Kenneth Cukier: Big data is better data

翻译人员: Simon Cai 校对人员: Amy Yang

America's favorite pie is?

美国人最爱的馅饼是什么？

Audience: Apple. Kenneth Cukier: Apple. Of course it is.

观众：苹果派 Kenneth Cukier：苹果派 毋庸置疑

How do we know it?

我们是怎么知道的？

Because of data.

因为数据

You look at supermarket sales.

当你观察超市的销售数据

You look at supermarket sales of 30-centimeter pies

会发现超市销售的30厘米冷冻馅饼中

that are frozen, and apple wins, no contest.

苹果派胜出, 毫无悬念

The majority of the sales are apple.

绝大多数的销售份额就是来自苹果派

But then supermarkets started selling

但是之后超市开始销售

smaller, 11-centimeter pies,

比较小的11厘米的馅饼

and suddenly, apple fell to fourth or fifth place.

突然间苹果派的销量下降到了第4或第5名

Why? What happened?

为什么？怎么了？

Okay, think about it.

好, 想象一下

When you buy a 30-centimeter pie,

当你准备买一个30厘米的馅饼时

the whole family has to agree,

全家都不得不同意(选择苹果派馅饼)

and apple is everyone's second favorite.

虽然苹果派只是每个人的次选项

(Laughter)

（笑声）

But when you buy an individual 11-centimeter pie,

但当你给自己选一个11厘米馅饼时

you can buy the one that you want.

你可以买你最爱吃的口味

You can get your first choice.

你会选你的首选项

You have more data.

你有了更多数据

You can see something

你可以知道些事情

that you couldn't see

这些事情在你只有少量数据时

when you only had smaller amounts of it.

你是无法知道的

Now, the point here is that more data

这里, 关键的是更多的数据

doesn't just let us see more,

不单单让我们知道更多

more of the same thing we were looking at.

知道更多我们正在关注的同样事物

More data allows us to see new.

更多的数据使我们能了解新的事情

It allows us to see better.

让我们更好地了解

It allows us to see different.

让我们有不同的视角

In this case, it allows us to see

在这个例子里 更多的数据让我们知道

what America's favorite pie is:

美国人最喜欢的馅饼

not apple.

不是苹果派

Now, you probably all have heard the term big data.

你或许听说过大数据这个词

In fact, you're probably sick of hearing the term

事实上, 你可能对这个词

big data.

已经心生厌恶

It is true that there is a lot of hype around the term,

确实, 大数据受到了空前的宣传炒作

and that is very unfortunate,

这很不应该

because big data is an extremely important tool

因为大数据是一个非常重要的工具

by which society is going to advance.

社会将由此而不断进步

In the past, we used to look at small data

过去我们习惯于处理小数据

and think about what it would mean

思考这些小数据的意义

to try to understand the world,

并以此来了解世界

and now we have a lot more of it,

现在我们有很多很多的数据

more than we ever could before.

数据量前所未有的巨大

What we find is that when we have

当我们掌握海量数据时

a large body of data, we can fundamentally do things

我们可以做一些事

that we couldn't do when we only had smaller amounts.

一些在只有较少数据时不可能办到的事

Big data is important, and big data is new,

大数据很重要, 它也是一个新兴事物

and when you think about it,

想象一下

the only way this planet is going to deal

能够帮助我们应对

with its global challenges —

世界性难题

to feed people, supply them with medical care,

像食物短缺 医疗短缺

supply them with energy, electricity,

能源短缺 电力短缺

and to make sure they're not burnt to a crisp

还有确保人类家园

because of global warming —

不会因为全球变暖而生灵涂炭

is because of the effective use of data.

的唯一办法是有效利用大数据

So what is new about big data? What is the big deal?

那么大数据新在何处, 重在何处呢？

Well, to answer that question, let's think about

为了回答这个问题, 让我们看一下

what information looked like,

信息看上去是什么样的

physically looked like in the past.

信息在以前是什么样的

In 1908, on the island of Crete,

1908年在克里特岛上 (注：位于地中海 为希腊第一大岛)

archaeologists discovered a clay disc.

考古学家发现了一个粘土做的盘子

They dated it from 2000 B.C., so it's 4,000 years old.

这是个公元前2000年的盘子 距今约有4000年的历史

Now, there's inscriptions on this disc,

盘子上有铭文

but we actually don't know what it means.

但是我们不知道它们是什么意思

It's a complete mystery, but the point is that

这完全是个谜团

this is what information used to look like

但这就是4000年前

4,000 years ago.

信息的样子

This is how society stored

这就是当时社会

and transmitted information.

存储和传递信息的方式

Now, society hasn't advanced all that much.

现代社会也没有什么很大的进步

We still store information on discs,

我们还是把数据存储在盘中 （注：指磁盘)

but now we can store a lot more information,

但我们可以存储更多的信息

more than ever before.

远远超过以前的信息容量

Searching it is easier. Copying it easier.

这些信息搜索和复制起来更简单

Sharing it is easier. Processing it is easier.

分享和处理起来也更便捷

And what we can do is we can reuse this information

我们也可以重新利用这些数据

for uses that we never even imagined

一些我们当初收集的时候

when we first collected the data.

从来没有料想过的用途

In this respect, the data has gone

从这个方面来说

数据已经从储存状态到了流动状态

from a stock to a flow,

from something that is stationary and static

从静态的统计性的数据

to something that is fluid and dynamic.

变成动态的数据流

There is, if you will, a liquidity to information.

这就是信息的流动性

The disc that was discovered off of Crete

克里特岛发现的粘土盘

that's 4,000 years old, is heavy,

有4000年的历史, 非常笨重

it doesn't store a lot of information,

但它不能记录太多的信息

and that information is unchangeable.

并且它所记录的信息是不能更改的

By contrast, all of the files

与此相反

爱德华·斯诺登从美国国家安全局

that Edward Snowden took

from the National Security Agency in the United States

所获得的文件

fits on a memory stick

可以放在一个

the size of a fingernail,

仅有指甲大小的存储盘里

and it can be shared at the speed of light.

并且可以以光速进行数据共享

More data. More.

更多数据 更多

Now, one reason why we have so much data in the world today

今天我们有这么多数据的一个原因是

is we are collecting things

我们一直在收集信息

that we've always collected information on,

就像我们一直在做的一样

but another reason why is we're taking things

另一个原因是我们记录了

that have always been informational

许多蕴含丰富信息的事物

but have never been rendered into a data format

但是从没把信息转换成数据形式

and we are putting it into data.

现在我们正在把信息转变成数据

Think, for example, the question of location.

举个例子, 定位问题

Take, for example, Martin Luther.

比如说马丁·路德

If we wanted to know in the 1500s

在16世纪 如果我们想知道

where Martin Luther was,

马丁·路德在哪里

we would have to follow him at all times,

我们必须一直跟着他

maybe with a feathery quill and an inkwell,

或许用羽毛笔和墨水

and record it,

把这些情况记录下来

but now think about what it looks like today.

那现今是什么样的情形呢？

You know that somewhere,

在某些地方

probably in a telecommunications carrier's database,

可能在电信运营商的数据库里

there is a spreadsheet or at least a database entry

有个电子数据表或者至少一个数据目录

that records your information

记录着所有关于你

of where you've been at all times.

任何时候在什么地点的信息

If you have a cell phone,

如果你有个手机

and that cell phone has GPS, but even if it doesn't have GPS,

这个手机有GPS, 或者即使没有GPS

it can record your information.

它还是可以记录你的信息

In this respect, location has been datafied.

从这方面来说, 位置信息被数据化了

Now think, for example, of the issue of posture,

再举个例子, 关于姿势

the way that you are all sitting right now,

你们现在坐着的姿势

the way that you sit,

你坐着的姿势

the way that you sit, the way that you sit.

你坐着的姿势 你坐着的姿势

It's all different, and it's a function of your leg length

这些都不一样 这是一个关于腿长

and your back and the contours of your back,

你的背部和背部轮廓的函数

and if I were to put sensors, maybe 100 sensors

如果我现在放一些传感器 或许100个

into all of your chairs right now,

在你的椅子里

I could create an index that's fairly unique to you,

我可以算出你的独一无二的参数

sort of like a fingerprint, but it's not your finger.

就像你的指纹 但不是针对你的手指

So what could we do with this?

那我们能用它来干什么呢？

Researchers in Tokyo are using it

东京的研究者把它

as a potential anti-theft device in cars.

运用在一个汽车防盗设施的雏形上

The idea is that the carjacker sits behind the wheel,

它的设想是盗贼坐在驾驶座上

tries to stream off, but the car recognizes

企图把车开走 但是汽车识别出

that a non-approved driver is behind the wheel,

驾驶座上的是个未授权驾驶人

and maybe the engine just stops, unless you

那汽车可能就会熄火

type in a password into the dashboard

除非你在仪表盘上输入密码

to say, "Hey, I have authorization to drive." Great.

来表明“我已获得授权”

What if every single car in Europe

如果欧洲的每辆汽车

had this technology in it?

都装备了这项技术会是怎样的情形？

What could we do then?

我们还能做些什么呢？

Maybe, if we aggregated the data,

或许如果我们整合数据

maybe we could identify telltale signs

我们可以识别示警信号

that best predict that a car accident

对于在下一个五秒钟内

is going to take place in the next five seconds.

可能发生的意外做出最佳预判

And then what we will have datafied

我们也可以进行数据化的是

is driver fatigue,

司机的疲劳度

and the service would be when the car senses

当汽车侦测到司机的坐姿

that the person slumps into that position,

倒成某一特定姿势时

automatically knows, hey, set an internal alarm

这个设备感知到并发出车内警告

that would vibrate the steering wheel, honk inside

可能是震动方向盘或语音提示

to say, "Hey, wake up,

“嗨，醒醒

pay more attention to the road."

集中精神在路况上”

These are the sorts of things we can do

这就是生活的更多方面数据化后

when we datafy more aspects of our lives.

我们能做的事情

So what is the value of big data?

那么大数据的价值在哪里？

Well, think about it.

好 思考一下

You have more information.

你有了更多地信息

You can do things that you couldn't do before.

你可以做你以前不能做的事

One of the most impressive areas

在运用这个概念的领域里

where this concept is taking place

让人印象最为最深刻的

is in the area of machine learning.

是机器学习

Machine learning is a branch of artificial intelligence,

机器学习是人工智能的一个分支

which itself is a branch of computer science.

人工智能又是计算机科学的一个分支

The general idea is that instead of

它的基本理念是

instructing a computer what do do,

把关于某个问题的一堆数据扔给电脑

we are going to simply throw data at the problem

让电脑自己找出解决方案

and tell the computer to figure it out for itself.

而不是教电脑应该做什么

And it will help you understand it

通过机器学习的原型

by seeing its origins.

可以帮助你来理解这个理念

In the 1950s, a computer scientist

20世纪50年代IBM的计算机科学家

at IBM named Arthur Samuel liked to play checkers,

亚瑟·塞缪尔想玩跳棋

so he wrote a computer program

所以他写了个程序

so he could play against the computer.

这样他就可以和电脑来玩

He played. He won.

开始他下一盘 赢一盘

He played. He won.

下一盘 赢一盘

He played. He won,

下一盘 赢一盘

because the computer only knew

因为电脑只知道

what a legal move was.

规则允许怎样走

Arthur Samuel knew something else.

亚瑟·塞缪尔还知道其他东西

Arthur Samuel knew strategy.

他知道下棋的策略

So he wrote a small sub-program alongside it

所以他又写了一个附加程序

operating in the background, and all it did

这个程序在后台运行

was score the probability

它的功能只是计算概率

that a given board configuration would likely lead

在一个给定的棋局里

to a winning board versus a losing board

每走一步后

after every move.

会获胜或者失败的概率

He plays the computer. He wins.

再和电脑下棋 还是下一盘 赢一盘

He plays the computer. He wins.

下一盘 赢一盘

He plays the computer. He wins.

下一盘 赢一盘

And then Arthur Samuel leaves the computer

后来亚瑟让电脑

to play itself.

自己和自己下棋

It plays itself. It collects more data.

电脑自己玩的时候收集了更多的数据

It collects more data. It increases the accuracy of its prediction.

收集的数据越多, 预测的准确率就越高

And then Arthur Samuel goes back to the computer

然后亚瑟又继续和电脑下棋

and he plays it, and he loses,

这次他下一盘 输一盘

and he plays it, and he loses,

下一盘 输一盘

and he plays it, and he loses,

下一盘 输一盘

and Arthur Samuel has created a machine

亚瑟创造了一个机器

that surpasses his ability in a task that he taught it.

它的能力超越了亚瑟开始时所教给它的

And this idea of machine learning

机器学习的理念

is going everywhere.

现在已经随处可见

How do you think we have self-driving cars?

你们觉得无人驾驶汽车(关键的技术)是什么？

Are we any better off as a society

是不是把所有交通规则输入软件

enshrining all the rules of the road into software?

就万事大吉了？不是

No. Memory is cheaper. No.

内存很便宜？不是

Algorithms are faster. No. Processors are better. No.

算法更快了 不是 处理器更强大了 不是

All of those things matter, but that's not why.

这些都有影响, 但不是真正的原因

It's because we changed the nature of the problem.

真正的原因是我们改变了问题的本质

We changed the nature of the problem from one

我们把问题的本质从

in which we tried to overtly and explicitly

试图明确无误地

explain to the computer how to drive

教会电脑怎样驾驶

to one in which we say,

变成我们对电脑说

"Here's a lot of data around the vehicle.

“这里有许多关于汽车的数据

You figure it out.

你自己搞定它

You figure it out that that is a traffic light,

你知道那是交通信号灯

that that traffic light is red and not green,

那是红灯不是绿灯

that that means that you need to stop

遇到红灯你必须停下来

and not go forward."

不能往前走”

Machine learning is at the basis

机器学习是许多

of many of the things that we do online:

网上在线应用的基础

search engines,

搜索引擎

Amazon's personalization algorithm,

亚马逊的个性化算法

computer translation,

电脑智能翻译

voice recognition systems.

语音识别系统

Researchers recently have looked at

研究者最近在研究

the question of biopsies,

关于活组织检查的问题

cancerous biopsies,

关于肿瘤活组织检查

and they've asked the computer to identify

他们让电脑

by looking at the data and survival rates

通过 (历史) 数据和存活率

to determine whether cells are actually

来判断这些细胞

cancerous or not,

是否是癌症细胞

and sure enough, when you throw the data at it,

果不其然 当你把数据交给电脑

through a machine-learning algorithm,

电脑通过自主学习

the machine was able to identify

可以寻找出

the 12 telltale signs that best predict

12个最佳的鉴别特征用来预测

that this biopsy of the breast cancer cells

乳腺癌细胞的活检切片

are indeed cancerous.

确实是癌症细胞

The problem: The medical literature

问题是医学文献

only knew nine of them.

只知道其中的九个鉴别特征

Three of the traits were ones

其他三个

that people didn't need to look for,

人们不会去寻找

but that the machine spotted.

但是电脑把它们找了出来

Now, there are dark sides to big data as well.

大数据也有黑暗的一面

It will improve our lives, but there are problems

它可以改善我们的生活

that we need to be conscious of,

但也会带来一些我们需要注意的问题

and the first one is the idea

首先就是

that we may be punished for predictions,

我们可能因为预测的结果而受到惩罚

that the police may use big data for their purposes,

警察可能会用大数据来实现目标

a little bit like "Minority Report."

有点像“少数派报告”

Now, it's a term called predictive policing,

现在有个词叫做预见性监管

or algorithmic criminology,

或者叫算法犯罪学

and the idea is that if we take a lot of data,

这个想法是如果我们掌握了大量数据

for example where past crimes have been,

比如以往犯罪发生的地点

we know where to send the patrols.

我们可以就知道把警力派到哪里

That makes sense, but the problem, of course,

这很合理 但问题是

is that it's not simply going to stop on location data,

数据分析不会仅限于地点数据

it's going to go down to the level of the individual.

它会进一步深入到个人层面

Why don't we use data about the person's

为什么我们不去分析

high school transcript?

某人的中学成绩单

Maybe we should use the fact that

或者我们可以了解

they're unemployed or not, their credit score,

他们的就职情况、信用记录

their web-surfing behavior,

他们的上网行为

whether they're up late at night.

他们是否熬夜

Their Fitbit, when it's able to identify biochemistries,

当可以通过健康腕带读取生化数据时

will show that they have aggressive thoughts.

就可以知道他们是否有激进的想法

We may have algorithms that are likely to predict

我们可以用算法来预测

what we are about to do,

我们将要做什么

and we may be held accountable

可能有些事情还没做

before we've actually acted.

我们就要承担责任

Privacy was the central challenge

个人隐私在小数据时代

in a small data era.

是主要挑战

In the big data age,

在大数据时代

the challenge will be safeguarding free will,

这个挑战将会成为保卫自由意愿

moral choice, human volition,

道德选择 、人类意志

human agency.

人类的能动性

There is another problem:

还有另一个问题

Big data is going to steal our jobs.

大数据会偷走我们的工作

Big data and algorithms are going to challenge

在21世纪

white collar, professional knowledge work

大数据和算法会威胁到

in the 21st century

白领和需要专业知识的工作

in the same way that factory automation

就像在20世纪工厂自动化

and the assembly line

和装配生产线的应用

challenged blue collar labor in the 20th century.

威胁到了蓝领们的工作岗位

Think about a lab technician

想象一下一个研究室技术员

who is looking through a microscope

他的工作就是通过一个显微镜

at a cancer biopsy

观察一个癌症活检组织

and determining whether it's cancerous or not.

来判定它是不是癌症的

The person went to university.

这个人上大学

The person buys property.

买房子

He or she votes.

他/她投票选举

He or she is a stakeholder in society.

他/她是这个社会的一份子

And that person's job,

然后这个人的工作

as well as an entire fleet

还有其他

of professionals like that person,

像他一样的专业人员

is going to find that their jobs are radically changed

将会发现他们的工作被彻底改变了

or actually completely eliminated.

或者彻底废除了

Now, we like to think

我们一直以为

that technology creates jobs over a period of time

在短时或者暂时的就业调整期后

after a short, temporary period of dislocation,

一段时间内科技会创造就业机会

and that is true for the frame of reference

这对于我们所处的参考系

with which we all live, the Industrial Revolution,

工业革命来说就是这样

because that's precisely what happened.

因为在工业革命时期事情就是这样的

But we forget something in that analysis:

但是我们忘记了一件事情

There are some categories of jobs

有些类型的职业

that simply get eliminated and never come back.

已经彻底消失了并且再也不会回来

The Industrial Revolution wasn't very good

如果你是一匹马

if you were a horse.

工业革命不是一件好事

So we're going to need to be careful

所以我们必须非常小心

and take big data and adjust it for our needs,

根据我们的需求和整个人类的需求

our very human needs.

来利用和适应大数据

We have to be the master of this technology,

我们必须是技术的主人

not its servant.

而不是技术的仆人

We are just at the outset of the big data era,

我们正在步入大数据时代

and honestly, we are not very good

老实说, 我们并不能很好地

at handling all the data that we can now collect.

处理所有我们现在能够收集到的数据

It's not just a problem for the National Security Agency.

这不仅仅是国家安全局的问题

Businesses collect lots of data, and they misuse it too,

许多企业也搜集并不恰当地使用数据

and we need to get better at this, and this will take time.

我们需要时间来纠正这个问题

It's a little bit like the challenge that was faced

这有点像原始人类面对火时

by primitive man and fire.

所面临的挑战

This is a tool, but this is a tool that,

火是一种工具

unless we're careful, will burn us.

但是如果使用不当就会引火烧身

Big data is going to transform how we live,

大数据即将改变我们的生活方式

how we work and how we think.

我们的工作方式和思考方式

It is going to help us manage our careers

它可以帮助我们管理事业

and lead lives of satisfaction and hope

帮助我们过想要的满足、充满希望

and happiness and health,

幸福和健康的生活

but in the past, we've often looked at information technology

但是在过去, 对于信息技术(IT)

and our eyes have only seen the T,

我们经常只看到了T

the technology, the hardware,

就是技术、硬件

because that's what was physical.

因为这是切实可见的东西

We now need to recast our gaze at the I,

现在我们需要把目光放在 I 上

the information,

信息

which is less apparent,

它不是那么切实可见

but in some ways a lot more important.

但某种程度上却更加重要

Humanity can finally learn from the information

在人类永无止境的探索过程中

that it can collect,

我们可以从我们能收集的信息中

as part of our timeless quest

来了解这个世界

to understand the world and our place in it,

以及人类在这个世界中所处的地位

and that's why big data is a big deal.

这就是为什么大数据非常重要

(Applause)

（掌声）