Nicholas Christakis: How social networks predict epidemics

翻译人员: Ming Wu 校对人员: Xiaoqiao Xie

For the last 10 years, I've been spending my time trying to figure out

过去十年间，我一直在想

how and why human beings

人们是怎样把自己放在社交网络中的

assemble themselves into social networks.

以及他们为什么要这么做。

And the kind of social network I'm talking about

这里的社交网络

is not the recent online variety,

不是最近网上最近流行的那种

but rather, the kind of social networks

而是，

that human beings have been assembling for hundreds of thousands of years,

自从人类从非洲大陆出现之后，

ever since we emerged from the African savannah.

人们几十万年来进行的社交活动。

So, I form friendships and co-worker

比如说，

and sibling and relative relationships with other people

我和其他人建立友谊，同事，兄弟，亲戚的关系，

who in turn have similar relationships with other people.

其他人也和另外其他人建立类似的关系。

And this spreads on out endlessly into a distance.

这样的关系无尽止地延伸出去。

And you get a network that looks like this.

这样你就有了一个像这样的网络。

Every dot is a person.

网络中的每一个点就是一个人。

Every line between them is a relationship between two people --

连接两点的每一条线就是两个人之间的关系 --

different kinds of relationships.

不同类型的关系。

And you can get this kind of vast fabric of humanity,

这样就得到了这种巨大的人际关系网，

in which we're all embedded.

我们都交织在网中。

And my colleague, James Fowler and I have been studying for quite sometime

我的同事，詹姆斯.福乐，和我一起已经研究了好些时间

what are the mathematical, social,

什么是支配这些网络的数学，社交，

biological and psychological rules

生物和心理规则

that govern how these networks are assembled

以及

and what are the similar rules

什么是基本的规则

that govern how they operate, how they affect our lives.

来支配这些网络的运作，和如何影响我们的生活。

But recently, we've been wondering

最近，我们在研究

whether it might be possible to take advantage of this insight,

是否有可能利用这种认识

to actually find ways to improve the world,

来发现改善这个世界的方法，

to do something better,

做一些好事

to actually fix things, not just understand things.

解决一些问题，而不只是理解而已。

So one of the first things we thought we would tackle

所以我们认为需要解决的一件首要的事情

would be how we go about predicting epidemics.

就是怎样预测传染病。

And the current state of the art in predicting an epidemic --

现在预测传染病的方法 --

if you're the CDC or some other national body --

如果是美国疾病控制预防中心或者其他国家级的机构 --

is to sit in the middle where you are

就是呆在原地

and collect data

从医生和实验室

from physicians and laboratories in the field

收集数据

that report the prevalence or the incidence of certain conditions.

来分析疾病的流行性和发病率。

So, so and so patients have been diagnosed with something,

所以，（如果）一些病人在一个地方被诊断了得病，

or other patients have been diagnosed,

或者其他病人在另一个地方得到诊断，

and all these data are fed into a central repository, with some delay.

所有这些数据，一定的延迟之后，都送到一个中心数据库。

And if everything goes smoothly,

如果一切顺利，

one to two weeks from now

一两个星期之后，

you'll know where the epidemic was today.

你就会知道发生在今天的传染病在何处。

And actually, about a year or so ago,

实际上，一年前左右，

there was this promulgation

曾经有过这样的一个，google流感趋势的想法

of the idea of Google Flu Trends, with respect to the flu,

关于流感，

where by looking at people's searching behavior today,

通过观察人们今天的搜索行为，

we could know where the flu --

我们能知道流感的发病区...

what the status of the epidemic was today,

传染病今天的状况，

what's the prevalence of the epidemic today.

以及传播的趋势。

But what I'd like to show you today

但我今天想给你展示的

is a means by which we might get

是一种方法

not just rapid warning about an epidemic,

通过这个方法，我们不只是得到关于传染病的警示，

but also actually

而且

early detection of an epidemic.

也能够及早发现传染病。

And, in fact, this idea can be used

事实上，这个想法不只能

not just to predict epidemics of germs,

预测病毒的传播，

but also to predict epidemics of all sorts of kinds.

也能预测很多事情的传播。

For example, anything that spreads by a form of social contagion

比如说，任何以社交形式传播的事情，

could be understood in this way,

都能用这种方法来理解，

from abstract ideas on the left

从左边这些抽象的事物，

like patriotism, or altruism, or religion

像爱国主义，或利他主义，或者宗教，

to practices

到具体的事物，

like dieting behavior, or book purchasing,

像饮食行为，或者买书，

or drinking, or bicycle-helmet [and] other safety practices,

或饮酒，或自行车头盔和其他的一些安全措施，

or products that people might buy,

或者人们可能买的产品，

purchases of electronic goods,

电子书的购买，

anything in which there's kind of an interpersonal spread.

任何能在人们之间传播的事情。

A kind of a diffusion of innovation

一种创新的传播

could be understood and predicted

可以用我即将展示的方法

by the mechanism I'm going to show you now.

来理解和预测。

So, as all of you probably know,

正如你们所有人也许知道的，

the classic way of thinking about this

考虑这个问题的传统方法

is the diffusion-of-innovation,

是创新扩散，

or the adoption curve.

或创新采用曲线。

So here on the Y-axis, we have the percent of the people affected,

这儿，Y轴上显示的是受到影响的人们的百分比，

and on the X-axis, we have time.

x轴上显示的是时间。

And at the very beginning, not too many people are affected,

一开始，受到影响的人不多，

and you get this classic sigmoidal,

得到的是S状的分布，

or S-shaped, curve.

或者说S形状的曲线。

And the reason for this shape is that at the very beginning,

形成这种形状的原因是这样的，在一开始，

let's say one or two people

假设一两个人

are infected, or affected by the thing

受到影响，

and then they affect, or infect, two people,

然后他们去影响两个人，

who in turn affect four, eight, 16 and so forth,

然后就是四个，八个，十六个，等等，

and you get the epidemic growth phase of the curve.

这样你就得到了这个曲线的传播增长阶段。

And eventually, you saturate the population.

最终，这个群体就饱和了。

There are fewer and fewer people

可以被影响的人

who are still available that you might infect,

就越来越少，

and then you get the plateau of the curve,

这时你就得到了曲线的平顶部分，

and you get this classic sigmoidal curve.

这样就形成了经典的S状曲线。

And this holds for germs, ideas,

这种方法可以用于病毒，观点，

product adoption, behaviors,

产品推广，行为，

and the like.

以及其他类似的情况。

But things don't just diffuse in human populations at random.

但是这些事物在人群中并非是随机传播的。

They actually diffuse through networks.

他们实际上是通过网络传播。

Because, as I said, we live our lives in networks,

正如我所说的，因为我们在网络中生存，

and these networks have a particular kind of a structure.

这些网络有一种特殊的结构。

Now if you look at a network like this --

现在，如果你看像这样的一个网络。。。

this is 105 people.

有105人。

And the lines represent -- the dots are the people,

这些线代表。。。这些点是人，

and the lines represent friendship relationships.

这些线代表朋友关系。

You might see that people occupy

你也许看到在这个网络中

different locations within the network.

人们占据了不同的地点。

And there are different kinds of relationships between the people.

人们之间也有不同的关系。

You could have friendship relationships, sibling relationships,

这些关系可以是朋友关系，同胞关系，

spousal relationships, co-worker relationships,

配偶关系，同事关系，

neighbor relationships and the like.

邻居关系，以及类似的关系。

And different sorts of things

不同的事物

spread across different sorts of ties.

通过不同的联系来传播。

For instance, sexually transmitted diseases

比如说，性病

will spread across sexual ties.

通过性关系传播。

Or, for instance, people's smoking behavior

或者，比如说，人们的吸烟行为

might be influenced by their friends.

可能会受到他们朋友的影响。

Or their altruistic or their charitable giving behavior

或者利他主义以及慈善施舍行为

might be influenced by their coworkers,

可能会受到同事的影响，

or by their neighbors.

或者邻居的影响。

But not all positions in the network are the same.

但并非网络中所有的位置都是一样的。

So if you look at this, you might immediately grasp

所以你看这里，就能立即理解

that different people have different numbers of connections.

不同的人有不同数量的连接。

Some people have one connection, some have two,

有些人有一个连接，有些人有两个，

some have six, some have 10 connections.

有些人有六个，有些人有十个连接。

And this is called the "degree" of a node,

这就叫做结点的度数，

or the number of connections that a node has.

或者一个结点有的连接数目。

But in addition, there's something else.

但是，还有些别的东西。

So, if you look at nodes A and B,

如果你看结点A和B，

they both have six connections.

都有六个连接关系。

But if you can see this image [of the network] from a bird's eye view,

但是如果拔高来看，

you can appreciate that there's something very different

你就能理解A和B是

about nodes A and B.

非常不一样。

So, let me ask you this -- I can cultivate this intuition by asking a question --

让我来问你 -- 用一个问题来说明这个直觉 ---

who would you rather be

如果有个致命的病毒正在网络中传播，

if a deadly germ was spreading through the network, A or B?

你更愿意是哪一个，A还是B?

(Audience: B.) Nicholas Christakis: B, it's obvious.

(观众：B) 尼古拉斯·克里斯塔吉斯: B， 这很明显。

B is located on the edge of the network.

B在网络的边缘。

Now, who would you rather be

现在，如果有个非常有料的流言在网络中传播，

if a juicy piece of gossip were spreading through the network?

你愿意是哪一个？

A. And you have an immediate appreciation

A。 你很快就看到

that A is going to be more likely

A更有可能

to get the thing that's spreading and to get it sooner

更快地得到正在传播的事物

by virtue of their structural location within the network.

因为他们在网络中的位置。

A, in fact, is more central,

A，实际上，（位置）更加中心，

and this can be formalized mathematically.

这个能在数学上来表示。

So, if we want to track something

如果我们要追踪

that was spreading through a network,

在网络中传播的事物，

what we ideally would like to do is to set up sensors

理想情况下我们会想设置感应器

on the central individuals within the network,

在网络的中心人物上，

including node A,

包括结点A,

monitor those people that are right there in the middle of the network,

以此来观察网络中心的人们的活动，

and somehow get an early detection

从而能够做到及早探测

of whatever it is that is spreading through the network.

正在网络中传播的东西。

So if you saw them contract a germ or a piece of information,

也就是说，假如你看到网络中心的人们感染病毒或得到了一些信息，

you would know that, soon enough,

你就能知道，很快

everybody was about to contract this germ

所有人都会被传染这种病毒

or this piece of information.

或得到这个消息。

And this would be much better

这种方法

than monitoring six randomly chosen people,

比不考虑群体的结构，监测六个随机选择的人，

without reference to the structure of the population.

要好的多。

And in fact, if you could do that,

实际上，如果能够这样做，

what you would see is something like this.

你就会看到像这样的情况。

On the left-hand panel, again, we have the S-shaped curve of adoption.

在左边，我们有S形状的传播曲线。

In the dotted red line, we show

这条红色的点线，我们表示的是

what the adoption would be in the random people,

在随机人群中的感染率，

and in the left-hand line, shifted to the left,

左手的线条，向左移动，

we show what the adoption would be

表现的是

in the central individuals within the network.

在网络的中心群体中的感染率。

On the Y-axis is the cumulative instances of contagion,

Y轴上是感染个体的累计总数，

and on the X-axis is the time.

X轴上是时间。

And on the right-hand side, we show the same data,

右边，我们显示同样的数据，

but here with daily incidence.

但是在每天的个体数。

And what we show here is -- like, here --

我们在这里要显示的是 -- 比如说，这里 --

very few people are affected, more and more and more and up to here,

很少的人受到影响，然后逐渐增多到这里，

and here's the peak of the epidemic.

这里是感染的高峰。

But shifted to the left is what's occurring in the central individuals.

但是移到左边，是在中心群体中的发展趋势。

And this difference in time between the two

两者之间在时间上的区别

is the early detection, the early warning we can get,

正是我们能够得到

about an impending epidemic

关于传染病在人群中的

in the human population.

早期预测， 早期示警。

The problem, however,

然而，这个方法的难处在于，

is that mapping human social networks

测绘人类的社交关系网

is not always possible.

并不总是可能的。

It can be expensive, not feasible,

这很昂贵，[很难]，

unethical,

不正当，

or, frankly, just not possible to do such a thing.

或者坦白说，就是没可能做这样的事情。

So, how can we figure out

那么，我们怎样能弄清楚

who the central people are in a network

哪些人在网络中心

without actually mapping the network?

而不需要通过测绘整个网络呢？

What we came up with

我们想出来的方法

was an idea to exploit an old fact,

是采用了一个古老的事实，

or a known fact, about social networks,

或者说关于社交网络的已知事实，

which goes like this:

这个事实是这样的：

Do you know that your friends

你知道你的朋友

have more friends than you do?

有比你更多的朋友吗？

Your friends have more friends than you do,

你的朋友比你有更多的朋友。

and this is known as the friendship paradox.

这个称为朋友的悖论。

Imagine a very popular person in the social network --

想像有一个在社交网络中非常受欢迎的人物--

like a party host who has hundreds of friends --

就像一个聚会的主持有几百个朋友--

and a misanthrope who has just one friend,

而一个憎恨人类的人只有一个朋友，

and you pick someone at random from the population;

然后你随机从人群中选个人；

they were much more likely to know the party host.

他们更有可能认识聚会的主持。

And if they nominate the party host as their friend,

如果他们选择晚会主持作为他们的朋友，

that party host has a hundred friends,

那么这个聚会主持就有一百个朋友，

therefore, has more friends than they do.

因此，就有比他们更多的朋友。

And this, in essence, is what's known as the friendship paradox.

这个就称为朋友悖论。

The friends of randomly chosen people

随机选择的人群的朋友

have higher degree, and are more central

比随机人群本身，

than the random people themselves.

有更高的（关系）度数，并且更加中心。

And you can get an intuitive appreciation for this

你能对这个理论有一个本能的理解

if you imagine just the people at the perimeter of the network.

如果想像网络周边的人群。

If you pick this person,

如果你选择（网络周边的）这个人

the only friend they have to nominate is this person,

他们能选择的唯一朋友就是这个人，

who, by construction, must have at least two

而这个，在这个结构中，一定有至少两个朋友，

and typically more friends.

通常情况下，有更多的朋友。

And that happens at every peripheral node.

这种情况发生在每个周边结点上。

And in fact, it happens throughout the network as you move in,

实际上，每当你加入一个网络的时候这个情况都会发生，

everyone you pick, when they nominate a random --

你选择的每个人，当他们随机选择。。。

when a random person nominates a friend of theirs,

当任意一个人选择他们的一个朋友，

you move closer to the center of the network.

你就向网络中心移动。

So, we thought we would exploit this idea

所以，我们想利用这个概念

in order to study whether we could predict phenomena within networks.

来研究是否能预测网络的现象。

Because now, with this idea

因为，有了这个概念，

we can take a random sample of people,

我们就选择一个随机人群，

have them nominate their friends,

让他们提供他们的朋友，

those friends would be more central,

他们的朋友就更加中心，

and we could do this without having to map the network.

这样我们就能选择网络的中心，而不用描绘整个网络。

And we tested this idea with an outbreak of H1N1 flu

我们用这个想法来测试H1N1流感的爆发

at Harvard College

在哈佛大学

in the fall and winter of 2009, just a few months ago.

2009年的秋冬，就是几个月前。

We took 1,300 randomly selected undergraduates,

我们随机选择了1300本科学生，

we had them nominate their friends,

让他们推举他们的朋友，

and we followed both the random students and their friends

然后我们跟踪随机的学生人群和他们的朋友

daily in time

每天按时

to see whether or not they had the flu epidemic.

观察他们是否传染上流感。

And we did this passively by looking at whether or not they'd gone to university health services.

我们观察的方法是看他们有没有去过大学健康服务机构。

And also, we had them [actively] email us a couple of times a week.

并且我们要求他们一个星期给我们发几次电子邮件。

Exactly what we predicted happened.

我们的预测一点不错的发生了。

So the random group is in the red line.

这个随机组在这个红线上。

The epidemic in the friends group has shifted to the left, over here.

朋友组的传染移到左边，这里

And the difference in the two is 16 days.

中间相差了16天。

By monitoring the friends group,

通过检测朋友组，

we could get 16 days advance warning

我们能够得到16天的预先示警

of an impending epidemic in this human population.

在这个人群的关于这个传染病的传播。

Now, in addition to that,

现在，在这个基础上，

if you were an analyst who was trying to study an epidemic

如果你是分析师，要研究一种传染病

or to predict the adoption of a product, for example,

或者预测一个产品的推广，

what you could do is you could pick a random sample of the population,

你能做的是选择一个随机的人群，

also have them nominate their friends and follow the friends

让他们任命他们的朋友，然后跟踪他们的朋友，

and follow both the randoms and the friends.

跟踪随机组和朋友组。

Among the friends, the first evidence you saw of a blip above zero

在朋友组中，你看到的第一个零上的尖峰信号

in adoption of the innovation, for example,

关于，比如说，创新科技的采纳，

would be evidence of an impending epidemic.

就是即将来临的流行趋势的信号。

Or you could see the first time the two curves diverged,

或者你能看到两条曲线第一次分离的地方，

as shown on the left.

就像左边显示的。

When did the randoms -- when did the friends take off

朋友组什么时候开始

and leave the randoms,

与随机组分离，

and [when did] their curve start shifting?

他们的曲线什么时候开始偏移？

And that, as indicated by the white line,

正如白线显示的，

occurred 46 days

发生在

before the peak of the epidemic.

流行高峰的46天之前。

So this would be a technique

通过这个技术

whereby we could get more than a month-and-a-half warning

我们能得到关于流感在特定人群中传播

about a flu epidemic in a particular population.

一个半月以上的预先示警。

I should say that

我应该说

how far advanced a notice one might get about something

能多早得到关于一些事情的消息

depends on a host of factors.

取决于很多因素。

It could depend on the nature of the pathogen --

它也许取决于病原体的本质---

different pathogens,

不同的病原体，

using this technique, you'd get different warning --

使用这种技术，你可能得到不同的示警---

or other phenomena that are spreading,

或者其他一些传播的现象，

or frankly, on the structure of the human network.

或者，直接的说，在人类网络的结构中。

Now in our case, although it wasn't necessary,

现在，在我们的例子中，尽管不是很必要，

we could also actually map the network of the students.

我们也能够描绘这个学生网络。

So, this is a map of 714 students

这是714个学生的映射图

and their friendship ties.

和他们朋友联系。

And in a minute now, I'm going to put this map into motion.

很快，我要使这个图动起来。

We're going to take daily cuts through the network

我们要通过这个网络作每日监控

for 120 days.

120天。

The red dots are going to be cases of the flu,

红点将会是流感的传染者，

and the yellow dots are going to be friends of the people with the flu.

黄点就是流感传染这人的朋友。

And the size of the dots is going to be proportional

这些点的大小

to how many of their friends have the flu.

和他们得流感朋友的数目成正比。

So bigger dots mean more of your friends have the flu.

越大的点意味着更多的朋友得了流感。

And if you look at this image -- here we are now in September the 13th --

你看这个图 --- 这儿是九月十三号 ---

you're going to see a few cases light up.

你看到几个病例出现。

You're going to see kind of blooming of the flu in the middle.

在中间你就会看到流感开始爆发。

Here we are on October the 19th.

这儿是十月十九日。

The slope of the epidemic curve is approaching now, in November.

传播曲线的坡度开始临近，在十一月。

Bang, bang, bang, bang, bang -- you're going to see lots of blooming in the middle,

砰，砰，砰，砰，砰，你将看到在中间的很多地方爆发，

and then you're going to see a sort of leveling off,

然后你会看到情况稳定下来，

fewer and fewer cases towards the end of December.

到十二月底就越来越少的病例发生。

And this type of a visualization

这样的图形表示

can show that epidemics like this take root

能显示，像这样的传染病先

and affect central individuals first,

影响中心个体

before they affect others.

在影响别人之前。

Now, as I've been suggesting,

现在，如我所说，

this method is not restricted to germs,

这个方法并不局限于病毒，

but actually to anything that spreads in populations.

实际上可以用于任何在人群中传播的事物。

Information spreads in populations,

信息在人群中传播。

norms can spread in populations,

规范在人群中传播。

behaviors can spread in populations.

行为能在人群中传播

And by behaviors, I can mean things like criminal behavior,

我说的行为，就是像犯罪的行为

or voting behavior, or health care behavior,

或选举行为，或者保健行为，

like smoking, or vaccination,

像抽烟，或免疫，

or product adoption, or other kinds of behaviors

或产品推广，或者其他类型的行为

that relate to interpersonal influence.

和人际之间影响相关的性为。

If I'm likely to do something that affects others around me,

如果我想做些事情来影响我周围的人，

this technique can get early warning or early detection

这个技术能得到早期示警，或早期预测，

about the adoption within the population.

关于人群的采纳。

The key thing is that for it to work,

要这个技术起作用，关键在于，

there has to be interpersonal influence.

人际之间的影响必须存在。

It cannot be because of some broadcast mechanism

它不能是像一些传播机制

affecting everyone uniformly.

统一地影响每一个人。

Now the same insights

现在同样的观察

can also be exploited -- with respect to networks --

可以用上 --- 关于网络 ---

can also be exploited in other ways,

能够用其他的方法来利用，

for example, in the use of targeting

比如，可以用来追踪

specific people for interventions.

特殊的人群。

So, for example, most of you are probably familiar

比如说，你们大部分人可能听过

with the notion of herd immunity.

群体免疫的概念。

So, if we have a population of a thousand people,

如果我们有一个一千人的群体，

and we want to make the population immune to a pathogen,

我们想使这个群体对某种病原体免疫，

we don't have to immunize every single person.

我们并不需要给每个人打免疫预防针。

If we immunize 960 of them,

如果我们使960人免疫，

it's as if we had immunized a hundred [percent] of them.

效果和使所有人免疫差不多。

Because even if one or two of the non-immune people gets infected,

因为即使一两个没有免疫的人感染了，

there's no one for them to infect.

也没有其他人让他们感染。

They are surrounded by immunized people.

这两个人周围的人都已经免疫。

So 96 percent is as good as 100 percent.

所以百分之96和百分之百效果一样好。

Well, some other scientists have estimated

一些其他的科学家已经预测了

what would happen if you took a 30 percent random sample

可能发生的情况，如果你从这一千人中取百分之三十的随机样本

of these 1000 people, 300 people and immunized them.

也就是三百个人，并且使他们免疫。

Would you get any population-level immunity?

这样能不能使整个群体免疫？

And the answer is no.

答案是不可能。

But if you took this 30 percent, these 300 people

但是如果你选择百分之三十的这三百人，

and had them nominate their friends

让他们举出他们的朋友

and took the same number of vaccine doses

然后使用同样数量的免疫针

and vaccinated the friends of the 300 --

使得这三百人的朋友免疫，

the 300 friends --

这三百个朋友，

you can get the same level of herd immunity

你就能达到群体免疫的效果

as if you had vaccinated 96 percent of the population

就好像给百分之九十六的人打预防针的效果一样

at a much greater efficiency, with a strict budget constraint.

同时效率更高，并且花费更少。

And similar ideas can be used, for instance,

同样的想法可以用于，比如说，

to target distribution of things like bed nets

解决像床罩这种物品在发展中国家

in the developing world.

的派发和分布。

If we could understand the structure of networks in villages,

如果我们了解村庄的网络结构，

we could target to whom to give the interventions

我们就能选择介入的目标

to foster these kinds of spreads.

来促进这些过程的进行。

Or, frankly, for advertising with all kinds of products.

或者，更加直接的说，来促销所有的产品。

If we could understand how to target,

如果我们能理解如何选择目标，

it could affect the efficiency

就能影响到我们达到目标

of what we're trying to achieve.

的效率。

And in fact, we can use data

实际上，我们能使用来源于各种渠道的

from all kinds of sources nowadays [to do this].

数据[来应用这个方法]。

This is a map of eight million phone users

这里是一个欧洲国家八百万人的

in a European country.

电话用户的网络图。

Every dot is a person, and every line represents

每一个点就是一个人，每一条线代表

a volume of calls between the people.

人们之间的通话数量。

And we can use such data, that's being passively obtained,

我们可以利用这些数据，被动方式得到的数据，

to map these whole countries

来描绘整个国家

and understand who is located where within the network.

从而了解那些人处在网络的中心。

Without actually having to query them at all,

不用实际上去询问每个人，

we can get this kind of a structural insight.

我们能得到这种的结构。

And other sources of information, as you're no doubt aware

其他来源的信息，你肯定也知道，

are available about such features, from email interactions,

也能提供这样的信息，例如电子邮件交互，

online interactions,

在线交流，

online social networks and so forth.

在线社交网络，等等。

And in fact, we are in the era of what I would call

实际上，我们处在一个我称为

"massive-passive" data collection efforts.

“大量被动”数据收集的时代。

They're all kinds of ways we can use massively collected data

有很多种不同的方法我们能使用大量收集的数据

to create sensor networks

来建立传感网络

to follow the population,

跟踪人群，

understand what's happening in the population,

了解在人群中正在发生的事件，

and intervene in the population for the better.

从而可以更好的介入。

Because these new technologies tell us

因为这些新技术告诉我们

not just who is talking to whom,

不仅仅是那些人与那些人交流，

but where everyone is,

同时也告诉我们每个人处在什么位置，

and what they're thinking based on what they're uploading on the Internet,

根据他们上传到互联网上的东西来知道他们的想法，

and what they're buying based on their purchases.

他们的购物记录告诉我们他们买了什么。

And all this administrative data can be pulled together

所有这些管理数据能一起使用处理

and processed to understand human behavior

来了解人类的行为

in a way we never could before.

以一种前所未能的方式。

So, for example, we could use truckers' purchases of fuel.

比如说，我们能用卡车司机的购油记录。

So the truckers are just going about their business,

卡车司机处理他们的生意

and they're buying fuel.

他们要买汽油作燃料。

And we see a blip up in the truckers' purchases of fuel,

我们看到卡车司机购油数据的零上尖峰信号，

and we know that a recession is about to end.

我们就知道经济衰退大概要结束了。

Or we can monitor the velocity

或者我们能监测

with which people are moving with their phones on a highway,

人们在高速公路上带着电话移动的速度，

and the phone company can see,

电话公司能看到，

as the velocity is slowing down,

如果速度慢下来，

that there's a traffic jam.

那么就发生了交通堵塞。

And they can feed that information back to their subscribers,

他们能把这个信息发给他们的用户，

but only to their subscribers on the same highway

只发给在同一条高速公路上

located behind the traffic jam!

处于交通堵塞地点之前的用户！

Or we can monitor doctors prescribing behaviors, passively,

或者我们监测医生开药的行为，以被动的形式，

and see how the diffusion of innovation with pharmaceuticals

看看在医生网络中

occurs within [networks of] doctors.

制药公司的新药的发行如何。

Or again, we can monitor purchasing behavior in people

或者，我们能监测人们的购物行为，

and watch how these types of phenomena

看看这些种类的现象

can diffuse within human populations.

在人群中是怎样传播的。

And there are three ways, I think,

我想，有三种方式，

that these massive-passive data can be used.

这些大量被动的数据能被收集。

One is fully passive,

一个方式是完全被动，

like I just described --

如我刚刚描述的 ---

as in, for instance, the trucker example,

例如在卡车司机的例子中，

where we don't actually intervene in the population in any way.

我们不需要以任何方式干涉这个群体的行为。

One is quasi-active,

另一种是类似于主动的方式，

like the flu example I gave,

比如说像我说的流感的例子，

where we get some people to nominate their friends

我们选一些人来举出他们的朋友

and then passively monitor their friends --

然后被动地监测他们的朋友 ---

do they have the flu, or not? -- and then get warning.

他们是不是感染了流感？ -- 然后得到示警。

Or another example would be,

或者另一个例子，

if you're a phone company, you figure out who's central in the network

如果你是电话公司，你能弄清楚谁在网络的中心，

and you ask those people, "Look, will you just text us your fever every day?

然后你可以问这些人，“你们能不能把你们每天的发烧情况给我们发过来？

Just text us your temperature."

只要发体温度数。”

And collect vast amounts of information about people's temperature,

然后收集人体体温的大量数据，

but from centrally located individuals.

但是只是网络中心个体的信息。

And be able, on a large scale,

这样就能，大面积地，

to monitor an impending epidemic

监测即将来临的传染病

with very minimal input from people.

而只需要人们给出非常少量的信息。

Or, finally, it can be more fully active --

最后的一种方式，就更加主动 ---

as I know subsequent speakers will also talk about today --

我知道随后的演讲者今天会说到的 --

where people might globally participate in wikis,

人们在哪儿参与维基，

or photographing, or monitoring elections,

摄影，看选举，

and upload information in a way that allows us to pool

上载信息，这样可以让我们收集

information in order to understand social processes

数据，来了解社交过程

and social phenomena.

和社会现象。

In fact, the availability of these data, I think,

实际上，我认为，这些数据的可用性，

heralds a kind of new era

预示了一个新的纪元

of what I and others would like to call

也就是我们所说的

"computational social science."

“计算社会学”。

It's sort of like when Galileo invented -- or, didn't invent --

这有点像伽利略发明的 -- 不是发明 --

came to use a telescope

使用望远镜

and could see the heavens in a new way,

能用一种新的方式看到天空，

or Leeuwenhoek became aware of the microscope --

或者莱文胡克开始了解微观世界 ---

or actually invented --

发明了显微镜 ---

and could see biology in a new way.

而能以新的方式审视生物学。

But now we have access to these kinds of data

但现在我们能够得到这些数据

that allow us to understand social processes

这使得我们能了解社交过程

and social phenomena

和社会现象

in an entirely new way that was never before possible.

以一种前所未能的新方式。

And with this science, we can

通过这门科学，我们能

understand how exactly

准确了解

the whole comes to be greater

整体是怎样优于

than the sum of its parts.

局部的总和。

And actually, we can use these insights

我们能用这些知识

to improve society and improve human well-being.

来改善社会和人类的生存。

Thank you.

谢谢。