessdate=19 May 2016|work=[[The Economist]]|date=31 March 2016|url-status=live|archiveurl=https://web.archive.org/web/20160508010311/http://www.economist.com/blogs/economist-explains/2016/04/economist-explains|archivedate=8 May 2016|df=dmy-all}}</ref><ref>{{cite news|url=https://www.nytimes.com/2016/02/29/technology/the-promise-of-artificial-intelligence-unfolds-in-small-steps.html|title=The Promise of Artificial Intelligence Unfolds in Small Steps|last=Lohr|first=Steve|work=[[The New York Times]]|date=28 February 2016|accessdate=29 February 2016|url-status=live|archiveurl=https://web.archive.org/web/20160229171843/http://www.nytimes.com/2016/02/29/technology/the-promise-of-artificial-intelligence-unfolds-in-small-steps.html|archivedate=29 February 2016|df=dmy-all}}</ref>和能源储存<ref>{{Cite web|url=https://www.cnbc.com/2019/06/14/the-business-using-ai-to-change-how-we-think-about-energy-storage.html|title=A Californian business is using A.I. to change the way we think about energy storage|last=Frangoul|first=Anmar|date=2019-06-14|website=CNBC|language=en|access-date=2019-11-05}}</ref>。

−

With social media sites overtaking TV as a source for news for young people and news organizations increasingly reliant on social media platforms for generating distribution,<ref>{{cite web|url=https://www.bbc.co.uk/news/uk-36528256|title=Social media 'outstrips TV' as news source for young people|date=15 June 2016|author=Wakefield, Jane|work=BBC News|url-status=live|archiveurl=https://web.archive.org/web/20160624000744/http://www.bbc.co.uk/news/uk-36528256|archivedate=24 June 2016|df=dmy-all}}</ref> major publishers now use artificial intelligence (AI) technology to post stories more effectively and generate higher volumes of traffic.<ref>{{cite web|url=https://www.bbc.co.uk/news/business-36837824|title=So you think you chose to read this article?|date=22 July 2016|author=Smith, Mark|work=BBC News|url-status=live|archiveurl=https://web.archive.org/web/20160725205007/http://www.bbc.co.uk/news/business-36837824|archivedate=25 July 2016|df=dmy-all}}</ref>

−

With social media sites overtaking TV as a source for news for young people and news organizations increasingly reliant on social media platforms for generating distribution, major publishers now use artificial intelligence (AI) technology to post stories more effectively and generate higher volumes of traffic.

随着社交媒体网站取代电视成为年轻人获取新闻的来源，以及新闻机构越来越依赖社交媒体平台来发布新闻<ref>{{cite web|url=https://www.bbc.co.uk/news/uk-36528256|title=Social media 'outstrips TV' as news source for young people|date=15 June 2016|author=Wakefield, Jane|work=BBC News|url-status=live|archiveurl=https://web.archive.org/web/20160624000744/http://www.bbc.co.uk/news/uk-36528256|archivedate=24 June 2016|df=dmy-all}}</ref>，大型出版商现在使用AI技术发布新闻，这样做效率更高且能带来更多的流量<ref>{{cite web|url=https://www.bbc.co.uk/news/business-36837824|title=So you think you chose to read this article?|date=22 July 2016|author=Smith, Mark|work=BBC News|url-status=live|archiveurl=https://web.archive.org/web/20160725205007/http://www.bbc.co.uk/news/business-36837824|archivedate=25 July 2016|df=dmy-all}}</ref>。

−

AI can also produce [[Deepfake]]s, a content-altering technology. ZDNet reports, "It presents something that did not actually occur," Though 88% of Americans believe Deepfakes can cause more harm than good, only 47% of them believe they can be targeted. The boom of election year also opens public discourse to threats of videos of falsified politician media.<ref>{{Cite web|url=https://www.zdnet.com/article/half-of-americans-do-not-believe-deepfake-news-could-target-them-online/|title=Half of Americans do not believe deepfake news could target them online|last=Brown|first=Eileen|website=ZDNet|language=en|access-date=2019-12-03}}</ref>

−

AI can also produce Deepfakes, a content-altering technology. ZDNet reports, "It presents something that did not actually occur," Though 88% of Americans believe Deepfakes can cause more harm than good, only 47% of them believe they can be targeted. The boom of election year also opens public discourse to threats of videos of falsified politician media.

AI还可以用来生成“深度虚假（DeepFake）”，这是一种内容改变技术。至顶网报道说，“它展示出一些并没有真正发生的事情。”尽管88% 的美国人认为换脸弊大于利，但只有47%的人认为自己会成为换脸对象。选举年的盛况也让公众开始讨论起虚假政治视频的害处<ref>{{Cite web|url=https://www.zdnet.com/article/half-of-americans-do-not-believe-deepfake-news-could-target-them-online/|title=Half of Americans do not believe deepfake news could target them online|last=Brown|first=Eileen|website=ZDNet|language=en|access-date=2019-12-03}}</ref>。

第635行：第525行：

[[File:Laproscopic Surgery Robot.jpg|thumb| A patient-side surgical arm of [[Da Vinci Surgical System]]]]

−

AI in healthcare is often used for classification, whether to automate initial evaluation of a CT scan or EKG or to identify high-risk patients for population health. The breadth of applications is rapidly increasing.

在医疗保健中，AI通常被用于分类，它既可以自动对 CT 扫描或心电图EKG进行初步评估，又可以在人口健康调查中识别高风险患者。AI的应用范围正在迅速扩大。

−

As an example, AI is being applied to the high-cost problem of dosage issues—where findings suggested that AI could save $16 billion. In 2016, a groundbreaking study in California found that a mathematical formula developed with the help of AI correctly determined the accurate dose of immunosuppressant drugs to give to organ patients.<ref>{{Cite news|url=https://hbr.org/2018/05/10-promising-ai-applications-in-health-care|title=10 Promising AI Applications in Health Care|date=2018-05-10|work=Harvard Business Review|access-date=2018-08-28|archive-url=https://web.archive.org/web/20181215015645/https://hbr.org/2018/05/10-promising-ai-applications-in-health-care|archive-date=15 December 2018|url-status=dead}}</ref>

[[File:X-ray of a hand with automatic bone age calculation.jpg|thumb|[[Projectional radiography|X-ray]] of a hand, with automatic calculation of [[bone age]] by computer software，一只手的X光射线图，自动计算了骨龄]]

−

As an example, AI is being applied to the high-cost problem of dosage issues—where findings suggested that AI could save $16 billion. In 2016, a groundbreaking study in California found that a mathematical formula developed with the help of AI correctly determined the accurate dose of immunosuppressant drugs to give to organ patients. X-ray of a hand, with automatic calculation of bone age by computer software]]

例如研究结果表明，AI在成本高昂的剂量问题上可以节省160亿美元。2016年，加利福尼亚州的一项开创性研究报道，在AI的辅助下得到的一个数学公式，给出了器官患者免疫抑制药的准确剂量<ref>{{Cite news|url=https://hbr.org/2018/05/10-promising-ai-applications-in-health-care|title=10 Promising AI Applications in Health Care|date=2018-05-10|work=Harvard Business Review|access-date=2018-08-28|archive-url=https://web.archive.org/web/20181215015645/https://hbr.org/2018/05/10-promising-ai-applications-in-health-care|archive-date=15 December 2018|url-status=dead}}</ref> 。

--[[用户:Thingamabob|Thingamabob]]（[[用户讨论:Thingamabob|讨论]]）例如研究结果表明，AI在高成本的剂量问题上可以节省160亿美元。为省译

−

Artificial intelligence is assisting doctors. According to Bloomberg Technology, Microsoft has developed AI to help doctors find the right treatments for cancer.<ref>{{cite news | author=Dina Bass | title=Microsoft Develops AI to Help Cancer Doctors Find the Right Treatments | url=https://www.bloomberg.com/news/articles/2016-09-20/microsoft-develops-ai-to-help-cancer-doctors-find-the-right-treatments | date=20 September 2016 | publisher=Bloomberg | url-status=live | archiveurl=https://web.archive.org/web/20170511103625/https://www.bloomberg.com/news/articles/2016-09-20/microsoft-develops-ai-to-help-cancer-doctors-find-the-right-treatments | archivedate=11 May 2017 | df=dmy-all | newspaper=Bloomberg.com }}</ref> There is a great amount of research and drugs developed relating to cancer. In detail, there are more than 800 medicines and vaccines to treat cancer. This negatively affects the doctors, because there are too many options to choose from, making it more difficult to choose the right drugs for the patients. Microsoft is working on a project to develop a machine called "Hanover"{{citation needed|date=July 2019}}. Its goal is to memorize all the papers necessary to cancer and help predict which combinations of drugs will be most effective for each patient. One project that is being worked on at the moment is fighting [[acute myeloid leukemia|myeloid leukemia]], a fatal cancer where the treatment has not improved in decades. Another study was reported to have found that artificial intelligence was as good as trained doctors in identifying skin cancers.<ref>{{Cite news|url=https://www.bbc.co.uk/news/health-38717928|title=Artificial intelligence 'as good as cancer doctors'|last=Gallagher|first=James|date=26 January 2017|work=BBC News|language=en-GB|access-date=26 January 2017|url-status=live|archiveurl=https://web.archive.org/web/20170126133849/http://www.bbc.co.uk/news/health-38717928|archivedate=26 January 2017|df=dmy-all}}</ref> Another study is using artificial intelligence to try to monitor multiple high-risk patients, and this is done by asking each patient numerous questions based on data acquired from live doctor to patient interactions.<ref>{{Citation|title=Remote monitoring of high-risk patients using artificial intelligence|date=18 Oct 1994|url=https://www.google.com/patents/US5357427|editor-last=Langen|editor2-last=Katz|editor3-last=Dempsey|editor-first=Pauline A.|editor2-first=Jeffrey S.|editor3-first=Gayle|issue=US5357427 A|accessdate=27 February 2017|url-status=live|archiveurl=https://web.archive.org/web/20170228090520/https://www.google.com/patents/US5357427|archivedate=28 February 2017|df=dmy-all}}</ref> One study was done with transfer learning, the machine performed a diagnosis similarly to a well-trained ophthalmologist, and could generate a decision within 30 seconds on whether or not the patient should be referred for treatment, with more than 95% accuracy.<ref>{{Cite journal|url=https://www.cell.com/action/captchaChallenge?redirectUri=%2Fcell%2Fpdf%2FS0092-8674%2818%2930154-5.pdf|title=Identifying Medical Diagnoses and Treatable Diseases by Image-Based Deep Learning|last=Kermany|first=D|last2=Goldbaum|first2=M|journal=Cell|access-date=2018-12-18|last3=Zhang|first3=Kang|volume=172|issue=5|pages=1122–1131.e9|pmid=29474911|year=2018|doi=10.1016/j.cell.2018.02.010}}</ref>

−

Artificial intelligence is assisting doctors. According to Bloomberg Technology, Microsoft has developed AI to help doctors find the right treatments for cancer. There is a great amount of research and drugs developed relating to cancer. In detail, there are more than 800 medicines and vaccines to treat cancer. This negatively affects the doctors, because there are too many options to choose from, making it more difficult to choose the right drugs for the patients. Microsoft is working on a project to develop a machine called "Hanover". Its goal is to memorize all the papers necessary to cancer and help predict which combinations of drugs will be most effective for each patient. One project that is being worked on at the moment is fighting myeloid leukemia, a fatal cancer where the treatment has not improved in decades. Another study was reported to have found that artificial intelligence was as good as trained doctors in identifying skin cancers. Another study is using artificial intelligence to try to monitor multiple high-risk patients, and this is done by asking each patient numerous questions based on data acquired from live doctor to patient interactions. One study was done with transfer learning, the machine performed a diagnosis similarly to a well-trained ophthalmologist, and could generate a decision within 30 seconds on whether or not the patient should be referred for treatment, with more than 95% accuracy.

AI还能协助医生。据彭博科技报道，微软已经开发出帮助医生找到正确的癌症治疗方法的AI<ref>{{cite news | author=Dina Bass | title=Microsoft Develops AI to Help Cancer Doctors Find the Right Treatments | url=https://www.bloomberg.com/news/articles/2016-09-20/microsoft-develops-ai-to-help-cancer-doctors-find-the-right-treatments | date=20 September 2016 | publisher=Bloomberg | url-status=live | archiveurl=https://web.archive.org/web/20170511103625/https://www.bloomberg.com/news/articles/2016-09-20/microsoft-develops-ai-to-help-cancer-doctors-find-the-right-treatments | archivedate=11 May 2017 | df=dmy-all | newspaper=Bloomberg.com }}</ref>。如今有大量的研究和药物开发与癌症有关，准确来说有800多种可以治疗癌症的药物和疫苗。这对医生来说并不是一件好事，因为选项太多，使得为病人选择合适的药物变得更难。微软正在开发一种名为“汉诺威”的机器。它的目标是记住所有与癌症有关的论文，并帮助预测哪些药物的组合对病人最有效。目前正在进行的一个项目是抗击髓系白血病，这是一种致命的癌症，几十年来治疗水平一直没有提高。据报道，另一项研究发现，AI在识别皮肤癌方面与训练有素的医生一样优秀<ref>{{Cite news|url=https://www.bbc.co.uk/news/health-38717928|title=Artificial intelligence 'as good as cancer doctors'|last=Gallagher|first=James|date=26 January 2017|work=BBC News|language=en-GB|access-date=26 January 2017|url-status=live|archiveurl=https://web.archive.org/web/20170126133849/http://www.bbc.co.uk/news/health-38717928|archivedate=26 January 2017|df=dmy-all}}</ref>。另一项研究是使用AI通过询问每个高风险患者多个问题来监测他们，这些问题是基于从医生与患者的互动中获得的数据产生的<ref>{{Citation|title=Remote monitoring of high-risk patients using artificial intelligence|date=18 Oct 1994|url=https://www.google.com/patents/US5357427|editor-last=Langen|editor2-last=Katz|editor3-last=Dempsey|editor-first=Pauline A.|editor2-first=Jeffrey S.|editor3-first=Gayle|issue=US5357427 A|accessdate=27 February 2017|url-status=live|archiveurl=https://web.archive.org/web/20170228090520/https://www.google.com/patents/US5357427|archivedate=28 February 2017|df=dmy-all}}</ref>。其中一项研究是通过转移学习完成的，机器进行的诊断类似于训练有素的眼科医生，可以在30秒内做出是否应该转诊治疗的决定，准确率超过95% <ref>{{Cite journal|url=https://www.cell.com/action/captchaChallenge?redirectUri=%2Fcell%2Fpdf%2FS0092-8674%2818%2930154-5.pdf|title=Identifying Medical Diagnoses and Treatable Diseases by Image-Based Deep Learning|last=Kermany|first=D|last2=Goldbaum|first2=M|journal=Cell|access-date=2018-12-18|last3=Zhang|first3=Kang|volume=172|issue=5|pages=1122–1131.e9|pmid=29474911|year=2018|doi=10.1016/j.cell.2018.02.010}}</ref>。

−

According to [[CNN]], a recent study by surgeons at the Children's National Medical Center in Washington successfully demonstrated surgery with an autonomous robot. The team supervised the robot while it performed soft-tissue surgery, stitching together a pig's bowel during open surgery, and doing so better than a human surgeon, the team claimed.<ref>{{cite news|author=Senthilingam, Meera|title=Are Autonomous Robots Your next Surgeons?|work=CNN|publisher=Cable News Network|date=12 May 2016|accessdate=4 December 2016|url=http://www.cnn.com/2016/05/12/health/robot-surgeon-bowel-operation/|url-status=live|archiveurl=https://web.archive.org/web/20161203154119/http://www.cnn.com/2016/05/12/health/robot-surgeon-bowel-operation|archivedate=3 December 2016|df=dmy-all}}</ref> IBM has created its own artificial intelligence computer, the [[IBM Watson]], which has beaten human intelligence (at some levels). Watson has struggled to achieve success and adoption in healthcare.<ref>{{Cite web|url=https://spectrum.ieee.org/biomedical/diagnostics/how-ibm-watson-overpromised-and-underdelivered-on-ai-health-care|title=Full Page Reload|website=IEEE Spectrum: Technology, Engineering, and Science News|language=en|access-date=2019-09-03}}</ref>

−

According to CNN, a recent study by surgeons at the Children's National Medical Center in Washington successfully demonstrated surgery with an autonomous robot. The team supervised the robot while it performed soft-tissue surgery, stitching together a pig's bowel during open surgery, and doing so better than a human surgeon, the team claimed. IBM has created its own artificial intelligence computer, the IBM Watson, which has beaten human intelligence (at some levels). Watson has struggled to achieve success and adoption in healthcare.

据 CNN 报道，华盛顿国家儿童医疗中心的外科医生最近的一项研究成功演示了一台自主机器人手术。研究组观看了机器人做软组织手术、在开放手术中缝合猪肠的整个过程，并认为比人类外科医生做得更好<ref>{{cite news|author=Senthilingam, Meera|title=Are Autonomous Robots Your next Surgeons?|work=CNN|publisher=Cable News Network|date=12 May 2016|accessdate=4 December 2016|url=http://www.cnn.com/2016/05/12/health/robot-surgeon-bowel-operation/|url-status=live|archiveurl=https://web.archive.org/web/20161203154119/http://www.cnn.com/2016/05/12/health/robot-surgeon-bowel-operation|archivedate=3 December 2016|df=dmy-all}}</ref> 。IBM已经创造了自己的AI计算机——IBM 沃森，它在某种程度上已经超越了人类智能。沃森一直在努力实现医疗保健领域的应用<ref>{{Cite web|url=https://spectrum.ieee.org/biomedical/diagnostics/how-ibm-watson-overpromised-and-underdelivered-on-ai-health-care|title=Full Page Reload|website=IEEE Spectrum: Technology, Engineering, and Science News|language=en|access-date=2019-09-03}}</ref>。

第670行：第544行：

===汽车 ===

−

~~{{Main|driverless cars}}~~

−

Advancements in AI have contributed to the growth of the automotive industry through the creation and evolution of self-driving vehicles. {{as of|2016}}, there are over 30 companies utilizing AI into the creation of [[self-driving car]]s. A few companies involved with AI include [[Tesla Motors|Tesla]], [[Google]], and [[Apple Inc.|Apple]].<ref>"33 Corporations Working On Autonomous Vehicles". CB Insights. N.p., 11 August 2016. 12 November 2016.</ref>

−

Advancements in AI have contributed to the growth of the automotive industry through the creation and evolution of self-driving vehicles. , there are over 30 companies utilizing AI into the creation of self-driving cars. A few companies involved with AI include Tesla, Google, and Apple.

在AI领域，自动驾驶汽车的创造和发展促进了汽车行业的增长。目前有超过30家公司利用AI开发自动驾驶汽车，包括特斯拉、谷歌和苹果等<ref>"33 Corporations Working On Autonomous Vehicles". CB Insights. N.p., 11 August 2016. 12 November 2016.</ref>。

−

Many components contribute to the functioning of self-driving cars. These vehicles incorporate systems such as braking, lane changing, collision prevention, navigation and mapping. Together, these systems, as well as high-performance computers, are integrated into one complex vehicle.<ref>West, Darrell M. "Moving forward: Self-driving vehicles in China, Europe, Japan, Korea, and the United States". Center for Technology Innovation at Brookings. N.p., September 2016. 12 November 2016.</ref>

−

Many components contribute to the functioning of self-driving cars. These vehicles incorporate systems such as braking, lane changing, collision prevention, navigation and mapping. Together, these systems, as well as high-performance computers, are integrated into one complex vehicle.

自动驾驶汽车的功能的实现需要很多组件。这些车辆集成了诸如刹车、换车道、防撞、导航和测绘等系统。这些系统和高性能计算机一起被装配到一辆复杂的车中.<ref>West, Darrell M. "Moving forward: Self-driving vehicles in China, Europe, Japan, Korea, and the United States". Center for Technology Innovation at Brookings. N.p., September 2016. 12 November 2016.</ref>。

−

Recent developments in autonomous automobiles have made the innovation of self-driving trucks possible, though they are still in the testing phase. The UK government has passed legislation to begin testing of self-driving truck platoons in 2018.<ref>{{cite journal|last1=Burgess|first1=Matt|title=The UK is about to Start Testing Self-Driving Truck Platoons|url=https://www.wired.co.uk/article/uk-trial-self-driving-trucks-platoons-roads|journal=Wired UK|accessdate=20 September 2017|url-status=live|archiveurl=https://web.archive.org/web/20170922055917/http://www.wired.co.uk/article/uk-trial-self-driving-trucks-platoons-roads|archivedate=22 September 2017|df=dmy-all|date=2017-08-24}}</ref> Self-driving truck platoons are a fleet of self-driving trucks following the lead of one non-self-driving truck, so the truck platoons aren't entirely autonomous yet. Meanwhile, the Daimler, a German automobile corporation, is testing the Freightliner Inspiration which is a semi-autonomous truck that will only be used on the highway.<ref>{{cite journal|last1=Davies|first1=Alex|title=World's First Self-Driving Semi-Truck Hits the Road|url=https://www.wired.com/2015/05/worlds-first-self-driving-semi-truck-hits-road/|journal=WIRED|accessdate=20 September 2017|url-status=live|archiveurl=https://web.archive.org/web/20171028222802/https://www.wired.com/2015/05/worlds-first-self-driving-semi-truck-hits-road/|archivedate=28 October 2017|df=dmy-all|date=2015-05-05}}</ref>

−

Recent developments in autonomous automobiles have made the innovation of self-driving trucks possible, though they are still in the testing phase. The UK government has passed legislation to begin testing of self-driving truck platoons in 2018. Self-driving truck platoons are a fleet of self-driving trucks following the lead of one non-self-driving truck, so the truck platoons aren't entirely autonomous yet. Meanwhile, the Daimler, a German automobile corporation, is testing the Freightliner Inspiration which is a semi-autonomous truck that will only be used on the highway.

自动驾驶汽车的最新发展使自动驾驶卡车的创新成为可能，尽管它们仍处于测试阶段。英国政府已通过立法，于2018年开始测试自动驾驶卡车列队行驶<ref>{{cite journal|last1=Burgess|first1=Matt|title=The UK is about to Start Testing Self-Driving Truck Platoons|url=https://www.wired.co.uk/article/uk-trial-self-driving-trucks-platoons-roads|journal=Wired UK|accessdate=20 September 2017|url-status=live|archiveurl=https://web.archive.org/web/20170922055917/http://www.wired.co.uk/article/uk-trial-self-driving-trucks-platoons-roads|archivedate=22 September 2017|df=dmy-all|date=2017-08-24}}</ref>。自动驾驶卡车队列是指一排自动驾驶卡车跟随一辆非自动驾驶卡车，所以卡车排还不是完全自动的。与此同时，德国汽车公司戴姆勒正在测试Freightliner Inspiration，这是一种只在高速公路上行驶的半自动卡车<ref>{{cite journal|last1=Davies|first1=Alex|title=World's First Self-Driving Semi-Truck Hits the Road|url=https://www.wired.com/2015/05/worlds-first-self-driving-semi-truck-hits-road/|journal=WIRED|accessdate=20 September 2017|url-status=live|archiveurl=https://web.archive.org/web/20171028222802/https://www.wired.com/2015/05/worlds-first-self-driving-semi-truck-hits-road/|archivedate=28 October 2017|df=dmy-all|date=2015-05-05}}</ref>。

−

One main factor that influences the ability for a driver-less automobile to function is mapping. In general, the vehicle would be pre-programmed with a map of the area being driven. This map would include data on the approximations of street light and curb heights in order for the vehicle to be aware of its surroundings. However, Google has been working on an algorithm with the purpose of eliminating the need for pre-programmed maps and instead, creating a device that would be able to adjust to a variety of new surroundings.<ref>McFarland, Matt. "Google's artificial intelligence breakthrough may have a huge impact on self-driving cars and much more". ''The Washington Post'' 25 February 2015. Infotrac Newsstand. 24 October 2016</ref> Some self-driving cars are not equipped with steering wheels or brake pedals, so there has also been research focused on creating an algorithm that is capable of maintaining a safe environment for the passengers in the vehicle through awareness of speed and driving conditions.<ref>"Programming safety into self-driving cars". National Science Foundation. N.p., 2 February 2015. 24 October 2016.</ref>

−

One main factor that influences the ability for a driver-less automobile to function is mapping. In general, the vehicle would be pre-programmed with a map of the area being driven. This map would include data on the approximations of street light and curb heights in order for the vehicle to be aware of its surroundings. However, Google has been working on an algorithm with the purpose of eliminating the need for pre-programmed maps and instead, creating a device that would be able to adjust to a variety of new surroundings. Some self-driving cars are not equipped with steering wheels or brake pedals, so there has also been research focused on creating an algorithm that is capable of maintaining a safe environment for the passengers in the vehicle through awareness of speed and driving conditions.

影响无人驾驶汽车性能的一个主要因素是地图。一般来说，一张行驶区域的地图会被预先写入车辆中。这张地图将包括街灯和路缘高度的近似数据，让车辆能够感知周围环境。然而谷歌一直在研究一种不需要预编程地图的算法，创造一种能够适应各种新环境的设备<ref>McFarland, Matt. "Google's artificial intelligence breakthrough may have a huge impact on self-driving cars and much more". ''The Washington Post'' 25 February 2015. Infotrac Newsstand. 24 October 2016</ref>。一些自动驾驶汽车没有配备方向盘或刹车踏板，因此也有研究致力于创建感知速度和驾驶条件的算法，为车内乘客提供一个安全的环境<ref>"Programming safety into self-driving cars". National Science Foundation. N.p., 2 February 2015. 24 October 2016.</ref>。

−

Another factor that is influencing the ability of a driver-less automobile is the safety of the passenger. To make a driver-less automobile, engineers must program it to handle high-risk situations. These situations could include a head-on collision with pedestrians. The car's main goal should be to make a decision that would avoid hitting the pedestrians and saving the passengers in the car. But there is a possibility the car would need to make a decision that would put someone in danger. In other words, the car would need to decide to save the pedestrians or the passengers.<ref>ArXiv, E. T. (26 October 2015). Why Self-Driving Cars Must Be Programmed to Kill. Retrieved 17 November 2017, from https://www.technologyreview.com/s/542626/why-self-driving-cars-must-be-programmed-to-kill/{{Dead link|date=October 2019 |bot=InternetArchiveBot |fix-attempted=yes }}</ref> The programming of the car in these situations is crucial to a successful driver-less automobile.

−

Another factor that is influencing the ability of a driver-less automobile is the safety of the passenger. To make a driver-less automobile, engineers must program it to handle high-risk situations. These situations could include a head-on collision with pedestrians. The car's main goal should be to make a decision that would avoid hitting the pedestrians and saving the passengers in the car. But there is a possibility the car would need to make a decision that would put someone in danger. In other words, the car would need to decide to save the pedestrians or the passengers. The programming of the car in these situations is crucial to a successful driver-less automobile.

衡量无人驾驶汽车能力的另一个因素是乘客的安全。工程师们必须对无人驾驶汽车进行编程，使其能够处理比如与行人正面相撞等高风险情况。这辆车的主要目标应该是做出一个避免撞到行人，保护车内的乘客的决定。但是有时汽车也可能也会不得不决定将某人置于危险之中。也就是说，汽车需要决定是拯救行人还是乘客<ref>ArXiv, E. T. (26 October 2015). Why Self-Driving Cars Must Be Programmed to Kill. Retrieved 17 November 2017, from https://www.technologyreview.com/s/542626/why-self-driving-cars-must-be-programmed-to-kill/{{Dead link|date=October 2019 |bot=InternetArchiveBot |fix-attempted=yes }}</ref>。汽车在这些情况下的编程对于一辆成功的无人驾驶汽车是至关重要的。

第713行：第562行：

===金融和经济 ===

−

[[Financial institution]]s have long used [[artificial neural network]] systems to detect charges or claims outside of the norm, flagging these for human investigation. The use of AI in [[banking]] can be traced back to 1987 when [[Security Pacific National Bank]] in US set-up a Fraud Prevention Task force to counter the unauthorized use of debit cards.<ref>{{Cite web|url=https://www.latimes.com/archives/la-xpm-1990-01-17-fi-233-story.html|title=Impact of Artificial Intelligence on Banking|last=Christy|first=Charles A.|website=latimes.com|access-date=2019-09-10|date=17 January 1990}}</ref> Programs like Kasisto and Moneystream are using AI in financial services.

−

Financial institutions have long used artificial neural network systems to detect charges or claims outside of the norm, flagging these for human investigation. The use of AI in banking can be traced back to 1987 when Security Pacific National Bank in US set-up a Fraud Prevention Task force to counter the unauthorized use of debit cards. Programs like Kasisto and Moneystream are using AI in financial services.

长期以来，金融机构一直使用人工神经网络系统来检测超出常规的费用或申诉，并将其标记起来等待人工调查。AI在银行业的应用可以追溯到1987年，当时美国国家安全太平洋银行成立了一个防防诈特别小组，以打击未经授权使用借记卡的行为<ref>{{Cite web|url=https://www.latimes.com/archives/la-xpm-1990-01-17-fi-233-story.html|title=Impact of Artificial Intelligence on Banking|last=Christy|first=Charles A.|website=latimes.com|access-date=2019-09-10|date=17 January 1990}}</ref>。Kasisto 和 Moneystream等程序正在把AI技术使用到金融服务领域。

−

Banks use artificial intelligence systems today to organize operations, maintain book-keeping, invest in stocks, and manage properties. AI can react to changes overnight or when business is not taking place.<ref name="Eleanor">{{cite web|url=https://www.icas.com/ca-today-news/how-accountancy-and-finance-are-using-artificial-intelligence|title=Accounting, automation and AI|first=Eleanor|last=O'Neill|website=icas.com|language=English|date=31 July 2016|access-date=18 November 2016|url-status=live|archiveurl=https://web.archive.org/web/20161118165901/https://www.icas.com/ca-today-news/how-accountancy-and-finance-are-using-artificial-intelligence|archivedate=18 November 2016|df=dmy-all}}</ref> In August 2001, robots beat humans in a simulated [[stock trader|financial trading]] competition.<ref>[http://news.bbc.co.uk/2/hi/business/1481339.stm Robots Beat Humans in Trading Battle.] {{webarchive|url=https://web.archive.org/web/20090909001249/http://news.bbc.co.uk/2/hi/business/1481339.stm |date=9 September 2009 }} BBC.com (8 August 2001)</ref> AI has also reduced fraud and financial crimes by [[Statistical software|monitoring]] [[behavioral pattern]]s of users for any abnormal changes or anomalies.<ref name="fsroundtable.org">{{Cite news|url=http://fsroundtable.org/cto-corner-artificial-intelligence-use-in-financial-services/|title=CTO Corner: Artificial Intelligence Use in Financial Services – Financial Services Roundtable|date=2 April 2015|work=Financial Services Roundtable|language=en-US|access-date=18 November 2016|url-status=dead|archiveurl=https://web.archive.org/web/20161118165842/http://fsroundtable.org/cto-corner-artificial-intelligence-use-in-financial-services/|archivedate=18 November 2016|df=dmy-all}}</ref><ref>{{Cite web|url=https://www.sas.com/en_ae/solutions/ai.html|title=Artificial Intelligence Solutions, AI Solutions|website=www.sas.com}}</ref><ref>{{Cite web|url=https://www.latimes.com/business/la-fi-palantir-sales-ipo-20190107-story.html|title=Palantir once mocked the idea of salespeople. Now it's hiring them|last=Chapman|first=Lizette|website=latimes.com|access-date=2019-02-28|date=7 January 2019}}</ref>

−

Banks use artificial intelligence systems today to organize operations, maintain book-keeping, invest in stocks, and manage properties. AI can react to changes overnight or when business is not taking place. In August 2001, robots beat humans in a simulated financial trading competition. AI has also reduced fraud and financial crimes by monitoring behavioral patterns of users for any abnormal changes or anomalies.

如今，银行使用AI系统来组织业务、记账、投资股票和管理房地产。AI可以对突然的变化和没有业务的情况做出反应<ref name="Eleanor">{{cite web|url=https://www.icas.com/ca-today-news/how-accountancy-and-finance-are-using-artificial-intelligence|title=Accounting, automation and AI|first=Eleanor|last=O'Neill|website=icas.com|language=English|date=31 July 2016|access-date=18 November 2016|url-status=live|archiveurl=https://web.archive.org/web/20161118165901/https://www.icas.com/ca-today-news/how-accountancy-and-finance-are-using-artificial-intelligence|archivedate=18 November 2016|df=dmy-all}}</ref>。2001年8月，机器人在一场模拟金融交易竞赛中击败了人类<ref>[http://news.bbc.co.uk/2/hi/business/1481339.stm Robots Beat Humans in Trading Battle.] {{webarchive|url=https://web.archive.org/web/20090909001249/http://news.bbc.co.uk/2/hi/business/1481339.stm |date=9 September 2009 }} BBC.com (8 August 2001)</ref>。AI还通过监测用户的行为模式发现异常变化或异常现象，减少了欺诈和金融犯罪<ref name="fsroundtable.org">{{Cite news|url=http://fsroundtable.org/cto-corner-artificial-intelligence-use-in-financial-services/|title=CTO Corner: Artificial Intelligence Use in Financial Services – Financial Services Roundtable|date=2 April 2015|work=Financial Services Roundtable|language=en-US|access-date=18 November 2016|url-status=dead|archiveurl=https://web.archive.org/web/20161118165842/http://fsroundtable.org/cto-corner-artificial-intelligence-use-in-financial-services/|archivedate=18 November 2016|df=dmy-all}}</ref><ref>{{Cite web|url=https://www.sas.com/en_ae/solutions/ai.html|title=Artificial Intelligence Solutions, AI Solutions|website=www.sas.com}}</ref><ref>{{Cite web|url=https://www.latimes.com/business/la-fi-palantir-sales-ipo-20190107-story.html|title=Palantir once mocked the idea of salespeople. Now it's hiring them|last=Chapman|first=Lizette|website=latimes.com|access-date=2019-02-28|date=7 January 2019}}</ref>。

−

AI is increasingly being used by [[Corporate finance|corporations]]. [[Jack Ma]] has controversially predicted that AI [[CEO]]'s are 30 years away.<ref>{{Cite web|url=https://money.cnn.com/2017/04/24/technology/alibaba-jack-ma-30-years-pain-robot-ceo/index.html|title=Jack Ma: In 30 years, the best CEO could be a robot|first=Sherisse|last=Pham|date=24 April 2017|website=CNNMoney}}</ref><ref>{{Cite web|url=https://venturebeat.com/2016/10/22/cant-find-a-perfect-ceo-create-an-ai-one-yourself/|title=Can't find a perfect CEO? Create an AI one yourself|date=22 October 2016}}</ref>

−

~~AI is increasingly being used by corporations. Jack Ma has controversially predicted that AI CEO's are 30 years away.~~

人AI正越来越多地被企业所使用。马云发表过一个有争议的预测：距离AI当上CEO还有30年的时间<ref>{{Cite web|url=https://money.cnn.com/2017/04/24/technology/alibaba-jack-ma-30-years-pain-robot-ceo/index.html|title=Jack Ma: In 30 years, the best CEO could be a robot|first=Sherisse|last=Pham|date=24 April 2017|website=CNNMoney}}</ref><ref>{{Cite web|url=https://venturebeat.com/2016/10/22/cant-find-a-perfect-ceo-create-an-ai-one-yourself/|title=Can't find a perfect CEO? Create an AI one yourself|date=22 October 2016}}</ref>。

−

The use of AI machines in the market in applications such as online trading and decision making has changed major economic theories.<ref>{{cite book |last1=Marwala |first1= Tshilidzi| last2=Hurwitz |first2= Evan |title=Artificial Intelligence and Economic Theory: Skynet in the Market |year=2017 |publisher=[[Springer Science+Business Media|Springer]] |location=London |isbn=978-3-319-66104-9}}</ref> For example, AI-based buying and selling platforms have changed the law of [[supply and demand]] in that it is now possible to easily estimate individualized demand and supply curves and thus individualized pricing. Furthermore, AI machines reduce [[information asymmetry]] in the market and thus making markets more efficient while reducing the volume of trades{{citation needed|date=July 2019}}. Furthermore, AI in the markets limits the consequences of behavior in the markets again making markets more efficient{{citation needed|date=July 2019}}. Other theories where AI has had impact include in [[rational choice]], [[rational expectations]], [[game theory]], [[Lewis turning point]], [[portfolio optimization]] and [[counterfactual thinking]]{{citation needed|date=July 2019}}.. In August 2019, the [[American Institute of Certified Public Accountants|AICPA]] introduced AI training course for accounting professionals.<ref>{{Cite web|url=https://www.mileseducation.com/finance/artificial_intelligence|title=Miles Education {{!}} Future Of Finance {{!}} Blockchain Fundamentals for F&A Professionals Certificate|website=www.mileseducation.com|access-date=2019-09-26|archive-url=https://web.archive.org/web/20190926102133/https://www.mileseducation.com/finance/artificial_intelligence|archive-date=26 September 2019|url-status=dead}}</ref>

−

The use of AI machines in the market in applications such as online trading and decision making has changed major economic theories. For example, AI-based buying and selling platforms have changed the law of supply and demand in that it is now possible to easily estimate individualized demand and supply curves and thus individualized pricing. Furthermore, AI machines reduce information asymmetry in the market and thus making markets more efficient while reducing the volume of trades. Furthermore, AI in the markets limits the consequences of behavior in the markets again making markets more efficient. Other theories where AI has had impact include in rational choice, rational expectations, game theory, Lewis turning point, portfolio optimization and counterfactual thinking.. In August 2019, the AICPA introduced AI training course for accounting professionals.

AI机器在市场上如在线交易和决策的应用改变了主流经济理论<ref>{{cite book |last1=Marwala |first1= Tshilidzi| last2=Hurwitz |first2= Evan |title=Artificial Intelligence and Economic Theory: Skynet in the Market |year=2017 |publisher=[[Springer Science+Business Media|Springer]] |location=London |isbn=978-3-319-66104-9}}</ref>。例如，基于AI的买卖平台改变了供求规律，因为现在可以通过AI很容易地估计个性化需求和供给曲线，从而实现个性化的定价。此外，AI减少了交易的信息不对称，在使市场更有效率的同时也减少了交易量。此外，AI限定了市场行为的后果，进一步提高了交易效率。受AI影响的其他理论包括理性选择、理性预期、博弈论、刘易斯转折点、投资组合优化和反事实思维。2019年8月，AICPA 为会计专业人员开设了 AI 培训课程<ref>{{Cite web|url=https://www.mileseducation.com/finance/artificial_intelligence|title=Miles Education {{!}} Future Of Finance {{!}} Blockchain Fundamentals for F&A Professionals Certificate|website=www.mileseducation.com|access-date=2019-09-26|archive-url=https://web.archive.org/web/20190926102133/https://www.mileseducation.com/finance/artificial_intelligence|archive-date=26 September 2019|url-status=dead}}</ref>。

第744行：第576行：

===网络安全 ===

−

~~{{More citations needed section|date=January 2020}}~~

−

The [[cybersecurity]] arena faces significant challenges in the form of large-scale hacking attacks of different types that harm organizations of all kinds and create billions of dollars in business damage. Artificial intelligence and Natural Language Processing (NLP) has begun to be used by security companies - for example, SIEM (Security Information and Event Management) solutions. The more advanced of these solutions use AI and NLP to automatically sort the data in networks into high risk and low-risk information. This enables security teams to focus on the attacks that have the potential to do real harm to the organization, and not become victims of attacks such as [[Denial-of-service attack|Denial of Service (DoS)]], [[Malware]] and others.

−

The cybersecurity arena faces significant challenges in the form of large-scale hacking attacks of different types that harm organizations of all kinds and create billions of dollars in business damage. Artificial intelligence and Natural Language Processing (NLP) has begun to be used by security companies - for example, SIEM (Security Information and Event Management) solutions. The more advanced of these solutions use AI and NLP to automatically sort the data in networks into high risk and low-risk information. This enables security teams to focus on the attacks that have the potential to do real harm to the organization, and not become victims of attacks such as Denial of Service (DoS), Malware and others.

网络安全领域面临着各种大规模黑客攻击的重大挑战，这些攻击损害到了很多组织，造成了数十亿美元的商业损失。网络安全公司已经开始使用AI和自然语言处理(NLP) ，例如，SIEM (Security Information and Event Management,安全信息和事件管理)解决方案。更高级的解决方案使用AI和自然语言处理将网络中的数据划分为高风险和低风险两类信息。这使得安全团队能够专注于对付那些有可能对组织造成真正伤害的攻击，不沦为分布式拒绝服务攻击（DoS）、恶意软件和其他攻击的受害者。

第756行：第581行：

===政务 ===

−

~~{{Main|Artificial intelligence in government}}~~

−

Artificial intelligence in government consists of applications and regulation. Artificial intelligence paired with [[facial recognition system]]s may be used for [[mass surveillance]]. This is already the case in some parts of China.<ref>{{Cite news|url=https://www.nytimes.com/2019/05/22/world/asia/china-surveillance-xinjiang.html|title=How China Uses High-Tech Surveillance to Subdue Minorities|first1=Chris|last1=Buckley|first2=Paul|last2=Mozur|date=22 May 2019|work=The New York Times}}</ref><ref>{{Cite web|url=http://social.techcrunch.com/2019/05/03/china-smart-city-exposed/|title=Security lapse exposed a Chinese smart city surveillance system}}</ref> An artificial intelligence has also competed in the Tama City [[AI mayor|mayoral elections]] in 2018.

−

Artificial intelligence in government consists of applications and regulation. Artificial intelligence paired with facial recognition systems may be used for mass surveillance. This is already the case in some parts of China. An artificial intelligence has also competed in the Tama City mayoral elections in 2018.

政务AI包括应用和管理。AI与人脸识别系统相结合可用于大规模监控。中国的一些地区已经开始使用这种技术<ref>{{Cite news|url=https://www.nytimes.com/2019/05/22/world/asia/china-surveillance-xinjiang.html|title=How China Uses High-Tech Surveillance to Subdue Minorities|first1=Chris|last1=Buckley|first2=Paul|last2=Mozur|date=22 May 2019|work=The New York Times}}</ref><ref>{{Cite web|url=http://social.techcrunch.com/2019/05/03/china-smart-city-exposed/|title=Security lapse exposed a Chinese smart city surveillance system}}</ref>。一个AI还参与了2018年Tama City市长选举的角逐。

−

In 2019, the tech city of Bengaluru in India is set to deploy AI managed traffic signal systems across the 387 traffic signals in the city. This system will involve use of cameras to ascertain traffic density and accordingly calculate the time needed to clear the traffic volume which will determine the signal duration for vehicular traffic across streets.<ref>{{Cite web|url=https://nextbigwhat.com/ai-traffic-signals-to-be-installed-in-bengaluru-soon/|title=AI traffic signals to be installed in Bengaluru soon|date=2019-09-24|website=NextBigWhat|language=en-US|access-date=2019-10-01}}</ref>

−

In 2019, the tech city of Bengaluru in India is set to deploy AI managed traffic signal systems across the 387 traffic signals in the city. This system will involve use of cameras to ascertain traffic density and accordingly calculate the time needed to clear the traffic volume which will determine the signal duration for vehicular traffic across streets.

2019年，印度硅谷班加罗尔将在该市的387个交通信号灯上部署AI控制的交通信号系统。这个系统将使用摄像头来确定交通密度，并据此计算清除交通量所需的时间，决定街道上的车辆交通灯的持续时间<ref>{{Cite web|url=https://nextbigwhat.com/ai-traffic-signals-to-be-installed-in-bengaluru-soon/|title=AI traffic signals to be installed in Bengaluru soon|date=2019-09-24|website=NextBigWhat|language=en-US|access-date=2019-10-01}}</ref>。

第776行：第589行：

===与法律有关的专业 ===

−

~~{{Main|Legal informatics#Artificial intelligence}}~~

−

Artificial intelligence (AI) is becoming a mainstay component of law-related professions. In some circumstances, this analytics-crunching technology is using algorithms and machine learning to do work that was previously done by entry-level lawyers.{{Citation needed|date=December 2019}}

−

Artificial intelligence (AI) is becoming a mainstay component of law-related professions. In some circumstances, this analytics-crunching technology is using algorithms and machine learning to do work that was previously done by entry-level lawyers.

AI正在成为法律相关专业的主要组成部分。一些情况下，人们会通过AI分析处理技术，使用算法和机器学习来完成以前由初级律师完成的工作。

−

In [[Electronic discovery|Electronic Discovery (eDiscovery)]], the industry has been focused on machine learning (predictive coding/technology assisted review), which is a subset of AI. To add to the soup of applications, Natural Language Processing (NLP) and Automated Speech Recognition (ASR) are also in vogue in the industry.<ref>{{Cite web|url=https://www.ft.com/content/fef40df0-4a6a-11e9-bde6-79eaea5acb64|title=AI learns to read Korean, so you don't have to|last=Croft|first=Jane|date=2019-05-02|website=Financial Times|language=en-GB|access-date=2019-12-19}}</ref>

−

In Electronic Discovery (eDiscovery), the industry has been focused on machine learning (predictive coding/technology assisted review), which is a subset of AI. To add to the soup of applications, Natural Language Processing (NLP) and Automated Speech Recognition (ASR) are also in vogue in the industry.

电子资料档案查询(eDiscovery)产业一直很关注机器学习(预测编码 / 技术辅助评审) ，这是AI的一个子领域。自然语言处理(NLP)和自动语音识别(ASR)也正在这个行业流行起来。

第795行：第597行：

===电子游戏 ===

−

~~{{Main|Artificial intelligence (video games)}}~~

−

In video games, artificial intelligence is routinely used to generate dynamic purposeful behavior in [[non-player character]]s (NPCs). In addition, well-understood AI techniques are routinely used for [[pathfinding]]. Some researchers consider NPC AI in games to be a "solved problem" for most production tasks. Games with more atypical AI include the AI director of ''[[Left 4 Dead]]'' (2008) and the neuroevolutionary training of platoons in ''[[Supreme Commander 2]]'' (2010).<ref>{{cite news|url=https://www.economist.com/news/science-and-technology/21721890-games-help-them-understand-reality-why-ai-researchers-video-games|title=Why AI researchers like video games|website=The Economist|url-status=live|archiveurl=https://web.archive.org/web/20171005051028/https://www.economist.com/news/science-and-technology/21721890-games-help-them-understand-reality-why-ai-researchers-video-games|archivedate=5 October 2017|df=dmy-all}}</ref><ref>Yannakakis, G. N. (2012, May). Game AI revisited. In Proceedings of the 9th conference on Computing Frontiers (pp. 285–292). ACM.</ref>

−

In video games, artificial intelligence is routinely used to generate dynamic purposeful behavior in non-player characters (NPCs). In addition, well-understood AI techniques are routinely used for pathfinding. Some researchers consider NPC AI in games to be a "solved problem" for most production tasks. Games with more atypical AI include the AI director of Left 4 Dead (2008) and the neuroevolutionary training of platoons in Supreme Commander 2 (2010).

−

在视频游戏中，AI通常被用来让非玩家角色( non-player characters，NPCs)中做出动态的目的性行为。此外，还常用简单的AI技术寻路。一些研究人员认为，对于大多数生产任务来说，游戏中的 NPC AI 是一个“已解决问题”。含更多非典型 AI 的游戏有《求生之路》（Left 4 Dead, 2008)中的 AI 导演和《最高指挥官2》(Supreme Commander 2, 2010)中的对一个野战排进行的神经演化训练<ref>{{cite news|url=https://www.economist.com/news/science-and-technology/21721890-games-help-them-understand-reality-why-ai-researchers-video-games|title=Why AI researchers like video games|website=The Economist|url-status=live|archiveurl=https://web.archive.org/web/20171005051028/https://www.economist.com/news/science-and-technology/21721890-games-help-them-understand-reality-why-ai-researchers-video-games|archivedate=5 October 2017|df=dmy-all}}</ref><ref>Yannakakis, G. N. (2012, May). Game AI revisited. In Proceedings of the 9th conference on Computing Frontiers (pp. 285–292). ACM.</ref>。

第811行：第605行：

===军事 ===

−

~~{{Further|Artificial intelligence arms race|Lethal autonomous weapon|Unmanned combat aerial vehicle}}~~

−

The United States and other nations are developing AI applications for a range of military functions.<ref name=":2">{{Cite book|last=Congressional Research Service|first=|url=https://fas.org/sgp/crs/natsec/R45178.pdf|title=Artificial Intelligence and National Security|publisher=Congressional Research Service|year=2019|isbn=|location=Washington, DC|pages=}}[[Template:PD-notice|PD-notice]]</ref> The main military applications of Artificial Intelligence and Machine Learning are to enhance C2, Communications, Sensors, Integration and Interoperability.<ref name="AI">{{cite web|title=Artificial intelligence as the basis of future control networks.|url=https://www.researchgate.net/publication/334573170|last=Slyusar|first=Vadym|date=2019|work=Preprint}}</ref> AI research is underway in the fields of intelligence collection and analysis, logistics, cyber operations, information operations, command and control, and in a variety of semiautonomous and autonomous vehicles.<ref name=":2" /> Artificial Intelligence technologies enable coordination of sensors and effectors, threat detection and identification, marking of enemy positions, target acquisition, coordination and deconfliction of distributed Join Fires between networked combat vehicles and tanks also inside Manned and Unmanned Teams (MUM-T).<ref name=AI /> AI has been incorporated into military operations in Iraq and Syria.<ref name=":2" />

−

The United States and other nations are developing AI applications for a range of military functions. The main military applications of Artificial Intelligence and Machine Learning are to enhance C2, Communications, Sensors, Integration and Interoperability. AI research is underway in the fields of intelligence collection and analysis, logistics, cyber operations, information operations, command and control, and in a variety of semiautonomous and autonomous vehicles. Artificial Intelligence technologies enable coordination of sensors and effectors, threat detection and identification, marking of enemy positions, target acquisition, coordination and deconfliction of distributed Join Fires between networked combat vehicles and tanks also inside Manned and Unmanned Teams (MUM-T). AI has been incorporated into military operations in Iraq and Syria.

美国和其他国家正在为一系列军事目的开发AI应用程序<ref name=":2">{{Cite book|last=Congressional Research Service|first=|url=https://fas.org/sgp/crs/natsec/R45178.pdf|title=Artificial Intelligence and National Security|publisher=Congressional Research Service|year=2019|isbn=|location=Washington, DC|pages=}}[[Template:PD-notice|PD-notice]]</ref>。AI和机器学习的主要军事应用是增强 C2、通信、传感器、集成和互操作性。情报收集和分析、后勤、网络操作、信息操作、指挥和控制以及各种半自动和自动车辆等领域正在进行AI研究<ref name="AI">{{cite web|title=Artificial intelligence as the basis of future control networks.|url=https://www.researchgate.net/publication/334573170|last=Slyusar|first=Vadym|date=2019|work=Preprint}}</ref> AI research is underway in the fields of intelligence collection and analysis, logistics, cyber operations, information operations, command and control, and in a variety of semiautonomous and autonomous vehicles.<ref name=":2" />。AI技术能够协调传感器和效应器、探测威胁和识别、标记敌人阵地、目标获取、协调和消除有人和无人小组(MUM-T)、联网作战车辆和坦克内部的分布式联合火力<ref name=AI /> 。伊拉克和叙利亚的军事行动就采用了AI。<ref name=":2" />

−

Worldwide annual military spending on robotics rose from US$5.1 billion in 2010 to US$7.5 billion in 2015.<ref>{{cite news|title=Getting to grips with military robotics|url=https://www.economist.com/news/special-report/21735478-autonomous-robots-and-swarms-will-change-nature-warfare-getting-grips|accessdate=7 February 2018|work=The Economist|date=25 January 2018|language=en}}</ref><ref>{{cite web|title=Autonomous Systems: Infographic|url=https://www.siemens.com/innovation/en/home/pictures-of-the-future/digitalization-and-software/autonomous-systems-infographic.html|website=siemens.com|accessdate=7 February 2018|language=en}}</ref> Military drones capable of autonomous action are widely considered a useful asset.<ref>{{Cite web|url=https://www.cnas.org/publications/reports/understanding-chinas-ai-strategy|title=Understanding China's AI Strategy|last=Allen|first=Gregory|date=February 6, 2019|website=www.cnas.org/publications/reports/understanding-chinas-ai-strategy|publisher=Center for a New American Security|archive-url=https://web.archive.org/web/20190317004017/https://www.cnas.org/publications/reports/understanding-chinas-ai-strategy|archive-date=March 17, 2019|url-status=|access-date=March 17, 2019}}</ref> Many artificial intelligence researchers seek to distance themselves from military applications of AI.<ref>{{cite news|last1=Metz|first1=Cade|title=Pentagon Wants Silicon Valley's Help on A.I.|url=https://www.nytimes.com/2018/03/15/technology/military-artificial-intelligence.html|accessdate=19 March 2018|work=The New York Times|date=15 March 2018}}</ref>

−

Worldwide annual military spending on robotics rose from US$5.1 billion in 2010 to US$7.5 billion in 2015. Military drones capable of autonomous action are widely considered a useful asset. Many artificial intelligence researchers seek to distance themselves from military applications of AI.

全球每年在机器人方面的军费开支从2010年的51亿美元增加到2015年的75亿美元<ref>{{cite news|title=Getting to grips with military robotics|url=https://www.economist.com/news/special-report/21735478-autonomous-robots-and-swarms-will-change-nature-warfare-getting-grips|accessdate=7 February 2018|work=The Economist|date=25 January 2018|language=en}}</ref><ref>{{cite web|title=Autonomous Systems: Infographic|url=https://www.siemens.com/innovation/en/home/pictures-of-the-future/digitalization-and-software/autonomous-systems-infographic.html|website=siemens.com|accessdate=7 February 2018|language=en}}</ref>。人们都认为具有自主行动能力的军用无人机很有价值<ref>{{Cite web|url=https://www.cnas.org/publications/reports/understanding-chinas-ai-strategy|title=Understanding China's AI Strategy|last=Allen|first=Gregory|date=February 6, 2019|website=www.cnas.org/publications/reports/understanding-chinas-ai-strategy|publisher=Center for a New American Security|archive-url=https://web.archive.org/web/20190317004017/https://www.cnas.org/publications/reports/understanding-chinas-ai-strategy|archive-date=March 17, 2019|url-status=|access-date=March 17, 2019}}</ref>。而许多AI研究人员则正在试图远离AI的军事应用。<ref>{{cite news|last1=Metz|first1=Cade|title=Pentagon Wants Silicon Valley's Help on A.I.|url=https://www.nytimes.com/2018/03/15/technology/military-artificial-intelligence.html|accessdate=19 March 2018|work=The New York Times|date=15 March 2018}}</ref>

第873行：第657行：

AI催生了许多在如视觉艺术等领域的创造性应用。在纽约现代艺术博物馆举办的“思考机器: 计算机时代的艺术与设计，1959-1989”展览概述了艺术、建筑和设计的历史中AI的应用<ref name="moma">{{Cite web|url=https://www.moma.org/calendar/exhibitions/3863|title=Thinking Machines: Art and Design in the Computer Age, 1959–1989|website=The Museum of Modern Art|language=en|access-date=2019-07-23}}</ref>。最近的展览展示了AI在艺术创作中的应用<ref name = wp1>[https://www.washingtonpost.com/news/innovations/wp/2016/03/10/googles-psychedelic-paint-brush-raises-the-oldest-question-in-art/ Retrieved July 29]</ref>，包括谷歌赞助的旧金山灰色地带基金会(Gray Area Foundation)的慈善拍卖会，艺术家们在拍卖会中尝试了 DeepDream 算法，以及2017年秋天在洛杉矶和法兰克福举办的“非人类: AI时代的艺术”展览.<ref name = sf>{{cite web|url=https://www.statefestival.org/program/2017/unhuman-art-in-the-age-of-ai |title=Unhuman: Art in the Age of AI – State Festival |publisher=Statefestival.org |date= |accessdate=2018-09-13}}</ref><ref name="artsy">{{Cite web|url=https://www.artsy.net/article/artsy-editorial-hard-painting-made-computer-human|title=It's Getting Hard to Tell If a Painting Was Made by a Computer or a Human|last=Chun|first=Rene|date=2017-09-21|website=Artsy|language=en|access-date=2019-07-23}}</ref>。2018年春天，计算机协会发行了一期主题为计算机和艺术的特刊，着重展示了机器学习在艺术中的作用。奥地利电子艺术博物馆和维也纳应用艺术博物馆于2019年开设了AI展览<ref name="Ars Electronica Exhibition ''Understanding AI''">{{Cite web|url=https://ars.electronica.art/center/en/exhibitions/ai/ |access-date=September 2019}}</ref><ref name="Museum of Applied Arts Exhibition ''Uncanny Values''">{{Cite web|url=https://www.mak.at/en/program/exhibitions/uncanny_values |access-date=October 2019|title=MAK Wien - MAK Museum Wien}}</ref>。2019年的电子艺术节 “Out of the box”将AI艺术在可持续社会转型中的作用变成了一个主题<ref name="European Platform for Digital Humanism">{{Cite web|url=https://ars.electronica.art/outofthebox/en/digital-humanism-conf/ |access-date=September 2019}}</ref>。

+

== 哲学和伦理学 ==

−

~~{{Main|Philosophy of artificial intelligence|Ethics of artificial intelligence}}~~

−

~~There are three philosophical questions related to AI:~~

−

~~There are three philosophical questions related to AI:~~

有三个与人工智能相关的哲学问题:

−

~~# Is [[artificial general intelligence]] possible? Can a machine solve any problem that a human being can solve using intelligence? Or are there hard limits to what a machine can accomplish?~~

−

~~# Are intelligent machines dangerous? How can we ensure that machines behave ethically and that they are used ethically?~~

−

# Can a machine have a [[mind]], [[consciousness]] and [[philosophy of mind|mental states]] in exactly the same sense that human beings do? Can a machine be [[Sentience|sentient]], and thus deserve certain rights? Can a machine [[intention]]ally cause harm?

−

# 通用人工智能可能实现吗？机器能解决任何人类使用智能就能解决的问题吗？或者一台机器所能完成的事情是否有严格的界限？

第900行：第669行：

===人工智能的局限性===

−

~~{{Main|Philosophy of AI|Turing test|Physical symbol systems hypothesis|Dreyfus' critique of AI|The Emperor's New Mind|AI effect}}~~

−

~~Can a machine be intelligent? Can it "think"?~~

−

~~Can a machine be intelligent? Can it "think"?~~

机器是智能的吗？它能“思考”吗？

第912行：第675行：

;''[[Computing Machinery and Intelligence|Alan Turing's "polite convention"]]'': We need not decide if a machine can "think"; we need only decide if a machine can act as intelligently as a human being. This approach to the philosophical problems associated with artificial intelligence forms the basis of the [[Turing test]].<ref name="Turing test"/>

−

Alan Turing's "polite convention": We need not decide if a machine can "think"; we need only decide if a machine can act as intelligently as a human being. This approach to the philosophical problems associated with artificial intelligence forms the basis of the Turing test.

;''阿兰 · 图灵的“礼貌惯例'': 阿兰 · 图灵的'''“礼貌惯例”''' : 我们不需要决定一台机器是否可以“思考”；我们只需要决定一台机器是否可以像人一样聪明地行动。这个对AI相关哲学问题的回应成为了图灵测试的基础。

第919行：第681行：

;''The [[Dartmouth Workshop|Dartmouth proposal]]'': "Every aspect of learning or any other feature of intelligence can be so precisely described that a machine can be made to simulate it." This conjecture was printed in the proposal for the Dartmouth Conference of 1956.<ref name="Dartmouth proposal"/>

−

The Dartmouth proposal: "Every aspect of learning or any other feature of intelligence can be so precisely described that a machine can be made to simulate it." This conjecture was printed in the proposal for the Dartmouth Conference of 1956.

;''达特茅斯提案'':达特茅斯会议提出: “可以通过准确地描述学习的每个方面或智能的任何特征，使得一台机器模拟学习和智能。”这个猜想被写在了1956年达特茅斯学院会议的提案中。

−

;''[[Physical symbol system|Newell and Simon's physical symbol system hypothesis]]'': "A physical symbol system has the necessary and sufficient means of general intelligent action." Newell and Simon argue that intelligence consists of formal operations on symbols.<ref name="Physical symbol system hypothesis"/> [[Hubert Dreyfus]] argued that, on the contrary, human expertise depends on unconscious instinct rather than conscious symbol manipulation and on having a "feel" for the situation rather than explicit symbolic knowledge. (See [[Dreyfus' critique of AI]].)<ref>

−

Newell and Simon's physical symbol system hypothesis: "A physical symbol system has the necessary and sufficient means of general intelligent action." Newell and Simon argue that intelligence consists of formal operations on symbols. Hubert Dreyfus argued that, on the contrary, human expertise depends on unconscious instinct rather than conscious symbol manipulation and on having a "feel" for the situation rather than explicit symbolic knowledge. (See Dreyfus' critique of AI.)

;纽厄尔和西蒙的物理符号系统假说: 物理符号系统是通往通用智能行为的充分必要途径。纽厄尔和西蒙认为智能由符号形式的运算组成。<ref name="Physical symbol system hypothesis"/> 休伯特·德雷福斯则相反地认为，人类的知识依赖于无意识的本能，而不是有意识的符号运算；依赖于对情境的“感觉”，而不是明确的符号知识。(参见德雷福斯对人工智能的批评。)<ref>Dreyfus criticized the [[necessary and sufficient|necessary]] condition of the [[physical symbol system]] hypothesis, which he called the "psychological assumption": "The mind can be viewed as a device operating on bits of information according to formal rules." {{Harv|Dreyfus|1992|p=156}}</ref><ref name="Dreyfus' critique"/>

−

;''Gödelian arguments'': [[Gödel]] himself,<ref name="Gödel himself"/> [[John Lucas (philosopher)|John Lucas]] (in 1961) and [[Roger Penrose]] (in a more detailed argument from 1989 onwards) made highly technical arguments that human mathematicians can consistently see the truth of their own "Gödel statements" and therefore have computational abilities beyond that of mechanical Turing machines.<ref name="The mathematical objection"/> However, some people do not agree with the "Gödelian arguments".<ref>{{cite web|author1=Graham Oppy|title=Gödel's Incompleteness Theorems|url=http://plato.stanford.edu/entries/goedel-incompleteness/#GdeArgAgaMec|website=[[Stanford Encyclopedia of Philosophy]]|accessdate=27 April 2016|date=20 January 2015|quote=These Gödelian anti-mechanist arguments are, however, problematic, and there is wide consensus that they fail.|author1-link=Graham Oppy}}</ref><ref>{{cite book|author1=Stuart J. Russell|author2-link=Peter Norvig|author2=Peter Norvig|title=Artificial Intelligence: A Modern Approach|date=2010|publisher=[[Prentice Hall]]|location=Upper Saddle River, NJ|isbn=978-0-13-604259-4|edition=3rd|chapter=26.1.2: Philosophical Foundations/Weak AI: Can Machines Act Intelligently?/The mathematical objection|quote=even if we grant that computers have limitations on what they can prove, there is no evidence that humans are immune from those limitations.|title-link=Artificial Intelligence: A Modern Approach|author1-link=Stuart J. Russell}}</ref><ref>Mark Colyvan. An introduction to the philosophy of mathematics. [[Cambridge University Press]], 2012. From 2.2.2, 'Philosophical significance of Gödel's incompleteness results': "The accepted wisdom (with which I concur) is that the Lucas-Penrose arguments fail."</ref>

第940行：第694行：

;''The [[artificial brain]] argument'': The brain can be simulated by machines and because brains are intelligent, simulated brains must also be intelligent; thus machines can be intelligent. [[Hans Moravec]], [[Ray Kurzweil]] and others have argued that it is technologically feasible to copy the brain directly into hardware and software and that such a simulation will be essentially identical to the original.<ref name="Brain simulation"/>

−

The artificial brain argument: The brain can be simulated by machines and because brains are intelligent, simulated brains must also be intelligent; thus machines can be intelligent. Hans Moravec, Ray Kurzweil and others have argued that it is technologically feasible to copy the brain directly into hardware and software and that such a simulation will be essentially identical to the original.

;''人工大脑的观点'': 因为大脑可以被机器模拟，且大脑是智能的，模拟的大脑也必须是智能的；因此机器可以是智能的。汉斯·莫拉维克、雷·库兹韦尔和其他人认为，技术层面直接将大脑复制到硬件和软件上是可行的，而且这些拷贝在本质上和原来的大脑是没有区别的。

第948行：第700行：

;''The [[AI effect]]'': Machines are ''already'' intelligent, but observers have failed to recognize it. When [[Deep Blue (chess computer)|Deep Blue]] beat [[Garry Kasparov]] in chess, the machine was acting intelligently. However, onlookers commonly discount the behavior of an artificial intelligence program by arguing that it is not "real" intelligence after all; thus "real" intelligence is whatever intelligent behavior people can do that machines still cannot. This is known as the AI Effect: "AI is whatever hasn't been done yet."

−

The AI effect: Machines are already intelligent, but observers have failed to recognize it. When Deep Blue beat Garry Kasparov in chess, the machine was acting intelligently. However, onlookers commonly discount the behavior of an artificial intelligence program by arguing that it is not "real" intelligence after all; thus "real" intelligence is whatever intelligent behavior people can do that machines still cannot. This is known as the AI Effect: "AI is whatever hasn't been done yet."

;''AI效应'': 机器本来就是智能的，但是观察者却没有意识到这一点。当深蓝在国际象棋比赛中击败加里 · 卡斯帕罗夫时，机器就在做出智能行为。然而，旁观者通常对AI程序的行为不屑一顾，认为它根本不是“真正的”智能; 因此，“真正的”智能就是人任何类能够做到但机器仍然做不到的智能行为。这就是众所周知的AI效应: “AI就是一切尚未完成的事情"。

第955行：第706行：

===潜在危害===

−

Widespread use of artificial intelligence could have [[unintended consequences]] that are dangerous or undesirable. Scientists from the [[Future of Life Institute]], among others, described some short-term research goals to see how AI influences the economy, the laws and ethics that are involved with AI and how to minimize AI security risks. In the long-term, the scientists have proposed to continue optimizing function while minimizing possible security risks that come along with new technologies.<ref>Russel, Stuart., Daniel Dewey, and Max Tegmark. Research Priorities for Robust and Beneficial Artificial Intelligence. AI Magazine 36:4 (2015). 8 December 2016.</ref>

−

Widespread use of artificial intelligence could have unintended consequences that are dangerous or undesirable. Scientists from the Future of Life Institute, among others, described some short-term research goals to see how AI influences the economy, the laws and ethics that are involved with AI and how to minimize AI security risks. In the long-term, the scientists have proposed to continue optimizing function while minimizing possible security risks that come along with new technologies.

AI的广泛使用可能会产生危险或导致意外后果。生命未来研究所(Future of Life Institute)等机构的科学家提出了一些短期研究目标，以此了解AI如何影响经济、与AI相关的法律和道德规范，以及如何将AI的安全风险降到最低。从长远来看，科学家们建议继续优化功能，同时最小化新技术带来的可能的安全风险。<ref>Russel, Stuart., Daniel Dewey, and Max Tegmark. Research Priorities for Robust and Beneficial Artificial Intelligence. AI Magazine 36:4 (2015). 8 December 2016.</ref>

薄荷

7,129

个编辑