A headset that reads your brainwaves | Tan Le

翻译人员: Halei Liu 校对人员: Xu Jiang

Up until now, our communication with machines

直到现在，我们和机器的沟通的方式

has always been limited

都被限制在

to conscious and direct forms.

一种有意识和直接的形式上。

Whether it's something simple

无论是简单得像

like turning on the lights with a switch,

打开灯的开关那样，

or even as complex as programming robotics,

还是复杂得像编写一个机器人那样，

we have always had to give a command to a machine,

我们必须给机器一段指令，

or even a series of commands,

或一系列的指令，

in order for it to do something for us.

才能让机器为我们做我们想做的事情。

Communication between people, on the other hand,

而人和人之间的交流

is far more complex and a lot more interesting

就远远要复杂和有趣些，

because we take into account

因为我们接受的

so much more than what is explicitly expressed.

远比外露的表现要多。

We observe facial expressions, body language,

我们观察脸部表情，肢体语言，

and we can intuit feelings and emotions

我们可以通过对话

from our dialogue with one another.

感受到感觉和情感。

This actually forms a large part

这些其实都是我们做决定的过程中的

of our decision-making process.

一大部分。

Our vision is to introduce

我们的视野是向人机互动科技

this whole new realm of human interaction

介绍一个崭新的人类互动的

into human-computer interaction

新领域，

so that computers can understand

这样的话，计算机就不仅仅

not only what you direct it to do,

按照你的指令工作，

but it can also respond

也能够根据你脸部表情

to your facial expressions

和感情

and emotional experiences.

做出反应。

And what better way to do this

如果要这样做的话

than by interpreting the signals

还有什么能比得上

naturally produced by our brain,

去理解我们脑部所发出的电波，

our center for control and experience.

我们的控制和体验中心。

Well, it sounds like a pretty good idea,

好，这听起来是一个十分好的主意，

but this task, as Bruno mentioned,

但是这项任务，就像Burno提过的，

isn't an easy one for two main reasons:

并不是那么简单，这主要有两个原因：

First, the detection algorithms.

第一，检测的算法。

Our brain is made up of

我们的大脑是由上亿个

billions of active neurons,

活跃的脑神经所组成，

around 170,000 km

轴突的长度总共有

of combined axon length.

170,000 千米。

When these neurons interact,

当这些脑神经互动时，

the chemical reaction emits an electrical impulse,

所产生的化学反应放出的电脉冲

which can be measured.

是可以被测量的。

The majority of our functional brain

我们大脑的主要功能

is distributed over

是分布在大脑

the outer surface layer of the brain,

的外表面层。

and to increase the area that's available for mental capacity,

从心理的能力来说要去增加这个区域是可能的，

the brain surface is highly folded.

大脑表面充满了褶皱。

Now this cortical folding

皮质折叠

presents a significant challenge

对于解析表面电脉冲来说

for interpreting surface electrical impulses.

是一个重大挑战。

Each individual's cortex

每一个单独的皮层

is folded differently,

其折叠的结构都是有区别的，

very much like a fingerprint.

就像我们的指纹一样。

So even though a signal

就算一个信号

may come from the same functional part of the brain,

可能来自于大脑同一功能的部分，

by the time the structure has been folded,

就在这个结构被折叠的时间里，

its physical location

它的实际位置

is very different between individuals,

是十分不同的，

even identical twins.

就算是双胞胎也一样。

There is no longer any consistency

表层的信号

in the surface signals.

没有什么持续性。

Our breakthrough was to create an algorithm

我们的突破是创造一种计算方法

that unfolds the cortex,

可以展开皮层，

so that we can map the signals

这样我们可以在更靠近源头的地方

closer to its source,

来接受信号，

and therefore making it capable of working across a mass population.

从而就可以在更广泛的人群中使用。

The second challenge

第二个挑战是

is the actual device for observing brainwaves.

观测脑部的实际装置。

EEG measurements typically involve

脑电图一般是

a hairnet with an array of sensors,

一系列传感器的阵列，

like the one that you can see here in the photo.

就和你们在照片上所看到的一样。

A technician will put the electrodes

技术人员使用导电胶或粘贴

onto the scalp

将电极

using a conductive gel or paste

放到头皮上

and usually after a procedure of preparing the scalp

这通常会有一个光磨损的过程

by light abrasion.

来准备头皮。

Now this is quite time consuming

这十分耗时

and isn't the most comfortable process.

而且过程也不舒适。

And on top of that, these systems

加上，这个系统

actually cost in the tens of thousands of dollars.

要花费上百万美元。

So with that, I'd like to invite onstage

我现在想邀请我们去年的

Evan Grant, who is one of last year's speakers,

一位演讲者Evan Grant，上台来。

who's kindly agreed

他很客气的同意了

to help me to demonstrate

来帮助我们来展示

what we've been able to develop.

我们的研究发展。

(Applause)

（掌声）

So the device that you see

这个装置如你们所看见的

is a 14-channel, high-fidelity

是一个14个通道，高保真

EEG acquisition system.

脑电采集系统。

It doesn't require any scalp preparation,

不需要任何的头皮处理过程，

no conductive gel or paste.

不用导流胶或导流膏。

It only takes a few minutes to put on

只需要几分钟来固定好

and for the signals to settle.

和稳定信号。

It's also wireless,

这也是无线的，

so it gives you the freedom to move around.

所以这样就可以自由的移动。

And compared to the tens of thousands of dollars

和上百万美元的

for a traditional EEG system,

传统脑电图系统比

this headset only costs

这个装置只用

a few hundred dollars.

几百美元。

Now on to the detection algorithms.

现在说说检测的算法。

So facial expressions --

所以脸部的表情 －－

as I mentioned before in emotional experiences --

就像我先前提到的感情表达一样 －－

are actually designed to work out of the box

都是意想不到的

with some sensitivity adjustments

通过一些敏感性的调整

available for personalization.

使之个性化。

But with the limited time we have available,

但是由于时间原因，

I'd like to show you the cognitive suite,

我向你们介绍一套认知系统，

which is the ability for you

系统所做的是

to basically move virtual objects with your mind.

让你用你的意念来移动物体。

Now, Evan is new to this system,

现在，Evan是第一次接触这个系统，

so what we have to do first

所以我们要先为他

is create a new profile for him.

创建新的个人信息。

He's obviously not Joanne -- so we'll "add user."

他显然不是Joanne －－所以我们选择“添加用户。”

Evan. Okay.

Evan。 搞定。

So the first thing we need to do with the cognitive suite

我们先要做的是

is to start with training

开始训练一个

a neutral signal.

中和的信号。

With neutral, there's nothing in particular

所谓中和，就是Evan不用

that Evan needs to do.

做任何事情。

He just hangs out. He's relaxed.

他只是放松。

And the idea is to establish a baseline

这个过程会给他建立一个地基

or normal state for his brain,

或是他大脑的普通模式，

because every brain is different.

因为每一人的大脑都是不一样的。

It takes eight seconds to do this,

这大概需要8秒的时间。

and now that that's done,

好，现在完成了，

we can choose a movement-based action.

我们可以选择一个以移动为主的动作。

So Evan, choose something

所以Evan选中一个

that you can visualize clearly in your mind.

他可以在他大脑中现形的物体。

Evan Grant: Let's do "pull."

Evan：让我们来作“拉近。”

Tan Le: Okay, so let's choose "pull."

Tan：好，让我们选中“拉近。”

So the idea here now

现在的目标是

is that Evan needs to

Evan要去想象

imagine the object coming forward

这个物体会向

into the screen,

屏幕靠近。

and there's a progress bar that will scroll across the screen

在他做的同时，屏幕上会显示

while he's doing that.

一个进度条。

The first time, nothing will happen,

第一次，什么都没有。

because the system has no idea how he thinks about "pull."

因为系统不知到他所想的“拉近”是什么。

But maintain that thought

但是持续这个想象

for the entire duration of the eight seconds.

程序8秒钟。

So: one, two, three, go.

所以，1，2，3，开始。

Okay.

好。

So once we accept this,

一旦我们接受这个，

the cube is live.

方块就活起来了。

So let's see if Evan

现在我们看Evan

can actually try and imagine pulling.

能不能想象一下“拉近。”

Ah, good job!

哦，干的好！

(Applause)

（掌声）

That's really amazing.

十分令人惊叹。

(Applause)

（掌声）

So we have a little bit of time available,

这样我们还有一点时间，

so I'm going to ask Evan

所以我们让Evan

to do a really difficult task.

做个难一点的任务。

And this one is difficult

这个比较难

because it's all about being able to visualize something

因为这个是要想象

that doesn't exist in our physical world.

一个不存在我们现实世界里的物体。

This is "disappear."

这是“消失。”

So what you want to do -- at least with movement-based actions,

所以，你要做的 －－ 先做一个运动为主的动作，

we do that all the time, so you can visualize it.

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。

But with "disappear," there's really no analogies --

但是“消失”，从没有过。

so Evan, what you want to do here

所以Even，你现在要做的是

is to imagine the cube slowly fading out, okay.

想象这个正方体会慢慢的消失掉，好吗。

Same sort of drill. So: one, two, three, go.

跟刚才一样。所以，1，2，3，开始。

Okay. Let's try that.

好，让我们试试。

Oh, my goodness. He's just too good.

哦，天哪。他太棒了。

Let's try that again.

我们在试一次。

EG: Losing concentration.

Even：分心了。

(Laughter)

（笑）

TL: But we can see that it actually works,

Tan：但是我们可以看到这是可行的，

even though you can only hold it

就算你只花了一点点的时间

for a little bit of time.

在这个上面。

As I said, it's a very difficult process

就像我说的，去想象“消失”

to imagine this.

是一个很难的过程。

And the great thing about it is that

了不起的事情是

we've only given the software one instance

我们只给了系统一个他如何想象“消失”

of how he thinks about "disappear."

的例子。

As there is a machine learning algorithm in this --

因为这里有一个机器解析的过程 －－

(Applause)

（掌声）

Thank you.

谢谢。

Good job. Good job.

做的好，做的好。

(Applause)

（掌声）

Thank you, Evan, you're a wonderful, wonderful

谢谢你，Even，你是这项技术的

example of the technology.

完美代表。

So, as you can see, before,

所以就像你之前看到的，

there is a leveling system built into this software

这个系统是被建入这个软件中

so that as Evan, or any user,

这样就算是Even，或者其他的用户，

becomes more familiar with the system,

都能更熟悉这个系统，

they can continue to add more and more detections,

他们可以不停的加入更多的探测方式，

so that the system begins to differentiate

而系统也会在不同

between different distinct thoughts.

的想法中区分不同的差别。

And once you've trained up the detections,

而且，一旦你训练好了探测功能，

these thoughts can be assigned or mapped

这个功能会可以被分享到

to any computing platform,

任何一种计算器平台，

application or device.

应用程序或装置中。

So I'd like to show you a few examples,

所以我想向你们展示一些例子，

because there are many possible applications

因为这个新界面有很多

for this new interface.

潜在的应用程序。

In games and virtual worlds, for example,

比如说在游戏和虚拟世界中，

your facial expressions

你的脸部表情

can naturally and intuitively be used

可以直观的被用来

to control an avatar or virtual character.

控制虚拟替身或人物。

Obviously, you can experience the fantasy of magic

显然的，你能体验到神奇的魔法

and control the world with your mind.

和用你的意念来控制世界。

And also, colors, lighting,

颜色，和灯光，

sound and effects

声音和特效，

can dynamically respond to your emotional state

也会根据你的感情模式做出相应的反应

to heighten the experience that you're having, in real time.

以此来提高你在现实中的体验。

And moving on to some applications

现在看看世界各地的开发者和研究家们

developed by developers and researchers around the world,

所开发的应用程序，

with robots and simple machines, for example --

用机器人和简单机械，比如说 －－

in this case, flying a toy helicopter

在这个例子里，通过想象提升来

simply by thinking "lift" with your mind.

简单的驾驶一个玩具直升机。

The technology can also be applied

这个科技也可以被应用到

to real world applications --

现实生活中 －－

in this example, a smart home.

比如，智能家庭。

You know, from the user interface of the control system

你知道的，通过人机界面中的控制系统

to opening curtains

来打开窗帘

or closing curtains.

或关闭窗帘。

And of course, also to the lighting --

当然还有照明 －－

turning them on

开灯

or off.

或者关灯。

And finally,

以及最后的，

to real life-changing applications,

可以改变生活的应用程序

such as being able to control an electric wheelchair.

就比如说控制电子轮椅。

In this example,

在这个例子里，

facial expressions are mapped to the movement commands.

脸部表情被用来控制移动命令。

Man: Now blink right to go right.

男子：现在眨右眼往右。

Now blink left to turn back left.

现在眨左眼往左。

Now smile to go straight.

现在微笑往前。

TL: We really -- Thank you.

Tan：我们十分感谢你 －－ 谢谢。

(Applause)

（掌声）

We are really only scratching the surface of what is possible today,

我们今天仅仅大致地揭开了这个系统潜力的一角。

and with the community's input,

随着用户群体的投入，

and also with the involvement of developers

开发者

and researchers from around the world,

以及世界各地的研究员的加盟，

we hope that you can help us to shape

我们希望你们可以帮助我们来

where the technology goes from here. Thank you so much.

探寻这项技术将何去何从。十分谢谢你们。