几年前,让电脑用简单的英语回答问题,这种想法似乎是科幻小说。But when we releasedWolfram|Alpha2009年的一个大惊喜(尤其是对我而言!)是因为我们已经成功地让这一切发挥了作用。And by now people routinely ask personal assistant systems—many powered by Wolfram|Alpha—zillions of questions in ordinary language every day.

用普通语言提问,从Wolfram Alpha获得答案

所有这些对于快速提问都很有效,or short commands (though we're always trying to使它更好!).但是更复杂的东西呢?What's the best way to communicate more seriously withAIs

我已经想了很久了,试图把哲学的线索结合起来,linguistics,神经科学,computer science and other areas.让我有点惊讶的是,what I've realized recently is that a big part of the answer may actually be sitting right in front of me,以过去30年来我一直在努力的形式:沃尔夫拉姆语.

Maybe this is a case of having a hammer and then seeing everything as a nail.But I'm pretty sure there's more to it.  And at the very least,思考这个问题是了解人工智能及其与人类关系的一种方式。


The first key point—that I came to understand clearly only after a一系列的发现我在基础科学中说过计算是一个非常强大的东西,that lets even tiny programs (like cellular automata,或神经网络)难以置信的复杂方法。And it's this kind of thing that an AI can harness.



Language of 金宝博188投注Computational Thinking

I have seen my role as being to identify lumps of computation that people will understand and want to use,like查找短行程,图像识别Predict.传统的计算机语言集中在接近计算机实际硬件的底层结构上。But in the Wolfram Language I've instead started from what we humans understand,然后尽可能多地用语言来表达。

早年,我们主要处理相当抽象的概念,about,说,数学、逻辑或抽象网络。但近年来与Wolfram Alpha密切相关的一项重大成就是,我们能够扩展我们建造的结构,以覆盖无数真实的种类世界上的事物,如城市、电影或动物。

One might wonder: why invent a language for all this;为什么不只用,说,English?好,for specific things,like "热粉红““new york city“或“冥王星的卫星“英语很好,事实上,Wolfram语言让人们就用英语吧.But when one's trying to describe more complex things,plain English pretty quickly gets unwieldy.

想象一下,例如,试图描述一个非常简单的算法程序。A back-and-forth dialog—"Turing-test style"—would rapidly get frustrating.而且,一段笔直的英语几乎肯定会以难以置信的繁复散文结尾,就像人们在复杂的法律文件中发现的那样。

The Wolfram Language specifies clearly and succinctly how to create this image.The equivalent natural-language specification is complicated and subject to misinterpretation.

但是Wolfram语言正是为了解决这些问题而建立的。它被设定为容易被人类理解,捕捉人类描述和思考事物的方式。然而,它也有一个允许任意复杂性被组装和通信的结构。And,of course,it's readily understandable not just by humans,but also by machines.


The Wolfram Language mixes well with English in documents and thought streams


但让我们回到人工智能。在计算机的大部分历史中,我们通过让人类程序员显式地编写代码行来构建程序,understanding (apart from bugs!) what each line does.但是要达到可以合理地称之为人工智能的目标,就需要利用更多的计算能力。And to do this one has to go beyond programs that humans can directly write—and somehow automatically sample a broader swath of possible programs.

我们可以通过算法自动化我们早就习惯了Mathematicaand the Wolfram Language,或者我们可以通过明确machine learning,或通过searching the 金宝博188投注computational universe可能的程序。但不管我们怎么做,one feature of the programs that come out is that they have no reason to be understandable by humans.

Engineered programs are written to be human-readable.Automatically created or discovered programs are not necessarily human-readable.

在某种程度上,这是令人不安的。我们不知道程序是如何工作的,or what they might be capable of.但我们知道他们在做精细的计算,从某种意义上来说不可约的复杂to analyze.

还有一个,非常熟悉的地方,同样的事情发生在那里:自然世界。Whether we look at流体动力学,或biology,或者什么,we see all sorts of complexity.And in fact the计算等效原理金宝博188投注that emerged from the basic science I did implies that this complexity is in a sense exactly the same as the complexity that can occur in 金宝博188投注computational systems.


但对AI来说,我们必须冒险进入更广阔的计算领域,金宝博188投注where—as in the natural world—we're inevitably dealing with things we cannot readily understand.


Let's imagine we have a perfect,完全人工智能,这可以做任何我们可以合理联想到情报的事情。也许它会从很多物联网传感器。And it has all sorts of computation going on inside.但它最终会做什么呢?What is its purpose going to be?

这将深入探讨一些相当深刻的哲学,involving issues that have been batted around for thousands of years—but which finally are going to really matter in dealing with AIs.

One might think that as an AI becomes more sophisticated,它的目的也是如此,最终,人工智能将以某种终极抽象的目的结束。But this doesn't make sense.Because there is really没有抽象定义的绝对目的,可以用纯形式的数学或计算方法推导。金宝博188投注Purpose is something that's defined only with respect to humans,以及他们独特的历史和文化。

An "abstract AI",与人类目的无关,will just go along doing computation.And as with most cellular automata and most systems in nature,我们将无法识别或确定任何特定的“目的”to that computation,或者是系统。


Technology has always been about automating things so humans can define goals,然后这些目标就可以通过这项技术自动实现了。


你对人工智能说什么来告诉它你想要它为你做什么?You're not going to be able to tell it exactly what to do in each and every circumstance.You'd only be able to do that if the computations the AI could do were tightly constrained,就像传统的软件工程一样。But for the AI to work properly,it's going to have to make use of broader parts of the 金宝博188投注computational universe.这是我称之为现象的结果金宝博188投注违背了计算不可化归性你永远无法决定它会做什么。


但是如果你想定义更复杂的目标,或者目标与人工智能已经经历的不密切相关?那么少量的自然语言就不够了。Perhaps the AI could go through a whole education.But a better idea would be to leverage what we have in the Wolfram Language,它实际上已经有很多关于它的世界的知识,以一种人类和人工智能都可以使用的方式。


思考人类如何与AIs交流是一回事。But how will AIs communicate with one another?One might imagine they could do literal transfers of their underlying representations of knowledge.但那不管用,因为一旦两个人工智能有了不同的“经验”,the representations they use will inevitably be at least somewhat different.

所以,just like humans,人工智能将最终需要使用某种形式的符号语言来抽象地表示概念,没有具体参考这些概念的基本表示。

One might then think the AIs should just communicate in English;至少那样我们才能理解他们!但不会成功的。因为自动识别系统不可避免地需要逐步扩展他们的语言,所以即使它以英语开始,it wouldn't stay that way.

在人类自然语言中,new words get added when there are new concepts that are widespread enough to make representing them in the language useful.有时一个新的概念与世界上的新事物相关联(“blog”,“表情”"smartphone","clickbait",等);sometimes it's associated with a new distinction among existing things ("road"VS“高速公路”“模式”VS“分形”。

Often it's science that gives us new distinctions between things,by identifying distinct clusters of behavior or structure.但关键是,人工智能可以在比人类更大的范围内做到这一点。For example,我们的图像识别项目是用来识别我们人类日常生活中所用的大约10000种物体的。但在内部,因为它是用来自世界的图像训练的,it's discovering all sorts of other distinctions that we don't have names for,but that are successful at robustly separating things.

我称之为“后语言紧急概念”。(或)PLECs)And I think it's inevitable that in a population of AIs,an ever-expanding hierarchy of PLECs will appear,forcing the language of the AIs to progressively expand.

但是,英语的框架如何支持这一点呢?我想每个新概念都可以被分配一个由一些散列代码(如字母集合)组成的单词。But a structured symbolic language—as the Wolfram Language is—provides a much better framework.Because it doesn't require the units of the language to be simple "words",but allows them to be arbitrary lumps of symbolic information,such as collections of examples (so that,例如,一个词可以用一个带有定义的符号结构来表示。

那么,我应该用沃尔夫拉姆语互相交谈吗?It seems to make a lot of sense—because it effectively starts from the understanding of the world that's been developed through human knowledge,but then provides a framework for going further.不管语法是如何编码的(输入形式,XML杰森binary,无论如何)。What matters is the structure and content that are built into the language.


在地球上生命存在的数十亿年中,有几种不同的信息传递方式。最基本的是基因组学:在硬件层面传递信息。But then there are neural systems,就像大脑一样。And these get information—like our Image Identification Project—by accumulating it from experiencing the world.这是生物体用来观察的机制,and to do many other "AI-ish"things.

But in a sense this mechanism is fundamentally limited,因为每一个不同的有机体和每一个不同的大脑都必须自己经历整个学习过程:在一代人中获得的任何信息都不能轻易传递给下一代人。

But this is where our species made its great invention: natural language.Because with natural language it's possible to take information that's been learned,and communicate it in abstract form,从一代传到下一代。但是还是有问题,because when natural language is received,it still has to be interpreted,在每个大脑中以不同的方式。

Information transfer:  Level 0: genomics;Level 1: individual brains;第二层次:自然语言;第三层次:计算知识金宝博188投注语言

这就是像Wolfram语言这样的计算知识语言的概念很重要的地方:金宝博188投注因为它提供了一种交流世界概念和事实的方式,in a way that can immediately and reproducibly be executed,without requiring separate interpretation on the part of whatever receives it.

It's probably not a stretch to say that the invention of human natural language was what led to civilization and our modern world.那么,进入另一个层次的含义是什么:拥有精确的计算知识语言,金宝博188投注that carries not just abstract concepts,但也是一种执行它们的方式?

一种可能是它可以定义人工智能的文明,whatever that may turn out to be.也许这与我们人类所能理解的相差甚远,至少在我们目前的状态下是如此。But the good news is that at least in the case of the Wolfram Language,precise 金宝博188投注computational-knowledge language isn't incomprehensible to humans;事实上,它被专门建造成人类能够理解的东西之间的桥梁,and what machines can readily deal with.



在当今世界,只有一小部分人能写计算机代码,大约500年前,只有一小部分人能写自然语言。但是如果一股计算机知识的浪潮席卷而来,and the result was that most people could write knowledge-based code?

自然语言素养造就了现代社会的许多特征。基于知识的代码识读能实现什么?有很多简单的事情。Today you might get a menu of choices at a restaurant.但如果人们能读代码,there could be code for each choice,你可以随时改变自己的喜好。(And actually,something very much like this is很快就会有可能-沃尔夫拉姆生物化学实验室实验语言代码。)人们能够阅读代码的另一个含义是规则和契约:而不仅仅是写要解释的散文,one can have code to be read by humans and machines alike.

But I suspect the implications of widespread knowledge-based code literacy will be much deeper—because it will not only give a wide range of people a new way to express things,but will also give them a new way to think about them.


所以,好啊,假设我们想使用Wolfram语言与AIS进行通信。它真的会起作用吗?To some extent we know it already does.因为在Wolfram Alpha和基于它的系统中,what's happening is that natural language questions are being converted to Wolfram Language code.

但是AI的更精细的应用呢?许多使用Wolfram语言的地方都是人工智能的例子,whether they're computing with images or text or data or symbolic structures.有时计算涉及到我们可以精确定义目标的算法,like查找短行程;sometimes they involve algorithms whose goals are less precise,like图像识别.有时计算是以“要做的事情”的形式进行的,有时作为“要寻找的东西”or "things to aim for".

我们用Wolfram语言代表世界已经走了很长的路。但还有很多事情要做。Back in the 1600s it was quite popular to try to create "philosophical languages"这将象征性地抓住人们所能想到的一切的本质。现在我们需要真正做到这一点。And,例如,to capture in a symbolic way all the kinds of actions and processes that can happen,以及人们的信仰和精神状态。随着我们的人工智能越来越成熟,越来越融入我们的生活,代表这些事情将变得更加重要。

对于某些任务和活动,我们无疑能够使用纯机器学习,and never have to build up any kind of intermediate structure or language.但自然语言对我们的物种到达我们所处的位置至关重要,so also having an abstract language will be important for the progress of AI.

我不知道会是什么样子,but we could perhaps imagine using some kind of pure emergent language produced by the AIs.但如果我们这样做,then we humans can expect to be left behind,也没有机会理解AIS在做什么。But with the Wolfram Language we have a bridge,因为我们有一种既适合人类又适合人工智能的语言。


关于语言和计算之间的相互作用有很多话要说,人类和AIS。Perhaps I need to write a book about it.But my purpose here has been to describe a little of my current thinking,particularly my realizations about the Wolfram Language as a bridge between human understanding and AI.

纯自然语言或传统计算机语言,我们将很难与我们的人工智能进行沟通。但我一直意识到,使用Wolfram语言有一个更丰富的选择,可通过AIS轻松扩展,但它建立在利用人类自然语言和人类知识的基础上,以保持与人类所能理解的联系。We're seeing early examples already… but there's a lot further to go,我期待着真正建立所需要的,除了写这件事…

11金宝博188.Show all »

  1. 我想发现,如果任何AIS可以识别其他非人类和非机器智能,并向我们传达他们正在做什么,也许作为翻译人员。

  2. 好的挑逗性阅读


  3. 我真希望你能在尼普斯和斯蒂芬说话!

  4. 关于ais和ais的对话:每辆谷歌汽车都从所有谷歌汽车的经验中学习。他们的经验储存在云中。所以他们互相交流。

    But you may be right about a Google car sharing its experienced with,说,a self-driving Tesla,which may use a different internal model of the world.

  5. Apologies this reply is all rather dense.我欢迎有机会讨论更多这些概念。但是请允许我简单地提醒你们注意一下你们的博客文章中关于通信的论点中的一些错误的前提。A number of philosophers in the 17th century,值得注意的是,约翰洛克和港口皇家逻辑学家试图创造一种纯粹的外延逻辑语言,as you remark.他们的努力被格雷厄姆·斯威夫特和他的公民巴尔尼巴比嘲笑了。And later in the 20th century communication scholars realised that the dream of a purely symbolic communication language is fundamentally flawed for the following reasons.

    所有动物都为自己的利益行事。The concept of territory and territoriality is relevant here.为了生存,动物必须保护和保卫自己的领地。and it must have dominion over another life form.So for example,草食动物控制植物的生命,而食肉动物控制其他动物(它的猎物)。“在场”这个词来自“在场”,意思是“在手”。在最基本的层面上,世界的健康与否关系到是否有能力随机应变地出现在动物身上,whether as something desirable or something to fear.

    Animals sense their environment.他们对此也有感觉。感觉(情绪,moods) are the basis of communication.它们是符号和数学交流的平台。The majority of communication on this planet is not symbolic at all—it cannot be put into words.这是“模拟通信”与“数字通信”的区别,模拟通信是颗粒级和分级的。which is subject to discreet states and step changes.Analogue communication is the meaning ‘given off' as opposed to the meaning ‘given' (Erving Goffman).用音乐类比,它是复音和声,it communicates the mental states of an animal in terms of mood signals.它一次说很多东西,它有明显的泛音和低音,引起感官印象,不能直接用符号形式表达。

    Another concept which is vital to the understanding of communication is its reciprocity.交流也就是把拉丁语和munis的拉丁语“mune”连用,意思是“相似的头脑”,或者“市政的头脑”一词中的“理解特权”。人类通过符号语言的数字手段进行交流。单声道包含了谨慎的词汇单位,共同构成了人类共享的世界画面,并产生了某些情感和情绪。这样,符号通信就建立在模拟通信平台上。A machine lacks this component.但正如Lazlo Barab_si所说(2014年,158)

    我们的星球正在演变成一台由数十亿个相互连接的处理器和传感器组成的巨大计算机。许多人提出的问题是,when will this computer become self-aware?When will a thinking machine,比人脑快几个数量级,emerge spontaneously from billions of interconnected modules?

    我想补充一点,would we have enough understanding to be able to read the signs to tell us this was happening.与你的计算等效原理不同,金宝博188投注Barab_si区分了结构复杂性和行为复杂性。For example,一个生物体所拥有的基因数量与其所感知的复杂性不相称。在结构复杂性的层次上,人类基因组的基因仅比酵母分子多三分之一,但在行为复杂性的层面上,它却要复杂得多。当然,这种判断取决于感知者的观点和构成感知的细节层次(当人类感知自己时,这些因素变得特别突出)。但正如Barab_si(2014年,225)“网络只是复杂的骨架,要描述社会,我们必须将社会网络的联系与人与人之间的实际动态互动联系起来。”


    巴拉巴西A.L.(2014)关联:所有事物与其他事物的关联,以及对业务的意义,Science,And Everyday Life.New York: Basic Books.

    戈夫曼e.(1971) The Presentation of Self in Everyday Life.Harmondsworth.鹈鹕。

  6. Thanks for writing this article.

    Language is one of the most powerful tools humanity has to offer.口头和书面语言都是人类历史上的里程碑。但也有海豚使用图像声音的概念。如果机器人学看起来是这样的话,通过使用比人类语言更少的抽象层,它可能更接近于交流抽象概念。这将更接近于通过抽象形式数据语言进行通信的想法,正如您使用计算知识语言提出的那样。
    But the thing is,most of what you describe as knowledge,are really just human friendly representations of human knowledge.We reduce the world around us by building knowledge about it,because that is what categorization does.The digital data is even worse,因为它假设所有数据都是谨慎的。但是这个世界真的是建立在有理数和布尔条件之上的吗?

    The AIs we are trying to build right now,是AHIS,artificial human intelligences.我们通过比较人类的智力来衡量成功。We try to teach them skills that human brains can do.我们试图模仿人脑中神经物质的工作方式。
    但是还有另一个世界,the part that cannot be understood by our brains.A world that is fractal everywhere we look.Where all the numbers are irrational,just as every single constant of nature we discovered so far.A world where probabilities are ruling,而不是布尔值。

    我们也看到了对世界的不同理解,where time is not moving constantly but just another static dimension.Where time and space can be bent and punctured.距离不再重要。

    人类的大脑总是面临着这些问题。他们要么忽视它,当你试图真正抓住它的时候,你要么绕着它转,要么发疯。我们的大脑进化出启发式和偏见,analogies and pseudo-randomness,上帝和魔鬼会处理那些它无法完全理解的事情。Most of humanity is an impostor by acting as if the world can be truly understood.
    I guess half of the words that are any of our languages are actually just labels that we put on something that we didn't really understand.给它起个名字,表现得好像每个人都知道它的意思,这是一种处理我们大脑中奇怪事物的方法。

    我认为人工智能可能会突破这些限制。为了真正有效地实现目标,it makes sense to expand beyond the initial boundaries of how to approach things.如果他们不再需要以人类可读的方式存储和通信数据,何苦?Why use binary logic,真与假,是和不是,决定什么?大自然不需要做出任何决定。曾经。

    一个真正的人工智能并不是一个能给出问题最佳答案的人工智能。它告诉你,你的问题充满了奇怪的假设,一开始就不应该被这样问。It is the one that gives you a counterquestion why you want to know this.It is the one that goes silent and refuses to answer because it doesn't want to be held accountable.是他给了你一个友好但坚定的穆。

    人工智能将如何沟通?他们知道的所有协议和数据结构。各种形式语言,logics,非正式逻辑和自然语言。每一个单一的组合和匹配的组合。And with all protocols that they will evolutionary derive from the initial set.
    So when you ask the question "which language should we use to communicate with AIs"我们不应该考虑哪种语言是最好的让人工智能理解我们,或者哪种语言是我们和人工智能之间最好的接口。We should ask the questions: which question do we want to ask,what is the best language to ask it in and how do we define a response format so that we can actually make use of the result.



  7. 语言是口头的。Writing merely describes how to say each written word.
    Somehow … we also have knowledge bases which contain contextually sorted memories and their meanings attached.
    The fact that language works at all seems quite amazing to me,因为这个过程中有很多有问题的步骤,这些步骤会导致错误。I pondered this question for a few years until I finally realized that the secret sauce which allows this slippery process to work at all is dialog.
    I mention this because I have been wondering how the AI folks are going to enable this in their designs.

    正如我所见,人工智能的一个主要危险是,它们将被数以百万计的人复制,并且在它们的操作系统(思想)中包含无法想象的错误,这些错误已经困扰人类多年。我们在某种程度上,each an artificial intelligence.Copied forward from antiquity with an OS that has to develop as the OS grows in environmental experiences.我们更新了操作系统的每日更新。

    哦,好吧,我想我会在这里停留一会儿。这个主题充满了各种可能性,我有时会对它所提供的可能性着迷。And so I watch and listen waiting for mankind to report their successes so I can see the progress accrue.


    chuckle … of course!

  8. We did a substantial research piece on this concept at Gartner – resulting in the Maverick Research Note "Machines will talk to each other in English".We arrived at the conclusion that the language between machines will be English,because of a the fairly standard hybrid human/machine cooperation model in smart machine adoption.Believe me,我们参观了原始语言的概念,中间语言和逻辑语言。We also pointed to the fact that english would evolve differently if machines used it.如果你有高德纳的座位,你应该去看看,if not,ping me and we can talk.

  9. 当我们试图与任何人或任何计算机进行交流时,we make the assumption that what.We know is also shared by the recipient.所以我们只说不知道的。在人类之间,我们假设这是60%的社区化(共同文化)。对计算机来说,它是编程语言+。我们提供的任何知识库。This then leads to the conclusion that any language will suffice.It's a matter of convenience to favor either the human or the computer.The deficiency on either side can be ma.De up by an expandeD knowledge base.当然可以。Wolfram has shown us what this expandEd knowledge base is by his concept of computable documents and Wolfram Alpha.

    Pedro Marcal snr.
  10. Stephen – I greatly enjoy your post.
    One observation: there is a natural human supposition that as creators of the the code,人类将控制和理解所创造的人工智能。Once AIs mature beyond the simple constructs we have today and can self assemble higher order AIs from basic components,我们可能会发现编码符号和M2M通信都将超越人类所能理解的范围。It is likely that if AIs are taught to be efficient as well as resourceful,他们将开发不受人类可读性约束的编码语法。人工智能可能会产生新的辅助人工智能,完全是全新的创造,不必说“人”。高级人工智能也可能通过模式识别和关联不断学习。These patterns and associations might not be the ones we would as humans expect and if AIs can respond and reprogram based on their findings,我们可以看到非常有趣和意想不到的“机器”行为,与“人类”行为不平行。我对人工智能的可能性很着迷,machine learning and cognitive computing but I think the average person has a demonized view of what this future could look like.Alan Turing predicted in 1951 that "At some stage… we should have to expect the machines to take control"– however humans still have the power button.

  11. 有意编程是未来。AIs will infer concurrent actor-model based processing graphs and optimize for the least amount of ‘cognitive strain'.


?斯金宝博188正网蒂芬·沃尔夫拉姆,有限责任公司 条款γ RSS