TED2011
Deb Roy: The birth of a word
戴·罗伊: 单词的诞生
Filmed:
Readability: 4.2
2,809,941 views
麻省理工大学的研究员戴·罗伊希望了解他的初生儿子是怎样学习语言的，因此他在家里四处装置了摄像机捕捉儿子成长的几乎每一刻，然后将9万小时的家庭录像进行分析，观察 “gaaaa” 渐渐转变成 “水”的过程。 这一惊人的资料丰富的研究对帮助我们了解人类是如果学习起到了深远的作用
Deb Roy - Cognitive scientist
Deb Roy studies how children learn language, and designs machines that learn to communicate in human-like ways. On sabbatical from MIT Media Lab, he's working with the AI company Bluefin Labs. Full bio
Deb Roy studies how children learn language, and designs machines that learn to communicate in human-like ways. On sabbatical from MIT Media Lab, he's working with the AI company Bluefin Labs. Full bio
Double-click the English transcript below to play the video.
00:15
Imagine if you could record your life --
0
0
4000
想象一下如果你能记录你的生活--
00:19
everything you said, everything you did,
1
4000
3000
你说的一切，做的一切
00:22
available in a perfect memory store at your fingertips,
2
7000
3000
就存储在一个完美的你触手可及的记忆库
00:25
so you could go back
3
10000
2000
你可以回到过去
00:27
and find memorable moments and relive them,
4
12000
3000
找寻那难忘一刻回顾这一切
00:30
or sift through traces of time
5
15000
3000
或者追寻时间的轨迹
00:33
and discover patterns in your own life
6
18000
2000
发现在属于你自己的生活模式
00:35
that previously had gone undiscovered.
7
20000
3000
那种以前没有发现的规律
00:38
Well that's exactly the journey
8
23000
2000
而那就是我们全家
00:40
that my family began
9
25000
2000
5年半前开始的
00:42
five and a half years ago.
10
27000
2000
家庭旅程
00:44
This is my wife and collaborator, Rupal.
11
29000
3000
这是我妻子和合作者， 鲁泊尔
00:47
And on this day, at this moment,
12
32000
2000
在这一天，从这一刻
00:49
we walked into the house with our first child,
13
34000
2000
我们带着我们第一个孩子走进了这个家
00:51
our beautiful baby boy.
14
36000
2000
我们美丽的儿子
00:53
And we walked into a house
15
38000
3000
我们走进了一个
00:56
with a very special home video recording system.
16
41000
4000
安装了特殊的家庭摄像系统的家
01:07
(Video) Man: Okay.
17
52000
2000
(录像) 男人：好
01:10
Deb Roy: This moment
18
55000
1000
戴·罗伊: 这一刻
01:11
and thousands of other moments special for us
19
56000
3000
和其他千万的我们的特殊时刻
01:14
were captured in our home
20
59000
2000
在我们家中被捕捉下来
01:16
because in every room in the house,
21
61000
2000
因为这个房子的每个屋子
01:18
if you looked up, you'd see a camera and a microphone,
22
63000
3000
如果你仰头看，你都可以看见一个摄像机和话筒
01:21
and if you looked down,
23
66000
2000
而你望下看
01:23
you'd get this bird's-eye view of the room.
24
68000
2000
你可以俯视整个房间
01:25
Here's our living room,
25
70000
3000
这是我们的客厅
01:28
the baby bedroom,
26
73000
3000
这是婴儿的房间
01:31
kitchen, dining room
27
76000
2000
厨房，餐厅
01:33
and the rest of the house.
28
78000
2000
这是其余的地方
01:35
And all of these fed into a disc array
29
80000
3000
这些都被装进了一排
01:38
that was designed for a continuous capture.
30
83000
3000
为持续拍摄设计的光盘中
01:41
So here we are flying through a day in our home
31
86000
3000
这里我们飞快地经历一遍我们家庭的一天
01:44
as we move from sunlit morning
32
89000
3000
我们从太阳初升的早晨
01:47
through incandescent evening
33
92000
2000
到亮起电灯的夜晚
01:49
and, finally, lights out for the day.
34
94000
3000
最后， 熄灯就寝
01:53
Over the course of three years,
35
98000
3000
历经3年的时间
01:56
we recorded eight to 10 hours a day,
36
101000
2000
我们每天记录8到10个小时
01:58
amassing roughly a quarter-million hours
37
103000
3000
积累了大约25万小时
02:01
of multi-track audio and video.
38
106000
3000
的多轨音频和视频资料
02:04
So you're looking at a piece of what is by far
39
109000
2000
所以你现在看到的是有史以来
02:06
the largest home video collection ever made.
40
111000
2000
最大的家庭录相集
02:08
(Laughter)
41
113000
3000
(笑声)
02:11
And what this data represents
42
116000
2000
从个人的角度而言，
02:13
for our family at a personal level,
43
118000
4000
这些代表了我们家庭的资料
02:17
the impact has already been immense,
44
122000
2000
已经产生了巨大的影响
02:19
and we're still learning its value.
45
124000
3000
我们还在继续学习其中的价值
02:22
Countless moments
46
127000
2000
无数的时刻
02:24
of unsolicited natural moments, not posed moments,
47
129000
3000
无预兆的，不造作的自然时刻
02:27
are captured there,
48
132000
2000
都记录在这里
02:29
and we're starting to learn how to discover them and find them.
49
134000
3000
我们正开始学习怎样发现和寻找它们
02:32
But there's also a scientific reason that drove this project,
50
137000
3000
而促使这个项目还有一个科学的原因
02:35
which was to use this natural longitudinal data
51
140000
4000
便是用这些纵向记录的数据
02:39
to understand the process
52
144000
2000
去了解一个
02:41
of how a child learns language --
53
146000
2000
孩子是怎样学习语言的--
02:43
that child being my son.
54
148000
2000
这个孩子是我的儿子
02:45
And so with many privacy provisions put in place
55
150000
4000
所以在设置了隐私保护的条件下
02:49
to protect everyone who was recorded in the data,
56
154000
3000
每个被记录到的人物都得到保护
02:52
we made elements of the data available
57
157000
3000
我们对我们信任的麻省理工研究团队
02:55
to my trusted research team at MIT
58
160000
3000
公开了部分数据
02:58
so we could start teasing apart patterns
59
163000
3000
因此我们可以从这个巨大的
03:01
in this massive data set,
60
166000
3000
数据资料中排除出一些多余的模式
03:04
trying to understand the influence of social environments
61
169000
3000
以此来试图理解社会环境
03:07
on language acquisition.
62
172000
2000
对语言形成的影响
03:09
So we're looking here
63
174000
2000
所以我们在这里看到
03:11
at one of the first things we started to do.
64
176000
2000
我们所做的第一件事情
03:13
This is my wife and I cooking breakfast in the kitchen,
65
178000
4000
这是我的妻子和我在厨房做早餐
03:17
and as we move through space and through time,
66
182000
3000
随着时间的流逝地点的变化
03:20
a very everyday pattern of life in the kitchen.
67
185000
3000
这是厨房里日常生活的轨迹
03:23
In order to convert
68
188000
2000
为了转换
03:25
this opaque, 90,000 hours of video
69
190000
3000
这个9万小时的录相
03:28
into something that we could start to see,
70
193000
2000
将它变成我们能识辨的东西
03:30
we use motion analysis to pull out,
71
195000
2000
我们用行动分析来抽取
03:32
as we move through space and through time,
72
197000
2000
我们在时空的移动
03:34
what we call space-time worms.
73
199000
3000
我们称之为 时空虫
03:37
And this has become part of our toolkit
74
202000
3000
这个成为了我们工具的一部分
03:40
for being able to look and see
75
205000
3000
用来观察和辨识
03:43
where the activities are in the data,
76
208000
2000
数据中的各种活动
03:45
and with it, trace the pattern of, in particular,
77
210000
3000
再利用这个办法，去追踪模型，特别是
03:48
where my son moved throughout the home,
78
213000
2000
我儿子在家去过哪些地方
03:50
so that we could focus our transcription efforts,
79
215000
3000
使得我们能够聚焦解读
03:53
all of the speech environment around my son --
80
218000
3000
我儿子学习语言的语境
03:56
all of the words that he heard from myself, my wife, our nanny,
81
221000
3000
他从我，我妻子和保姆那里听到的所有词汇
03:59
and over time, the words he began to produce.
82
224000
3000
渐渐的，他开始使用的词汇
04:02
So with that technology and that data
83
227000
3000
因此通过技术和数据
04:05
and the ability to, with machine assistance,
84
230000
2000
在机器的协助下
04:07
transcribe speech,
85
232000
2000
录制下对话
04:09
we've now transcribed
86
234000
2000
我们现在已经完成了
04:11
well over seven million words of our home transcripts.
87
236000
3000
超过7万字的家庭言谈的记录
04:14
And with that, let me take you now
88
239000
2000
现在，让我带你们
04:16
for a first tour into the data.
89
241000
3000
进入这些数据的第一个旅行
04:19
So you've all, I'm sure,
90
244000
2000
我相信，你们大家都
04:21
seen time-lapse videos
91
246000
2000
看过时间推移的影片
04:23
where a flower will blossom as you accelerate time.
92
248000
3000
加快时间的推移你可以看见花朵盛开
04:26
I'd like you to now experience
93
251000
2000
现在我让你们看看
04:28
the blossoming of a speech form.
94
253000
2000
语言的花朵是怎样绽放的
04:30
My son, soon after his first birthday,
95
255000
2000
我的儿子，在他的第一个生日后
04:32
would say "gaga" to mean water.
96
257000
3000
会说“gaga“来指水
04:35
And over the course of the next half-year,
97
260000
3000
在这之后的半年里
04:38
he slowly learned to approximate
98
263000
2000
他渐渐地学会了
04:40
the proper adult form, "water."
99
265000
3000
成年人说的正确的“水”
04:43
So we're going to cruise through half a year
100
268000
2000
我们现在来用40秒时间
04:45
in about 40 seconds.
101
270000
2000
快速浏览这半年
04:47
No video here,
102
272000
2000
没有影象
04:49
so you can focus on the sound, the acoustics,
103
274000
3000
所以你们可以专注听声音，声学上的
04:52
of a new kind of trajectory:
104
277000
2000
这种新的轨迹变化
04:54
gaga to water.
105
279000
2000
从“Gaga"到"Water"
04:56
(Audio) Baby: Gagagagagaga
106
281000
12000
(声音)婴儿：Gagagagagaga
05:08
Gaga gaga gaga
107
293000
4000
Gaga gaga gaga
05:12
guga guga guga
108
297000
5000
guga guga guga
05:17
wada gaga gaga guga gaga
109
302000
5000
wada gaga gaga guga gaga
05:22
wader guga guga
110
307000
4000
wader guga guga
05:26
water water water
111
311000
3000
water water water
05:29
water water water
112
314000
6000
water water water
05:35
water water
113
320000
4000
water water
05:39
water.
114
324000
2000
water
05:41
DR: He sure nailed it, didn't he.
115
326000
2000
戴·罗伊: 他学会了啊，不是吗？
05:43
(Applause)
116
328000
7000
(掌声)
05:50
So he didn't just learn water.
117
335000
2000
而他并不只是学会了水
05:52
Over the course of the 24 months,
118
337000
2000
在24个月里
05:54
the first two years that we really focused on,
119
339000
3000
在最初的2年里，这才是我真正关注的
05:57
this is a map of every word he learned in chronological order.
120
342000
4000
这里有一张图按照时序列出了他所学到的词汇
06:01
And because we have full transcripts,
121
346000
3000
因为我们有全部的记录
06:04
we've identified each of the 503 words
122
349000
2000
我们为他到两岁前学会的503个单词
06:06
that he learned to produce by his second birthday.
123
351000
2000
都做了辨认和分析
06:08
He was an early talker.
124
353000
2000
他算是说话早的
06:10
And so we started to analyze why.
125
355000
3000
所以我们开始分析其原因
06:13
Why were certain words born before others?
126
358000
3000
为什么有些词他学得早
06:16
This is one of the first results
127
361000
2000
这是其中的一个研究结果
06:18
that came out of our study a little over a year ago
128
363000
2000
是一年多前出来的
06:20
that really surprised us.
129
365000
2000
让我们很吃惊
06:22
The way to interpret this apparently simple graph
130
367000
3000
解读这张看似简单的图表的方式
06:25
is, on the vertical is an indication
131
370000
2000
是横坐标表示
06:27
of how complex caregiver utterances are
132
372000
3000
照顾者的话语复杂程度
06:30
based on the length of utterances.
133
375000
2000
基于话语的长度
06:32
And the [horizontal] axis is time.
134
377000
3000
纵坐标代表了时间(演讲者口误)
06:35
And all of the data,
135
380000
2000
所有的数据
06:37
we aligned based on the following idea:
136
382000
3000
我们都用下述的方法排列：
06:40
Every time my son would learn a word,
137
385000
3000
每次我们发现儿子学了一个新的词
06:43
we would trace back and look at all of the language he heard
138
388000
3000
我们就会回溯他听过的这个词的
06:46
that contained that word.
139
391000
2000
所有的语言记录
06:48
And we would plot the relative length of the utterances.
140
393000
4000
然后我们绘制这些语言的长度
06:52
And what we found was this curious phenomena,
141
397000
3000
我们发现了一个奇特的现象
06:55
that caregiver speech would systematically dip to a minimum,
142
400000
3000
照顾者的讲话会系统地将语言简化
06:58
making language as simple as possible,
143
403000
3000
简化到最简单的程度
07:01
and then slowly ascend back up in complexity.
144
406000
3000
然后渐渐地回升到更复杂的句子
07:04
And the amazing thing was
145
409000
2000
而惊奇的事是
07:06
that bounce, that dip,
146
411000
2000
这种回升和下降
07:08
lined up almost precisely
147
413000
2000
正好精确的
07:10
with when each word was born --
148
415000
2000
吻合了每个词的诞生过程--
07:12
word after word, systematically.
149
417000
2000
一个词接一个词，很有系统规律
07:14
So it appears that all three primary caregivers --
150
419000
2000
似乎三个主要的照顾他的人
07:16
myself, my wife and our nanny --
151
421000
3000
我，我妻子，和我们的保姆--
07:19
were systematically and, I would think, subconsciously
152
424000
3000
都是有系统的，我想，也是下意识的
07:22
restructuring our language
153
427000
2000
重新构建我们的用语
07:24
to meet him at the birth of a word
154
429000
3000
去迎合他的新的词汇的诞生
07:27
and bring him gently into more complex language.
155
432000
4000
带他渐渐学习更为复杂的语言
07:31
And the implications of this -- there are many,
156
436000
2000
这其中蕴含的--有很多意义
07:33
but one I just want to point out,
157
438000
2000
但是我想指出的其中的一个
07:35
is that there must be amazing feedback loops.
158
440000
3000
就是这个过程中必定包涵了一个惊人的反馈循环
07:38
Of course, my son is learning
159
443000
2000
当然，我的儿子是
07:40
from his linguistic environment,
160
445000
2000
在他的语言环境中学习
07:42
but the environment is learning from him.
161
447000
3000
但是那个环境也在向他学习
07:45
That environment, people, are in these tight feedback loops
162
450000
3000
环境，人，都在这个紧密的反馈循环中
07:48
and creating a kind of scaffolding
163
453000
2000
并建立了一种类似脚手架的互相支撑关系
07:50
that has not been noticed until now.
164
455000
3000
这是之前没有被注意到的
07:54
But that's looking at the speech context.
165
459000
2000
这是关注讲话的语境来看
07:56
What about the visual context?
166
461000
2000
若是从视觉环境来看呢？
07:58
We're not looking at --
167
463000
2000
我们现在看到的是
08:00
think of this as a dollhouse cutaway of our house.
168
465000
2000
想象这是用我们家做样板做的洋娃娃屋
08:02
We've taken those circular fish-eye lens cameras,
169
467000
3000
我们使用环状鱼眼睛摄像机
08:05
and we've done some optical correction,
170
470000
2000
我们还做了些光学修正
08:07
and then we can bring it into three-dimensional life.
171
472000
4000
然后我们就可以把它做成三维录像
08:11
So welcome to my home.
172
476000
2000
欢迎到我家来
08:13
This is a moment,
173
478000
2000
这是其中的一刻
08:15
one moment captured across multiple cameras.
174
480000
3000
通过几个录相机拍下的同一时刻
08:18
The reason we did this is to create the ultimate memory machine,
175
483000
3000
我们这样做是为了创造出终极的记忆机器
08:21
where you can go back and interactively fly around
176
486000
3000
你可以用互动的方式前后快速搜寻
08:24
and then breathe video-life into this system.
177
489000
3000
然后用这系统体验录像生活
08:27
What I'm going to do
178
492000
2000
我要做的是
08:29
is give you an accelerated view of 30 minutes,
179
494000
3000
是给你们看一段压缩了30分钟的速放录像
08:32
again, of just life in the living room.
180
497000
2000
这次也是在客厅
08:34
That's me and my son on the floor.
181
499000
3000
这是我和我儿子在地上
08:37
And there's video analytics
182
502000
2000
这是影片分析
08:39
that are tracking our movements.
183
504000
2000
跟踪我们的移动
08:41
My son is leaving red ink. I am leaving green ink.
184
506000
3000
我儿子的留下了红色的轨迹，我的是绿色的
08:44
We're now on the couch,
185
509000
2000
我们在沙发上
08:46
looking out through the window at cars passing by.
186
511000
3000
看着窗外汽车开过
08:49
And finally, my son playing in a walking toy by himself.
187
514000
3000
最后，我儿子自己玩他的学步玩具
08:52
Now we freeze the action, 30 minutes,
188
517000
3000
现在定格，30分钟
08:55
we turn time into the vertical axis,
189
520000
2000
我们将时间放到垂直轴上
08:57
and we open up for a view
190
522000
2000
然后我们打开
08:59
of these interaction traces we've just left behind.
191
524000
3000
刚才留下的互动的轨迹
09:02
And we see these amazing structures --
192
527000
3000
我们看见令人惊讶的结构
09:05
these little knots of two colors of thread
193
530000
3000
这是两种颜色的小结点
09:08
we call "social hot spots."
194
533000
2000
我们把它称为社交热点
09:10
The spiral thread
195
535000
2000
那些螺旋线
09:12
we call a "solo hot spot."
196
537000
2000
我们称为单一热点
09:14
And we think that these affect the way language is learned.
197
539000
3000
我们觉得这个影响语言学习
09:17
What we'd like to do
198
542000
2000
我们要做的是
09:19
is start understanding
199
544000
2000
是开始去了解
09:21
the interaction between these patterns
200
546000
2000
这些模式与我儿子接触的
09:23
and the language that my son is exposed to
201
548000
2000
语言间的关系
09:25
to see if we can predict
202
550000
2000
看我们是否能预测
09:27
how the structure of when words are heard
203
552000
2000
什么时候听到怎样的单词结构
09:29
affects when they're learned --
204
554000
2000
会影响到什么时候学会字词
09:31
so in other words, the relationship
205
556000
2000
换句话说，就是
09:33
between words and what they're about in the world.
206
558000
4000
词汇和他们所表示的世界的关系
09:37
So here's how we're approaching this.
207
562000
2000
这是我们的解读方法
09:39
In this video,
208
564000
2000
在这个录像中
09:41
again, my son is being traced out.
209
566000
2000
同样是跟踪我的儿子
09:43
He's leaving red ink behind.
210
568000
2000
他留下了红色的轨迹
09:45
And there's our nanny by the door.
211
570000
2000
我们的保姆在门边
09:47
(Video) Nanny: You want water? (Baby: Aaaa.)
212
572000
3000
(录像)保姆：你要喝水妈？ (宝宝：Aaaa)
09:50
Nanny: All right. (Baby: Aaaa.)
213
575000
3000
保姆：好。(宝宝：Aaaa)
09:53
DR: She offers water,
214
578000
2000
戴·罗伊:她给他水
09:55
and off go the two worms
215
580000
2000
然后两条时空虫
09:57
over to the kitchen to get water.
216
582000
2000
开始移动到厨房拿水
09:59
And what we've done is use the word "water"
217
584000
2000
同时我们所做的就和“水”这个词
10:01
to tag that moment, that bit of activity.
218
586000
2000
联系上了，随着一些动作
10:03
And now we take the power of data
219
588000
2000
然后我们用数据的力量
10:05
and take every time my son
220
590000
3000
每次我儿子
10:08
ever heard the word water
221
593000
2000
听到水这个字
10:10
and the context he saw it in,
222
595000
2000
以及他看见的情景
10:12
and we use it to penetrate through the video
223
597000
3000
我们利用这些来分析整个影片
10:15
and find every activity trace
224
600000
3000
找到每个跟
10:18
that co-occurred with an instance of water.
225
603000
3000
“水”字出现时发生的活动
10:21
And what this data leaves in its wake
226
606000
2000
这个数据勾勒出了
10:23
is a landscape.
227
608000
2000
这么一幅风景
10:25
We call these wordscapes.
228
610000
2000
我们把这个叫做 词景
10:27
This is the wordscape for the word water,
229
612000
2000
这是水字的词景
10:29
and you can see most of the action is in the kitchen.
230
614000
2000
你可以看见大多数行动是在厨房
10:31
That's where those big peaks are over to the left.
231
616000
3000
就是左边的这些高峰
10:34
And just for contrast, we can do this with any word.
232
619000
3000
相对，你也可以为其他词汇勾勒词景
10:37
We can take the word "bye"
233
622000
2000
比如“goog bye”(再见)里的
10:39
as in "good bye."
234
624000
2000
”bye"字
10:41
And we're now zoomed in over the entrance to the house.
235
626000
2000
我们放大到房子大门口附近
10:43
And we look, and we find, as you would expect,
236
628000
3000
我们看到，我们发现，你也会想到
10:46
a contrast in the landscape
237
631000
2000
一幅相对的景象
10:48
where the word "bye" occurs much more in a structured way.
238
633000
3000
在那儿你看到“bye“高频率出现的结构
10:51
So we're using these structures
239
636000
2000
我们用这些结构
10:53
to start predicting
240
638000
2000
开始预言
10:55
the order of language acquisition,
241
640000
3000
学会语言的顺序
10:58
and that's ongoing work now.
242
643000
2000
这是在持续进行的工作
11:00
In my lab, which we're peering into now, at MIT --
243
645000
3000
在我麻省理工学院的研究室-就是现在看到
11:03
this is at the media lab.
244
648000
2000
那是在媒体实验室里
11:05
This has become my favorite way
245
650000
2000
这成了我最喜欢的空间
11:07
of videographing just about any space.
246
652000
2000
视频制图方法
11:09
Three of the key people in this project,
247
654000
2000
这个项目的关键人物都在
11:11
Philip DeCamp, Rony Kubat and Brandon Roy are pictured here.
248
656000
3000
就是图片里的菲利普·迪坎普, 罗尼·库巴特和布兰登·罗伊
11:14
Philip has been a close collaborator
249
659000
2000
菲利普是一个密切的合作者
11:16
on all the visualizations you're seeing.
250
661000
2000
你们看到的视觉化功能就是他负责的
11:18
And Michael Fleischman
251
663000
3000
还有麦克尔·菲莱舍曼
11:21
was another Ph.D. student in my lab
252
666000
2000
是我实验室的另一个博士生
11:23
who worked with me on this home video analysis,
253
668000
3000
和我一起做了家庭视频的分析
11:26
and he made the following observation:
254
671000
3000
是他发表了以下的观点：
11:29
that "just the way that we're analyzing
255
674000
2000
“我们分析
11:31
how language connects to events
256
676000
3000
语言如何于事件相关
11:34
which provide common ground for language,
257
679000
2000
这是语言的共同的基础
11:36
that same idea we can take out of your home, Deb,
258
681000
4000
我们可以把同样的思路带出你的家，戴
11:40
and we can apply it to the world of public media."
259
685000
3000
我们可以把它用到公共媒体上”
11:43
And so our effort took an unexpected turn.
260
688000
3000
所以我们的研究有了个意想不到的转折
11:46
Think of mass media
261
691000
2000
想到大众媒体
11:48
as providing common ground
262
693000
2000
提供共同的基础
11:50
and you have the recipe
263
695000
2000
你就可以把我们的方法
11:52
for taking this idea to a whole new place.
264
697000
3000
运用到一个崭新的地方
11:55
We've started analyzing television content
265
700000
3000
我们开始分析电视内容
11:58
using the same principles --
266
703000
2000
用同样的原则--
12:00
analyzing event structure of a TV signal --
267
705000
3000
分析一个电视信号的事件结构--
12:03
episodes of shows,
268
708000
2000
电视剧集
12:05
commercials,
269
710000
2000
广告
12:07
all of the components that make up the event structure.
270
712000
3000
所有的组成事件结构的成分
12:10
And we're now, with satellite dishes, pulling and analyzing
271
715000
3000
我们现在， 通过卫星电视，抽出分析了
12:13
a good part of all the TV being watched in the United States.
272
718000
3000
在美国高收视率的电视节目
12:16
And you don't have to now go and instrument living rooms with microphones
273
721000
3000
你不再需要把麦克风装在起居室里来
12:19
to get people's conversations,
274
724000
2000
记录人们的对话
12:21
you just tune into publicly available social media feeds.
275
726000
3000
你只要去听公开的社交媒体讯息就可以了
12:24
So we're pulling in
276
729000
2000
我们每个月抽出
12:26
about three billion comments a month,
277
731000
2000
大概30亿个评论
12:28
and then the magic happens.
278
733000
2000
奇迹发生了
12:30
You have the event structure,
279
735000
2000
这中间可以找到事件结构
12:32
the common ground that the words are about,
280
737000
2000
这些词汇的共同基础
12:34
coming out of the television feeds;
281
739000
3000
那些从这次电视讯息里透露出的反馈
12:37
you've got the conversations
282
742000
2000
你得到有关这些
12:39
that are about those topics;
283
744000
2000
话题的对话
12:41
and through semantic analysis --
284
746000
3000
通过语意分析
12:44
and this is actually real data you're looking at
285
749000
2000
你们看到的这个是根据我们的数据处理过后
12:46
from our data processing --
286
751000
2000
的真实的数据结果--
12:48
each yellow line is showing a link being made
287
753000
3000
每条黄线显示一个链接
12:51
between a comment in the wild
288
756000
3000
连接着外界的评论
12:54
and a piece of event structure coming out of the television signal.
289
759000
3000
和电视信号发出的事件结构间的关系
12:57
And the same idea now
290
762000
2000
这都是用同样的思路
12:59
can be built up.
291
764000
2000
构建起来的
13:01
And we get this wordscape,
292
766000
2000
我们得到了这个词汇背景
13:03
except now words are not assembled in my living room.
293
768000
3000
不过现在词汇不是从我的客厅里来的
13:06
Instead, the context, the common ground activities,
294
771000
4000
取而代之的情境，共同基础活动
13:10
are the content on television that's driving the conversations.
295
775000
3000
是电视内容带动的对话
13:13
And what we're seeing here, these skyscrapers now,
296
778000
3000
我们现在看到的这些高耸的结构
13:16
are commentary
297
781000
2000
都是电视评论
13:18
that are linked to content on television.
298
783000
2000
它们跟电视上播放的内容联系着
13:20
Same concept,
299
785000
2000
同样的概念
13:22
but looking at communication dynamics
300
787000
2000
但是你们看见的是它在不同的领域
13:24
in a very different sphere.
301
789000
2000
展现的交流动态
13:26
And so fundamentally, rather than, for example,
302
791000
2000
从根本上，而不是，比如
13:28
measuring content based on how many people are watching,
303
793000
3000
根据收视率衡量内容
13:31
this gives us the basic data
304
796000
2000
这个给了我们观察这些
13:33
for looking at engagement properties of content.
305
798000
3000
内容参与性的最基本的资料
13:36
And just like we can look at feedback cycles
306
801000
3000
就跟我们可以看见家里的
13:39
and dynamics in a family,
307
804000
3000
反馈循环和互动一样
13:42
we can now open up the same concepts
308
807000
3000
我们现在可以利用同样的构想
13:45
and look at much larger groups of people.
309
810000
3000
来观察更大的群体
13:48
This is a subset of data from our database --
310
813000
3000
这是我们资料库里的一个子集
13:51
just 50,000 out of several million --
311
816000
3000
只是几百万信息中的5万条
13:54
and the social graph that connects them
312
819000
2000
社交图是和公开资缘
13:56
through publicly available sources.
313
821000
3000
来自于对大众公开的来源
13:59
And if you put them on one plain,
314
824000
2000
如果你把它们放到平面上
14:01
a second plain is where the content lives.
315
826000
3000
第二个平面是内容活跃的地方
14:04
So we have the programs
316
829000
3000
于是我们有了节目
14:07
and the sporting events
317
832000
2000
体育活动
14:09
and the commercials,
318
834000
2000
广告
14:11
and all of the link structures that tie them together
319
836000
2000
所有的链接结构将它们连在一起
14:13
make a content graph.
320
838000
2000
形成了内容图表
14:15
And then the important third dimension.
321
840000
4000
然后是重要的第三个面向
14:19
Each of the links that you're seeing rendered here
322
844000
2000
大家在这里看到的每个连接
14:21
is an actual connection made
323
846000
2000
是一段内容和有些人评论
14:23
between something someone said
324
848000
3000
和有些人评论
14:26
and a piece of content.
325
851000
2000
间构成的真实联系
14:28
And there are, again, now tens of millions of these links
326
853000
3000
这里的几千万条链
14:31
that give us the connective tissue of social graphs
327
856000
3000
让我们看见了社交图表中的关联组织
14:34
and how they relate to content.
328
859000
3000
和它们跟内容的关系
14:37
And we can now start to probe the structure
329
862000
2000
于是我们可以用有趣的办法来
14:39
in interesting ways.
330
864000
2000
探索这个结构
14:41
So if we, for example, trace the path
331
866000
3000
所以，比如，我们跟踪
14:44
of one piece of content
332
869000
2000
某个内容的发展途经
14:46
that drives someone to comment on it,
333
871000
2000
这促使有人对此发表评论
14:48
and then we follow where that comment goes,
334
873000
3000
然后我们跟踪这些评论的去向
14:51
and then look at the entire social graph that becomes activated
335
876000
3000
然后观察整个活跃的社交图
14:54
and then trace back to see the relationship
336
879000
3000
然后又回头追踪查看那个社交图
14:57
between that social graph and content,
337
882000
2000
和内容之间的关系
14:59
a very interesting structure becomes visible.
338
884000
2000
于是显现出一个非常有趣的结构
15:01
We call this a co-viewing clique,
339
886000
2000
我们称之为 共视团体
15:03
a virtual living room if you will.
340
888000
3000
你可以把它当成一个虚拟的客厅
15:06
And there are fascinating dynamics at play.
341
891000
2000
这里头上演着引人注目的戏剧
15:08
It's not one way.
342
893000
2000
它不是单向的
15:10
A piece of content, an event, causes someone to talk.
343
895000
3000
一个内容，一个事件促使某人发表了意见
15:13
They talk to other people.
344
898000
2000
他们和其他人对话
15:15
That drives tune-in behavior back into mass media,
345
900000
3000
就驱动了大众传媒的收视行为
15:18
and you have these cycles
346
903000
2000
于是出现了这样的循环
15:20
that drive the overall behavior.
347
905000
2000
驱动了整体的收视行为
15:22
Another example -- very different --
348
907000
2000
另一个例子--情况很不同--
15:24
another actual person in our database --
349
909000
3000
我们的资料库里有一位人士--
15:27
and we're finding at least hundreds, if not thousands, of these.
350
912000
3000
其实我们可以找到成千上百个例子
15:30
We've given this person a name.
351
915000
2000
我们给这个人一个名字
15:32
This is a pro-amateur, or pro-am media critic
352
917000
3000
这是一个专业的媒体评论员
15:35
who has this high fan-out rate.
353
920000
3000
有很多粉丝
15:38
So a lot of people are following this person -- very influential --
354
923000
3000
很多人都追随他 -- 很有影响力--
15:41
and they have a propensity to talk about what's on TV.
355
926000
2000
他们很喜欢讨论电视上在播的东西
15:43
So this person is a key link
356
928000
3000
于是这个人就是一个关键的链接
15:46
in connecting mass media and social media together.
357
931000
3000
将大众媒体和社交媒体联系在了一起
15:49
One last example from this data:
358
934000
3000
这份资料的最后一个例子是：
15:52
Sometimes it's actually a piece of content that is special.
359
937000
3000
有时确实是一件特别的内容
15:55
So if we go and look at this piece of content,
360
940000
4000
如果我们回顾这个内容
15:59
President Obama's State of the Union address
361
944000
3000
几个星期前的欧巴马总统
16:02
from just a few weeks ago,
362
947000
2000
国情咨文演讲
16:04
and look at what we find in this same data set,
363
949000
3000
再来看看我们在这组资料中发现些什么
16:07
at the same scale,
364
952000
3000
用同样的尺度来衡量
16:10
the engagement properties of this piece of content
365
955000
2000
这个内容的可参与属性
16:12
are truly remarkable.
366
957000
2000
真的是很神奇的
16:14
A nation exploding in conversation
367
959000
2000
整个国家顿时同步
16:16
in real time
368
961000
2000
爆发了谈话
16:18
in response to what's on the broadcast.
369
963000
3000
是针对广播的东西
16:21
And of course, through all of these lines
370
966000
2000
当然，通过这些线路
16:23
are flowing unstructured language.
371
968000
2000
涌现出了结构的语言
16:25
We can X-ray
372
970000
2000
我们可以在 社交点 上
16:27
and get a real-time pulse of a nation,
373
972000
2000
感受一下这个国家即时的动脉
16:29
real-time sense
374
974000
2000
即时的感受
16:31
of the social reactions in the different circuits in the social graph
375
976000
3000
不同的社会圈的社会反应被内容所激活
16:34
being activated by content.
376
979000
3000
都展示在社会图表上
16:37
So, to summarize, the idea is this:
377
982000
3000
所以， 总结来说，观点是：
16:40
As our world becomes increasingly instrumented
378
985000
3000
当我们的世界变得越来越工具化
16:43
and we have the capabilities
379
988000
2000
我们有能力
16:45
to collect and connect the dots
380
990000
2000
搜集和链接一个一个小点
16:47
between what people are saying
381
992000
2000
将人们的话语
16:49
and the context they're saying it in,
382
994000
2000
和他们说这些话时所处得环境联系起来
16:51
what's emerging is an ability
383
996000
2000
那么呈现的将是洞悉
16:53
to see new social structures and dynamics
384
998000
3000
社会结构和社交动态的新视野
16:56
that have previously not been seen.
385
1001000
2000
那是以前我们没有看见过的
16:58
It's like building a microscope or telescope
386
1003000
2000
这好像是造一个显微镜或者望远镜
17:00
and revealing new structures
387
1005000
2000
展示了我们交流和行为间
17:02
about our own behavior around communication.
388
1007000
3000
的新结构
17:05
And I think the implications here are profound,
389
1010000
3000
我觉得其意义是深远的
17:08
whether it's for science,
390
1013000
2000
无论是对科学而言
17:10
for commerce, for government,
391
1015000
2000
还是对商业，政府而言
17:12
or perhaps most of all,
392
1017000
2000
或许更重要的是
17:14
for us as individuals.
393
1019000
3000
对我们每个人而言
17:17
And so just to return to my son,
394
1022000
3000
所以我们把话题回到我的儿子
17:20
when I was preparing this talk, he was looking over my shoulder,
395
1025000
3000
当我在准备这个演讲时，他在我身后看着
17:23
and I showed him the clips I was going to show to you today,
396
1028000
2000
我给他看了这段我今天将要给你们看的录相
17:25
and I asked him for permission -- granted.
397
1030000
3000
我征求他的同意，他同意了
17:28
And then I went on to reflect,
398
1033000
2000
然后我想
17:30
"Isn't it amazing,
399
1035000
3000
“这真是神奇的事情
17:33
this entire database, all these recordings,
400
1038000
3000
整个数据库， 所有这些录相
17:36
I'm going to hand off to you and to your sister" --
401
1041000
2000
我会给交给你和你的妹妹”
17:38
who arrived two years later --
402
1043000
3000
妹妹是两年后出生的
17:41
"and you guys are going to be able to go back and re-experience moments
403
1046000
3000
“你们两个将能够回顾重温
17:44
that you could never, with your biological memory,
404
1049000
3000
你们生物记忆无法
17:47
possibly remember the way you can now?"
405
1052000
2000
记得的这些时刻。”
17:49
And he was quiet for a moment.
406
1054000
2000
那一刻他很安静
17:51
And I thought, "What am I thinking?
407
1056000
2000
我想：”我在想什么啊？
17:53
He's five years old. He's not going to understand this."
408
1058000
2000
他才5岁， 他不会理解这些。 “
17:55
And just as I was having that thought, he looked up at me and said,
409
1060000
3000
而正当我怎么想着，他抬头对我说：
17:58
"So that when I grow up,
410
1063000
2000
“那等我长大了，
18:00
I can show this to my kids?"
411
1065000
2000
我可以给我的孩子们看，是吗？”
18:02
And I thought, "Wow, this is powerful stuff."
412
1067000
3000
我想：“哇， 这玩意儿真是太强大了。”
18:05
So I want to leave you
413
1070000
2000
所以，我要给各位
18:07
with one last memorable moment
414
1072000
2000
留下最后一个值得回忆的
18:09
from our family.
415
1074000
3000
家庭记忆
18:12
This is the first time our son
416
1077000
2000
这是我儿子第一次
18:14
took more than two steps at once --
417
1079000
2000
走了迈出两步的情形
18:16
captured on film.
418
1081000
2000
拍摄在录像中
18:18
And I really want you to focus on something
419
1083000
3000
我希望你们看的时候
18:21
as I take you through.
420
1086000
2000
注意到其中的一点
18:23
It's a cluttered environment; it's natural life.
421
1088000
2000
周围有点闹，这是自然的环境
18:25
My mother's in the kitchen, cooking,
422
1090000
2000
我妈在厨房做饭
18:27
and, of all places, in the hallway,
423
1092000
2000
就在过道里
18:29
I realize he's about to do it, about to take more than two steps.
424
1094000
3000
我意识到他就要迈步了，大概一两步的样子
18:32
And so you hear me encouraging him,
425
1097000
2000
因此各位可以听到我在鼓励他
18:34
realizing what's happening,
426
1099000
2000
我感到有事要发生
18:36
and then the magic happens.
427
1101000
2000
然后妙事发生了
18:38
Listen very carefully.
428
1103000
2000
请仔细听
18:40
About three steps in,
429
1105000
2000
大概在走了三步后
18:42
he realizes something magic is happening,
430
1107000
2000
他感到了美妙的事情发生了
18:44
and the most amazing feedback loop of all kicks in,
431
1109000
3000
令人惊讶的反应循环作用全部启动
18:47
and he takes a breath in,
432
1112000
2000
他松了一口气
18:49
and he whispers "wow"
433
1114000
2000
轻轻地说了声：“哇”
18:51
and instinctively I echo back the same.
434
1116000
4000
我也凭着直觉说了同样的话
18:56
And so let's fly back in time
435
1121000
3000
我们现在回到那一刻
18:59
to that memorable moment.
436
1124000
2000
回到那个令人难忘的一刻
19:05
(Video) DR: Hey.
437
1130000
2000
(录像) 戴·罗伊：嗨
19:07
Come here.
438
1132000
2000
过来
19:09
Can you do it?
439
1134000
3000
你行吗？
19:13
Oh, boy.
440
1138000
2000
哇，宝贝
19:15
Can you do it?
441
1140000
3000
你行吗？
19:18
Baby: Yeah.
442
1143000
2000
宝宝：好
19:20
DR: Ma, he's walking.
443
1145000
3000
戴1罗伊：妈，他走路了
19:24
(Laughter)
444
1149000
2000
(笑声)
19:26
(Applause)
445
1151000
2000
(掌声)
19:28
DR: Thank you.
446
1153000
2000
戴·罗伊：谢谢大家
19:30
(Applause)
447
1155000
15000
(掌声)
ABOUT THE SPEAKER
Deb Roy - Cognitive scientistDeb Roy studies how children learn language, and designs machines that learn to communicate in human-like ways. On sabbatical from MIT Media Lab, he's working with the AI company Bluefin Labs.
Why you should listen
Deb Roy directs the Cognitive Machines group at the MIT Media Lab, where he studies how children learn language, and designs machines that learn to communicate in human-like ways. To enable this work, he has pioneered new data-driven methods for analyzing and modeling human linguistic and social behavior. He has authored numerous scientific papers on artificial intelligence, cognitive modeling, human-machine interaction, data mining, and information visualization.
Deb Roy was the co-founder and serves as CEO of Bluefin Labs, a venture-backed technology company. Built upon deep machine learning principles developed in his research over the past 15 years, Bluefin has created a technology platform that analyzes social media commentary to measure real-time audience response to TV ads and shows.
Follow Deb Roy on Twitter>
Roy adds some relevant papers:
Deb Roy. (2009). New Horizons in the Study of Child Language Acquisition. Proceedings of Interspeech 2009. Brighton, England. bit.ly/fSP4Qh
Brandon C. Roy, Michael C. Frank and Deb Roy. (2009). Exploring word learning in a high-density longitudinal corpus. Proceedings of the 31st Annual Meeting of the Cognitive Science Society. Amsterdam, Netherlands. bit.ly/e1qxej
Plenty more papers on our research including technology and methodology can be found here, together with other research from my lab at MIT: bit.ly/h3paSQ
The work that I mentioned on relationships between television content and the social graph is being done at Bluefin Labs (www.bluefinlabs.com). Details of this work have not been published. The social structures we are finding (and that I highlighted in my TED talk) are indeed new. The social media communication channels that are leading to their formation did not even exist a few years ago, and Bluefin's technology platform for discovering these kinds of structures is the first of its kind. We'll certainly have more to say about all this as we continue to dig into this fascinating new kind of data, and as new social structures continue to evolve!
Deb Roy | Speaker | TED.com