其他
TED演讲：婴儿令人惊讶的逻辑思维
演讲题目：The surprisingly logical minds of babies
TED视频TED演讲稿
Twain meant it as a joke, of course, but he's right: There's something fascinating about science. From a few bones, we infer the existence of dinosuars. From spectral lines, the composition of nebulae. From fruit flies, the mechanisms of heredity, and from reconstructed images of blood flowing through the brain,吐温当然是在开玩笑，但他没说错：科学就是这么神奇。从几块骨头，我们能推测出恐龙的存在。从几条光谱带，我们能推测星云的构成物质。分析果蝇，我们能推导出遗传机制，分析大脑血液流动的图像，
or in my case, from the behavior of very young children, we try to say something about the fundamental mechanisms of human cognition. In particular, in my lab in the Department of Brain and Cognitive Sciences at MIT, I have spent the past decade trying to understand the mystery of how children learn so much from so little so quickly.或者，从我的研究方向来说，分析儿童的行为，我们尝试搞清楚人类认知的基本机制。尤其在我们麻省理工学院大脑和认知科学系实验室，过去十年我一直在研究一个问题，为什么小孩子能从无到有快速地学会很多东西。
Because, it turns out that the fascinating thing about science is also a fascinating thing about children, which, to put a gentler spin on Mark Twain, is precisely their ability to draw rich, abstract inferences rapidly and accurately from sparse, noisy data.因为，科学的奇妙之处，恰恰也是小孩子的奇妙之处，从马克·吐温的话引申出来，准确地说，就是他们都能从少量的、充满干扰的数据中迅速而准确地得出丰富的理论推断。我今天只举两个例子。
I'm going to give you just two examples today. One is about a problem of generalization, and the other is about a problem of causal reasoning. And although I'm going to talk about work in my lab, this work is inspired by and indebted to a field. I'm grateful to mentors, colleagues, and collaborators around the world.一个关于归纳总结，另一个关于因果推理。尽管我今天要谈的是我的实验室里的工作，但它的灵感来源于整个（认知科学）领域。我要感谢世界各地的导师、同事和合作者们。
Let me start with the problem of generalization. Generalizing from small samples of data is the bread and butter of science.我先从归纳总结开始讲起。从少量的数据样本进行归纳总结是科学的立身之本。
We poll a tiny fraction of the electorate and we predict the outcome of national elections. We see how a handful of patients responds to treatment in a clinical trial, and we bring drugs to a national market. But this only works if our sample is randomly drawn from the population.我们调查一小部分选民的投票结果，就能推测出大选结果。我们分析临床试验中一部分病人对治疗方案的反应，然后向全国市场推广新药。但这要求我们抽取样本要完全随机。
If our sample is cherry-picked in some way -- say, we poll only urban voters, or say, in our clinical trials for treatments for heart disease, we include only men -- the results may not generalize to the broader population.如果样本是刻意挑选的，比如说，只抽取城市选民，或者，在治疗心脏病的临床试验中，只抽取男性患者，那结果可能不适用于整个人群。
So scientists care whether evidence is randomly sampled or not, but what does that have to do with babies? Well, babies have to generalize from small samples of data all the time. They see a few rubber ducks and learn that they float, or a few balls and learn that they bounce.因此科学家非常重视样本的抽取是否随机，那婴儿会不会重视呢？实际上，婴儿一直在对少量数据样本进行归纳总结。他们见过几只橡胶鸭子，知道它们能浮起来，见过几个球，知道它们能在地上弹跳。
And they develop expectations about ducks and balls that they're going to extend to rubber ducks and balls for the rest of their lives. And the kinds of generalizations babies have to make about ducks and balls they have to make about almost everything: shoes and ships and sealing wax and cabbages and kings.他们对鸭子和球产生了预判并会在今后的人生中将这种预判延伸到（所有）橡胶鸭子和球身上。这种针对鸭子和球的归纳总结法，婴儿几乎要用在所有东西上：鞋子、船、封蜡、卷心菜和国王。
So do babies care whether the tiny bit of evidence they see is plausibly representative of a larger population? Let's find out. I'm going to show you two movies, one from each of two conditions of an experiment, and because you're going to see just two movies, you're going to see just two babies, and any two babies differ from each other in innumerable ways.那么婴儿会不会在乎他们看到的这几个样本是不是具有代表性呢？我们来看一看。我将给你们放两段视频，每一段各反映一个实验里的一种情况，因为只有两段视频，所以你们只能看到两个婴儿，而任意两个婴儿之间都是千差万别的。
But these babies, of course, here stand in for groups of babies, and the differences you're going to see represent average group differences in babies' behavior across conditions. In each movie, you're going to see a baby doing maybe just exactly what you might expect a baby to do, and we can hardly make babies more magical than they already are.当然，这两个婴儿，各代表一类婴儿，你们即将看到的差别，代表了婴儿在不同情况下普遍的行为差异。在每段视频中，婴儿的所作所为，可能会跟你所预期的一样，婴儿是如此神奇，可能超乎你的想象。
But to my mind the magical thing, and what I want you to pay attention to, is the contrast between these two conditions, because the only thing that differs between these two movies is the statistical evidence the babies are going to observe.但在我看来神奇的是，我也希望大家能注意到，就是两种情况之间的差别，因为两段视频唯一的不同之处就是婴儿需要观察的统计学证据。
We're going to show babies a box of blue and yellow balls, and my then-graduate student, now colleague at Stanford, Hyowon Gweon, is going to pull three blue balls in a row out of this box, and when she pulls those balls out, she's going to squeeze them, and the balls are going to squeak. And if you're a baby, that's like a TED Talk. It doesn't get better than that.我们会给婴儿看一个盒子，里面装满了蓝色和黄色的球，我当时的研究生学生，现在是斯坦福大学的同事，权孝媛。会从盒子里连续拿出三个蓝色的球，当她把球拿出来的时候，她会捏它们，球会发出声音。对孩子来说，这就像TED演讲。真的没什么区别。
But the important point is it's really easy to pull three blue balls in a row out of a box of mostly blue balls. You could do that with your eyes closed. It's plausibly a random sample from this population. And if you can reach into a box at random and pull out things that squeak, then maybe everything in the box squeaks.重要的一点是，从一个几乎全都是蓝色球的盒子里，连续拿出三个蓝色的球非常容易。闭上眼睛都能做到。这是一个真正的随机取样。如果你从一个盒子里随机取出来的东西能捏响，那也许这个盒子里所有的东西都能捏响。
So maybe babies should expect those yellow balls to squeak as well. Now, those yellow balls have funny sticks on the end, so babies could do other things with them if they wanted to. They could pound them or whack them. But let's see what the baby does.因此，婴儿也许会觉得黄色的球也能捏响。这些黄色的球在尾端有一根棍子，因此婴儿还可以对它做其他动作。比如说打它或者掰它。让我们来看婴儿会怎么做。
(Video) Hyowon Gweon: See this? (Ball squeaks) Did you see that? (Ball squeaks) Cool. See this one? (Ball squeaks) Wow.（视频）权孝媛：看到没？（球被捏响）听到了吗？（球被捏响）酷。看到这个球没？（球被捏响）哇。
Laura Schulz: Told you.劳拉·舒尔茨：我就说嘛。
(Video) HG: See this one? (Ball squeaks) Hey Clara, this one's for you. You can go ahead and play.（视频）权孝媛：看这个。（球被捏响）克拉拉，这个球给你。拿着玩吧。
LS: I don't even have to talk, right? All right, it's nice that babies will generalize properties of blue balls to yellow balls, and it's impressive that babies can learn from imitating us, but we've known those things about babies for a very long time.劳拉·舒尔茨：我都不必解释了，对吗？好的，婴儿能从蓝色球的特性推导出黄色球的特性这非常棒，而且婴儿通过模仿我们进行学习，令人印象深刻，但婴儿的这些特点我们早就知道了。
The really interesting question is what happens when we show babies exactly the same thing, and we can ensure it's exactly the same because we have a secret compartment and we actually pull the balls from there, but this time, all we change is the apparent population from which that evidence was drawn.真正有意思的是，我们将上述实验完全重复一遍，我们之所以能保证两次实验完全一样，是因为装球的箱子有一个隔层，实际上我们是从那个隔层里往外拿球，但是这一次，我们更改了样品库的外观，也就是说盒子里的球看起来不同了。
This time, we're going to show babies three blue balls pulled out of a box of mostly yellow balls, and guess what? You [probably won't] randomly draw three blue balls in a row out of a box of mostly yellow balls.这一次，我们还是给婴儿看三个蓝色的球，但是装球的箱子里几乎全是黄色的球，猜猜结果会怎样？从几乎全是黄色球的箱子里连续拿出三个蓝色的球，也许很难。
That is not plausibly randomly sampled evidence. That evidence suggests that maybe Hyowon was deliberately sampling the blue balls. Maybe there's something special about the blue balls. Maybe only the blue balls squeak. Let's see what the baby does.这不是令人信服的随机取样。也许孝媛是故意选的蓝色的球。也许蓝色的球有些特别之处。也许只有蓝色的球能捏响。我们来看婴儿会怎么做。
(Video) HG: See this? (Ball squeaks) See this toy? (Ball squeaks) Oh, that was cool. See? (Ball squeaks) Now this one's for you to play. You can go ahead and play.（视频）权孝媛：看到了吗？（球被捏响）再看这个。（球被捏响）哦，太酷了。看！（球被捏响）这个是给你的。拿去玩吧。
LS: So you just saw two 15-month-old babies do entirely different things based only on the probability of the sample they observed. Let me show you the experimental results.劳拉·舒尔茨：2个15个月大的婴儿，仅仅基于他们观察到的取样几率，做出了完全不同的反应。让我们来看一下实验结果。
On the vertical axis, you'll see the percentage of babies who squeezed the ball in each condition, and as you'll see, babies are much more likely to generalize the evidence when it's plausibly representative of the population than when the evidence is clearly cherry-picked.在纵轴上，你看到的是在不同情况下，会去捏球的婴儿的百分比，如图表所示，当婴儿认为取样具有代表性，而不是特意选取的时候，他们有更高几率去捏黄色的球。
And this leads to a fun prediction: Suppose you pulled just one blue ball out of the mostly yellow box. You [probably won't] pull three blue balls in a row at random out of a yellow box, but you could randomly sample just one blue ball. 这个结果能导致一个有趣的推测：假设你从几乎全是黄色球的箱子里拿出一个蓝色球。你也许很难从很多黄球的箱子里连续拿出三个蓝色球，但随机拿出一个还是有可能的。
That's not an improbable sample. And if you could reach into a box at random and pull out something that squeaks, maybe everything in the box squeaks.这不是一个小概率事件。如果你从箱子里随机抽出一个东西，而这个东西能捏响，那可能箱子里所有东西都能捏响。
So even though babies are going to see much less evidence for squeaking, and have many fewer actions to imitate in this one ball condition than in the condition you just saw, we predicted that babies themselves would squeeze more, and that's exactly what we found.因此，尽管婴儿们在接下来的“只拿一个球”的实验中，看到的证据更少，可模仿的动作也更少，但我们推测婴儿们捏球的几率会升高，结果正是如此。
So 15-month-old babies, in this respect, like scientists, care whether evidence is randomly sampled or not, and they use this to develop expectations about the world: what squeaks and what doesn't, what to explore and what to ignore.15个月大的婴儿，在这个实验中，跟科学家一样，十分看重取样是否真正随机，他们通过这种方法来发展对世界的预判：什么能捏响，什么不能，什么值得探究，什么可以忽略。
Let me show you another example now, this time about a problem of causal reasoning. And it starts with a problem of confounded evidence that all of us have, which is that we are part of the world. And this might not seem like a problem to you, but like most problems, it's only a problem when things go wrong.下面我们来看另一个实验，关于因果推论的实验。这个实验源于一个让我们所有人都感到困惑的事实：我们是这个世界的一部分。也许在你看来这根本不算个问题，但就像许多其他问题一样，只有问题出现时，它才算一个问题。
Take this baby, for instance. Things are going wrong for him. He would like to make this toy go, and he can't. I'll show you a few-second clip. And there's two possibilities, broadly: Maybe he's doing something wrong, or maybe there's something wrong with the toy.以下面这个婴儿为例。他就碰到了点问题。他想把玩具弄响，但是没有成功。我给你们放几秒视频。大体而言，有两种可能：也许他玩的方法不对，或者玩具坏了。
So in this next experiment, we're going to give babies just a tiny bit of statistical data supporting one hypothesis over the other, and we're going to see if babies can use that to make different decisions about what to do.因此在接下来的实验中，我们会给婴儿少量统计学数据，这些数据能支持某一种可能性，我们再看婴儿能否依据这些数据，作出不同的决定。
Here's the setup. Hyowon is going to try to make the toy go and succeed. I am then going to try twice and fail both times, and then Hyowon is going to try again and succeed, and this roughly sums up my relationship to my graduate students in technology across the board.实验是这样的。孝媛尝试弄响这个玩具，她成功了。然后我也开始玩，但两次都失败了，然后孝媛再次尝试，她又成功了，也许这是我跟孝媛在科技水平上差距的很好体现。
But the important point here is it provides a little bit of evidence that the problem isn't with the toy, it's with the person.这里的关键点在于，它提供了一点点证据证明问题不在于玩具，而在于人。
Some people can make this toy go, and some can't. Now, when the baby gets the toy, he's going to have a choice. His mom is right there, so he can go ahead and hand off the toy and change the person, but there's also going to be another toy at the end of that cloth, and he can pull the cloth towards him and change the toy. So let's see what the baby does.有的人能让玩具发出声音，有的人则不能。当婴儿拿到玩具之后，他要做出选择。他妈妈就在旁边，他可以将玩具交给妈妈，换一个人，同时在那块布的尽头放着另一个玩具，他可以将布拖过来，换一个玩具。我们来看看他会怎么做。
(Video) HG: Two, three. Go! (Music) LS: One, two, three, go! Arthur, I'm going to try again. One, two, three, go! YG: Arthur, let me try again, okay? One, two, three, go! (Music) Look at that. Remember these toys? See these toys? Yeah, I'm going to put this one over here, and I'm going to give this one to you. You can go ahead and play.（视频）孝媛：二、三，开始！（音乐）劳拉·舒尔茨：一、二、三，开始！亚瑟，我再试一次。一、二、三，开始！孝媛：亚瑟，让我再试一次，好吗？一、二、三，开始！（音乐）看啊。记得这些玩具吗？看到了吗？我把这个玩具放在这里，把这个玩具给你。你可以自己玩了。
LS: Okay, Laura, but of course, babies love their mommies. Of course babies give toys to their mommies when they can't make them work. So again, the really important question is what happens when we change the statistical data ever so slightly. This time, babies are going to see the toy work and fail in exactly the same order, but we're changing the distribution of evidence.劳拉·舒尔茨：好吧，劳拉，但是，小朋友都爱自己的妈妈呀。他玩不转玩具的时候肯定会把玩具交给妈妈。那么，让我们看看把这少量的统计学数据进行更换会怎么样。这一次，玩具响和不响的顺序跟刚才一样，但分布情况跟刚才不同。
This time, Hyowon is going to succeed once and fail once, and so am I. And this suggests it doesn't matter who tries this toy, the toy is broken. It doesn't work all the time. Again, the baby's going to have a choice. Her mom is right next to her, so she can change the person, and there's going to be another toy at the end of the cloth. Let's watch what she does.这一次，孝媛会成功一次，失败一次，我也一样。那就表明跟人没关系，是这个玩具有问题。它时好时坏。同样的，婴儿要做出选择。她妈妈就在她旁边，她可以换人来试，同样有另一个玩具放在布的另一头。我们来看她会如何选择。
(Video) HG: Two, three, go! (Music) Let me try one more time. One, two, three, go! Hmm.（视频）孝媛：二、三，开始！（音乐）我再试一次。一、二、三，开始！嗯？
LS: Let me try, Clara. One, two, three, go! Hmm, let me try again. One, two, three, go! (Music) HG: I'm going to put this one over here, and I'm going to give this one to you. You can go ahead and play.劳拉·舒尔茨：克拉拉，让我试一下吧。一、二、三，开始！嗯，我再试一次。一、二、三，开始！（音乐）孝媛：我把这个放在这边，把这个给你。你可以玩了。
LS: Let me show you the experimental results. On the vertical axis, you'll see the distribution of children's choices in each condition, and you'll see that the distribution of the choices children make depends on the evidence they observe.劳拉·舒尔茨：我们来看看实验结果。在纵轴上，显示的是在不同情况下婴儿所做选择的比例，我们可以看到，婴儿们做出的选择跟他们观察到的证据有关。
So in the second year of life, babies can use a tiny bit of statistical data to decide between two fundamentally different strategies for acting in the world: asking for help and exploring.因此，在出生后的第二年，婴儿已经可以利用少量统计数据来决定如何从两种不同的基本策略中做出选择从而在这个世界生存：求助和探索。
I've just shown you two laboratory experiments out of literally hundreds in the field that make similar points, because the really critical point is that children's ability to make rich inferences from sparse data underlies all the species-specific cultural learning that we do.我刚刚向大家展示的两个实验是从几百个类似实验中挑选出来的，它们得出了相似的结论，因为真正重要的一点是孩子们从很少的数据中推导出丰富结果的能力构成了我们研究物种特异性文化的基础。
Children learn about new tools from just a few examples. They learn new causal relationships from just a few examples. They even learn new words, in this case in American Sign Language.孩子能通过几个示范就掌握工具的用法。能通过几个例子就掌握新的因果关系。他们甚至能学会新的词语，这里我指的是美国手语。
I want to close with just two points. If you've been following my world, the field of brain and cognitive sciences, for the past few years, three big ideas will have come to your attention. The first is that this is the era of the brain.我想用两个观点来结束演讲。如果在过去几年，你一直在关注我们的领域，关注大脑和认知科学，那么你一定注意到了这三个观点。首先，现在是大脑的时代。
And indeed, there have been staggering discoveries in neuroscience: localizing functionally specialized regions of cortex, turning mouse brains transparent, activating neurons with light.实际上，神经系统科学已经取得了不错的进展：确定大脑皮层各区域的作用，让小白鼠的大脑透明化，利用光线触发神经元（活动）。
A second big idea is that this is the era of big data and machine learning, and machine learning promises to revolutionize our understanding of everything from social networks to epidemiology. And maybe, as it tackles problems of scene understanding and natural language processing, to tell us something about human cognition.第二个大的观点是现在是大数据和机器学习的时代，机器学习预示了我们对事物的理解将发生革命性的变化，无论是对社交网络还是流行病学。也许，随着它被用于场景理解和自然语言处理，能帮助我们更好地研究人类认知。
And the final big idea you'll have heard is that maybe it's a good idea we're going to know so much about brains and have so much access to big data, because left to our own devices, humans are fallible, we take shortcuts, we err, we make mistakes, we're biased, and in innumerable ways, we get the world wrong.最后一个你可能注意到的观点是我们能深入了解大脑，能深入运用大数据，是一件非常好的事情，因为人类天性随意，我们容易犯错，喜欢走捷径，我们闯祸，我们惹麻烦，我们心存偏见，而且从许多方面来讲，我们会错误理解这个世界。
I think these are all important stories, and they have a lot to tell us about what it means to be human, but I want you to note that today I told you a very different story. 我认为这些书都很重要，能帮我们理解身为人类意味着什么，但我想强调的是，今天我讲的是一个完全不同的故事。
It's a story about minds and not brains, and in particular, it's a story about the kinds of computations that uniquely human minds can perform, which involve rich, structured knowledge and the ability to learn from small amounts of data, the evidence of just a few examples.它讲的是思维而不是大脑，确切的说，是关于人类思维所特有的一种计算能力，这种能力让我们学识渊博，帮助我们从少量数据和证据中进行学习。
And fundamentally, it's a story about how starting as very small children and continuing out all the way to the greatest accomplishments of our culture, we get the world right.从本质上来说，这是一个关于成长的故事，小孩子如何一天天成长，取得巨大成就，为我们的文化做贡献，我们对世界的理解又是正确的。
Folks, human minds do not only learn from small amounts of data. Human minds think of altogether new ideas. Human minds generate research and discovery, and human minds generate art and literature and poetry and theater, and human minds take care of other humans: our old, our young, our sick. We even heal them.朋友们，人类的思维不光能从少量数据中进行学习。人类思维能提炼全新的观点。人类思维进行研究和发现，人类思维还能创作艺术、文学、诗歌和戏剧，人类思维还会关注其他人类：尊老爱幼，救死扶伤。让他们痊愈。
In the years to come, we're going to see technological innovations beyond anything I can even envision, but we are very unlikely to see anything even approximating the computational power of a human child in my lifetime or in yours.在未来几年，我们将看到超出我们想象的技术创新，但是我们很可能看不到哪怕仅仅是接近人类小孩计算能力的技术出现，可能我们的有生之年都看不到。
If we invest in these most powerful learners and their development, in babies and children and mothers and fathers and caregivers and teachers the ways we invest in our other most powerful and elegant forms of technology, engineering and design, we will not just be dreaming of a better future, we will be planning for one.如果我们对这些最强大的学习者和他们的发展进行投资，也就是对婴儿和儿童，对他们的父母，对他们的看护和老师，就像我们对技术、工程和设计等最强大和优雅的门类进行投资一样，那我们将不仅梦想着更好的未来，而是按计划在实现它。
Thank you very much.非常感谢大家。
TED演讲：改变自以为高效的“疯狂忙碌”
TED演讲：老师如何帮助学生渡过创伤
TED演讲：记忆是怎么形成和消失的？
TED演讲：女人不一定要生孩子！
更多资源：
更多资源：
欢迎在文章底部写下精彩留言哦！
欢迎在文章底部写下精彩留言哦！
特 别 声 明
公众号内提供的资源均为免费分享，大多来源于互联网，仅供于学习&交流，如有侵权、不妥请联系我们删除。同时，由于网络政策不确定性，相关资源随时会被删除或屏蔽，欲观从速，一旦被删则缘尽不可恢复！
主编/慕容风 监制/Richard
视频编辑/李小可 安雅制图/慕容秋水©爱天涯
特 别 声 明
公众号内提供的资源均为免费分享，大多来源于互联网，仅供于学习&交流，如有侵权、不妥请联系我们删除。同时，由于网络政策不确定性，相关资源随时会被删除或屏蔽，欲观从速，一旦被删则缘尽不可恢复！
主编/慕容风 监制/Richard特 别 声 明
公众号内提供的资源均为免费分享，大多来源于互联网，仅供于学习&交流，如有侵权、不妥请联系我们删除。同时，由于网络政策不确定性，相关资源随时会被删除或屏蔽，欲观从速，一旦被删则缘尽不可恢复！
视频编辑/李小可 安雅制图/慕容秋水©爱天涯