A headset that reads your brainwaves | Tan Le

377,362 views ・ 2010-07-22

TED


请双击下面的英文字幕来播放视频。

翻译人员: Halei Liu 校对人员: Xu Jiang
00:16
Up until now, our communication with machines
0
16260
2000
直到现在,我们和机器的沟通的方式
00:18
has always been limited
1
18260
2000
都被限制在
00:20
to conscious and direct forms.
2
20260
2000
一种有意识和直接的形式上。
00:22
Whether it's something simple
3
22260
2000
无论是简单得像
00:24
like turning on the lights with a switch,
4
24260
2000
打开灯的开关那样,
00:26
or even as complex as programming robotics,
5
26260
3000
还是复杂得像编写一个机器人那样,
00:29
we have always had to give a command to a machine,
6
29260
3000
我们必须给机器一段指令,
00:32
or even a series of commands,
7
32260
2000
或一系列的指令,
00:34
in order for it to do something for us.
8
34260
3000
才能让机器为我们做我们想做的事情。
00:37
Communication between people, on the other hand,
9
37260
2000
而人和人之间的交流
00:39
is far more complex and a lot more interesting
10
39260
3000
就远远要复杂和有趣些,
00:42
because we take into account
11
42260
2000
因为我们接受的
00:44
so much more than what is explicitly expressed.
12
44260
3000
远比外露的表现要多。
00:47
We observe facial expressions, body language,
13
47260
3000
我们观察脸部表情,肢体语言,
00:50
and we can intuit feelings and emotions
14
50260
2000
我们可以通过对话
00:52
from our dialogue with one another.
15
52260
3000
感受到感觉和情感。
00:55
This actually forms a large part
16
55260
2000
这些其实都是我们做决定的过程中的
00:57
of our decision-making process.
17
57260
2000
一大部分。
00:59
Our vision is to introduce
18
59260
2000
我们的视野是向人机互动科技
01:01
this whole new realm of human interaction
19
61260
3000
介绍一个崭新的人类互动的
01:04
into human-computer interaction
20
64260
2000
新领域,
01:06
so that computers can understand
21
66260
2000
这样的话,计算机就不仅仅
01:08
not only what you direct it to do,
22
68260
2000
按照你的指令工作,
01:10
but it can also respond
23
70260
2000
也能够根据你脸部表情
01:12
to your facial expressions
24
72260
2000
和感情
01:14
and emotional experiences.
25
74260
2000
做出反应。
01:16
And what better way to do this
26
76260
2000
如果要这样做的话
01:18
than by interpreting the signals
27
78260
2000
还有什么能比得上
01:20
naturally produced by our brain,
28
80260
2000
去理解我们脑部所发出的电波,
01:22
our center for control and experience.
29
82260
3000
我们的控制和体验中心。
01:25
Well, it sounds like a pretty good idea,
30
85260
2000
好,这听起来是一个十分好的主意,
01:27
but this task, as Bruno mentioned,
31
87260
2000
但是这项任务,就像Burno提过的,
01:29
isn't an easy one for two main reasons:
32
89260
3000
并不是那么简单,这主要有两个原因:
01:32
First, the detection algorithms.
33
92260
3000
第一,检测的算法。
01:35
Our brain is made up of
34
95260
2000
我们的大脑是由上亿个
01:37
billions of active neurons,
35
97260
2000
活跃的脑神经所组成,
01:39
around 170,000 km
36
99260
3000
轴突的长度总共有
01:42
of combined axon length.
37
102260
2000
170,000 千米。
01:44
When these neurons interact,
38
104260
2000
当这些脑神经互动时,
01:46
the chemical reaction emits an electrical impulse,
39
106260
2000
所产生的化学反应放出的电脉冲
01:48
which can be measured.
40
108260
2000
是可以被测量的。
01:50
The majority of our functional brain
41
110260
3000
我们大脑的主要功能
01:53
is distributed over
42
113260
2000
是分布在大脑
01:55
the outer surface layer of the brain,
43
115260
2000
的外表面层。
01:57
and to increase the area that's available for mental capacity,
44
117260
3000
从心理的能力来说要去增加这个区域是可能的,
02:00
the brain surface is highly folded.
45
120260
3000
大脑表面充满了褶皱。
02:03
Now this cortical folding
46
123260
2000
皮质折叠
02:05
presents a significant challenge
47
125260
2000
对于解析表面电脉冲来说
02:07
for interpreting surface electrical impulses.
48
127260
3000
是一个重大挑战。
02:10
Each individual's cortex
49
130260
2000
每一个单独的皮层
02:12
is folded differently,
50
132260
2000
其折叠的结构都是有区别的,
02:14
very much like a fingerprint.
51
134260
2000
就像我们的指纹一样。
02:16
So even though a signal
52
136260
2000
就算一个信号
02:18
may come from the same functional part of the brain,
53
138260
3000
可能来自于大脑同一功能的部分,
02:21
by the time the structure has been folded,
54
141260
2000
就在这个结构被折叠的时间里,
02:23
its physical location
55
143260
2000
它的实际位置
02:25
is very different between individuals,
56
145260
2000
是十分不同的,
02:27
even identical twins.
57
147260
3000
就算是双胞胎也一样。
02:30
There is no longer any consistency
58
150260
2000
表层的信号
02:32
in the surface signals.
59
152260
2000
没有什么持续性。
02:34
Our breakthrough was to create an algorithm
60
154260
2000
我们的突破是创造一种计算方法
02:36
that unfolds the cortex,
61
156260
2000
可以展开皮层,
02:38
so that we can map the signals
62
158260
2000
这样我们可以在更靠近源头的地方
02:40
closer to its source,
63
160260
2000
来接受信号,
02:42
and therefore making it capable of working across a mass population.
64
162260
3000
从而就可以在更广泛的人群中使用。
02:46
The second challenge
65
166260
2000
第二个挑战是
02:48
is the actual device for observing brainwaves.
66
168260
3000
观测脑部的实际装置。
02:51
EEG measurements typically involve
67
171260
2000
脑电图一般是
02:53
a hairnet with an array of sensors,
68
173260
3000
一系列传感器的阵列,
02:56
like the one that you can see here in the photo.
69
176260
3000
就和你们在照片上所看到的一样。
02:59
A technician will put the electrodes
70
179260
2000
技术人员使用导电胶或粘贴
03:01
onto the scalp
71
181260
2000
将电极
03:03
using a conductive gel or paste
72
183260
2000
放到头皮上
03:05
and usually after a procedure of preparing the scalp
73
185260
3000
这通常会有一个光磨损的过程
03:08
by light abrasion.
74
188260
2000
来准备头皮。
03:10
Now this is quite time consuming
75
190260
2000
这十分耗时
03:12
and isn't the most comfortable process.
76
192260
2000
而且过程也不舒适。
03:14
And on top of that, these systems
77
194260
2000
加上,这个系统
03:16
actually cost in the tens of thousands of dollars.
78
196260
3000
要花费上百万美元。
03:20
So with that, I'd like to invite onstage
79
200260
3000
我现在想邀请我们去年的
03:23
Evan Grant, who is one of last year's speakers,
80
203260
2000
一位演讲者Evan Grant,上台来。
03:25
who's kindly agreed
81
205260
2000
他很客气的同意了
03:27
to help me to demonstrate
82
207260
2000
来帮助我们来展示
03:29
what we've been able to develop.
83
209260
2000
我们的研究发展。
03:31
(Applause)
84
211260
6000
(掌声)
03:37
So the device that you see
85
217260
2000
这个装置如你们所看见的
03:39
is a 14-channel, high-fidelity
86
219260
2000
是一个14个通道,高保真
03:41
EEG acquisition system.
87
221260
2000
脑电采集系统。
03:43
It doesn't require any scalp preparation,
88
223260
3000
不需要任何的头皮处理过程,
03:46
no conductive gel or paste.
89
226260
2000
不用导流胶或导流膏。
03:48
It only takes a few minutes to put on
90
228260
3000
只需要几分钟来固定好
03:51
and for the signals to settle.
91
231260
2000
和稳定信号。
03:53
It's also wireless,
92
233260
2000
这也是无线的,
03:55
so it gives you the freedom to move around.
93
235260
3000
所以这样就可以自由的移动。
03:58
And compared to the tens of thousands of dollars
94
238260
3000
和上百万美元的
04:01
for a traditional EEG system,
95
241260
3000
传统脑电图系统比
04:04
this headset only costs
96
244260
2000
这个装置只用
04:06
a few hundred dollars.
97
246260
2000
几百美元。
04:08
Now on to the detection algorithms.
98
248260
3000
现在说说检测的算法。
04:11
So facial expressions --
99
251260
2000
所以脸部的表情 --
04:13
as I mentioned before in emotional experiences --
100
253260
2000
就像我先前提到的感情表达一样 --
04:15
are actually designed to work out of the box
101
255260
2000
都是意想不到的
04:17
with some sensitivity adjustments
102
257260
2000
通过一些敏感性的调整
04:19
available for personalization.
103
259260
3000
使之个性化。
04:22
But with the limited time we have available,
104
262260
2000
但是由于时间原因,
04:24
I'd like to show you the cognitive suite,
105
264260
2000
我向你们介绍一套认知系统,
04:26
which is the ability for you
106
266260
2000
系统所做的是
04:28
to basically move virtual objects with your mind.
107
268260
3000
让你用你的意念来移动物体。
04:32
Now, Evan is new to this system,
108
272260
2000
现在,Evan是第一次接触这个系统,
04:34
so what we have to do first
109
274260
2000
所以我们要先为他
04:36
is create a new profile for him.
110
276260
2000
创建新的个人信息。
04:38
He's obviously not Joanne -- so we'll "add user."
111
278260
3000
他显然不是Joanne --所以我们选择“添加用户。”
04:41
Evan. Okay.
112
281260
2000
Evan。 搞定。
04:43
So the first thing we need to do with the cognitive suite
113
283260
3000
我们先要做的是
04:46
is to start with training
114
286260
2000
开始训练一个
04:48
a neutral signal.
115
288260
2000
中和的信号。
04:50
With neutral, there's nothing in particular
116
290260
2000
所谓中和,就是Evan不用
04:52
that Evan needs to do.
117
292260
2000
做任何事情。
04:54
He just hangs out. He's relaxed.
118
294260
2000
他只是放松。
04:56
And the idea is to establish a baseline
119
296260
2000
这个过程会给他建立一个地基
04:58
or normal state for his brain,
120
298260
2000
或是他大脑的普通模式,
05:00
because every brain is different.
121
300260
2000
因为每一人的大脑都是不一样的。
05:02
It takes eight seconds to do this,
122
302260
2000
这大概需要8秒的时间。
05:04
and now that that's done,
123
304260
2000
好,现在完成了,
05:06
we can choose a movement-based action.
124
306260
2000
我们可以选择一个以移动为主的动作。
05:08
So Evan, choose something
125
308260
2000
所以Evan选中一个
05:10
that you can visualize clearly in your mind.
126
310260
2000
他可以在他大脑中现形的物体。
05:12
Evan Grant: Let's do "pull."
127
312260
2000
Evan:让我们来作“拉近。”
05:14
Tan Le: Okay, so let's choose "pull."
128
314260
2000
Tan:好,让我们选中“拉近。”
05:16
So the idea here now
129
316260
2000
现在的目标是
05:18
is that Evan needs to
130
318260
2000
Evan要去想象
05:20
imagine the object coming forward
131
320260
2000
这个物体会向
05:22
into the screen,
132
322260
2000
屏幕靠近。
05:24
and there's a progress bar that will scroll across the screen
133
324260
3000
在他做的同时,屏幕上会显示
05:27
while he's doing that.
134
327260
2000
一个进度条。
05:29
The first time, nothing will happen,
135
329260
2000
第一次,什么都没有。
05:31
because the system has no idea how he thinks about "pull."
136
331260
3000
因为系统不知到他所想的“拉近”是什么。
05:34
But maintain that thought
137
334260
2000
但是持续这个想象
05:36
for the entire duration of the eight seconds.
138
336260
2000
程序8秒钟。
05:38
So: one, two, three, go.
139
338260
3000
所以,1,2,3,开始。
05:49
Okay.
140
349260
2000
好。
05:51
So once we accept this,
141
351260
2000
一旦我们接受这个,
05:53
the cube is live.
142
353260
2000
方块就活起来了。
05:55
So let's see if Evan
143
355260
2000
现在我们看Evan
05:57
can actually try and imagine pulling.
144
357260
3000
能不能想象一下“拉近。”
06:00
Ah, good job!
145
360260
2000
哦,干的好!
06:02
(Applause)
146
362260
3000
(掌声)
06:05
That's really amazing.
147
365260
2000
十分令人惊叹。
06:07
(Applause)
148
367260
4000
(掌声)
06:11
So we have a little bit of time available,
149
371260
2000
这样我们还有一点时间,
06:13
so I'm going to ask Evan
150
373260
2000
所以我们让Evan
06:15
to do a really difficult task.
151
375260
2000
做个难一点的任务。
06:17
And this one is difficult
152
377260
2000
这个比较难
06:19
because it's all about being able to visualize something
153
379260
3000
因为这个是要想象
06:22
that doesn't exist in our physical world.
154
382260
2000
一个不存在我们现实世界里的物体。
06:24
This is "disappear."
155
384260
2000
这是“消失。”
06:26
So what you want to do -- at least with movement-based actions,
156
386260
2000
所以,你要做的 -- 先做一个运动为主的动作,
06:28
we do that all the time, so you can visualize it.
157
388260
3000
我们在现实中一直在做这个动作,所以我们可以看到这个动作。
06:31
But with "disappear," there's really no analogies --
158
391260
2000
但是“消失”,从没有过。
06:33
so Evan, what you want to do here
159
393260
2000
所以Even,你现在要做的是
06:35
is to imagine the cube slowly fading out, okay.
160
395260
3000
想象这个正方体会慢慢的消失掉,好吗。
06:38
Same sort of drill. So: one, two, three, go.
161
398260
3000
跟刚才一样。所以,1,2,3,开始。
06:50
Okay. Let's try that.
162
410260
3000
好,让我们试试。
06:53
Oh, my goodness. He's just too good.
163
413260
3000
哦,天哪。他太棒了。
06:57
Let's try that again.
164
417260
2000
我们在试一次。
07:04
EG: Losing concentration.
165
424260
2000
Even:分心了。
07:06
(Laughter)
166
426260
2000
(笑)
07:08
TL: But we can see that it actually works,
167
428260
2000
Tan:但是我们可以看到这是可行的,
07:10
even though you can only hold it
168
430260
2000
就算你只花了一点点的时间
07:12
for a little bit of time.
169
432260
2000
在这个上面。
07:14
As I said, it's a very difficult process
170
434260
3000
就像我说的,去想象“消失”
07:17
to imagine this.
171
437260
2000
是一个很难的过程。
07:19
And the great thing about it is that
172
439260
2000
了不起的事情是
07:21
we've only given the software one instance
173
441260
2000
我们只给了系统一个他如何想象“消失”
07:23
of how he thinks about "disappear."
174
443260
3000
的例子。
07:26
As there is a machine learning algorithm in this --
175
446260
3000
因为这里有一个机器解析的过程 --
07:29
(Applause)
176
449260
4000
(掌声)
07:33
Thank you.
177
453260
2000
谢谢。
07:35
Good job. Good job.
178
455260
3000
做的好,做的好。
07:38
(Applause)
179
458260
2000
(掌声)
07:40
Thank you, Evan, you're a wonderful, wonderful
180
460260
3000
谢谢你,Even,你是这项技术的
07:43
example of the technology.
181
463260
3000
完美代表。
07:46
So, as you can see, before,
182
466260
2000
所以就像你之前看到的,
07:48
there is a leveling system built into this software
183
468260
3000
这个系统是被建入这个软件中
07:51
so that as Evan, or any user,
184
471260
2000
这样就算是Even,或者其他的用户,
07:53
becomes more familiar with the system,
185
473260
2000
都能更熟悉这个系统,
07:55
they can continue to add more and more detections,
186
475260
3000
他们可以不停的加入更多的探测方式,
07:58
so that the system begins to differentiate
187
478260
2000
而系统也会在不同
08:00
between different distinct thoughts.
188
480260
3000
的想法中区分不同的差别。
08:04
And once you've trained up the detections,
189
484260
2000
而且,一旦你训练好了探测功能,
08:06
these thoughts can be assigned or mapped
190
486260
2000
这个功能会可以被分享到
08:08
to any computing platform,
191
488260
2000
任何一种计算器平台,
08:10
application or device.
192
490260
2000
应用程序或装置中。
08:12
So I'd like to show you a few examples,
193
492260
2000
所以我想向你们展示一些例子,
08:14
because there are many possible applications
194
494260
2000
因为这个新界面有很多
08:16
for this new interface.
195
496260
2000
潜在的应用程序。
08:19
In games and virtual worlds, for example,
196
499260
2000
比如说在游戏和虚拟世界中,
08:21
your facial expressions
197
501260
2000
你的脸部表情
08:23
can naturally and intuitively be used
198
503260
2000
可以直观的被用来
08:25
to control an avatar or virtual character.
199
505260
3000
控制虚拟替身或人物。
08:29
Obviously, you can experience the fantasy of magic
200
509260
2000
显然的,你能体验到神奇的魔法
08:31
and control the world with your mind.
201
511260
3000
和用你的意念来控制世界。
08:36
And also, colors, lighting,
202
516260
3000
颜色,和灯光,
08:39
sound and effects
203
519260
2000
声音和特效,
08:41
can dynamically respond to your emotional state
204
521260
2000
也会根据你的感情模式做出相应的反应
08:43
to heighten the experience that you're having, in real time.
205
523260
3000
以此来提高你在现实中的体验。
08:47
And moving on to some applications
206
527260
2000
现在看看世界各地的开发者和研究家们
08:49
developed by developers and researchers around the world,
207
529260
3000
所开发的应用程序,
08:52
with robots and simple machines, for example --
208
532260
3000
用机器人和简单机械,比如说 --
08:55
in this case, flying a toy helicopter
209
535260
2000
在这个例子里,通过想象提升来
08:57
simply by thinking "lift" with your mind.
210
537260
3000
简单的驾驶一个玩具直升机。
09:00
The technology can also be applied
211
540260
2000
这个科技也可以被应用到
09:02
to real world applications --
212
542260
2000
现实生活中 --
09:04
in this example, a smart home.
213
544260
2000
比如,智能家庭。
09:06
You know, from the user interface of the control system
214
546260
3000
你知道的,通过人机界面中的控制系统
09:09
to opening curtains
215
549260
2000
来打开窗帘
09:11
or closing curtains.
216
551260
3000
或关闭窗帘。
09:22
And of course, also to the lighting --
217
562260
3000
当然还有照明 --
09:25
turning them on
218
565260
3000
开灯
09:28
or off.
219
568260
2000
或者关灯。
09:30
And finally,
220
570260
2000
以及最后的,
09:32
to real life-changing applications,
221
572260
2000
可以改变生活的应用程序
09:34
such as being able to control an electric wheelchair.
222
574260
3000
就比如说控制电子轮椅。
09:37
In this example,
223
577260
2000
在这个例子里,
09:39
facial expressions are mapped to the movement commands.
224
579260
3000
脸部表情被用来控制移动命令。
09:42
Man: Now blink right to go right.
225
582260
3000
男子:现在眨右眼往右。
09:50
Now blink left to turn back left.
226
590260
3000
现在眨左眼往左。
10:02
Now smile to go straight.
227
602260
3000
现在微笑往前。
10:08
TL: We really -- Thank you.
228
608260
2000
Tan:我们十分感谢你 -- 谢谢。
10:10
(Applause)
229
610260
5000
(掌声)
10:15
We are really only scratching the surface of what is possible today,
230
615260
3000
我们今天仅仅大致地揭开了这个系统潜力的一角。
10:18
and with the community's input,
231
618260
2000
随着用户群体的投入,
10:20
and also with the involvement of developers
232
620260
2000
开发者
10:22
and researchers from around the world,
233
622260
3000
以及世界各地的研究员的加盟,
10:25
we hope that you can help us to shape
234
625260
2000
我们希望你们可以帮助我们来
10:27
where the technology goes from here. Thank you so much.
235
627260
3000
探寻这项技术将何去何从。十分谢谢你们。
关于本网站

这个网站将向你介绍对学习英语有用的YouTube视频。你将看到来自世界各地的一流教师教授的英语课程。双击每个视频页面上显示的英文字幕,即可从那里播放视频。字幕会随着视频的播放而同步滚动。如果你有任何意见或要求,请使用此联系表与我们联系。

https://forms.gle/WvT1wiN1qDtmnspy7