has always been limited

直到現在，我們與機器的溝通

to conscious and direct forms.

仍局限於

Whether it's something simple

有意識和直接的模式

like turning on the lights with a switch,

不論是一些簡單的事情

or even as complex as programming robotics,

如用開關開燈

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

或一些複雜的程式來控制機械人

or even a series of commands,

我們都要給機器輸入一個

in order for it to do something for us.

甚至一系列的指令

Communication between people, on the other hand,

才能命令它執行一些動作

is far more complex and a lot more interesting

相反的，人與人的溝通

because we take into account

就更加複雜和有趣得多

so much more than what is explicitly expressed.

因為我們會考慮到

We observe facial expressions, body language,

言語未表達的言外之意

and we can intuit feelings and emotions

我們會觀察表情、肢體語言

from our dialogue with one another.

在對話中我們會用直覺來

This actually forms a large part

感受對方的感覺和情緒

of our decision-making process.

這些都是做決定時

Our vision is to introduce

一些重要的因素

this whole new realm of human interaction

我們的願景是引進

into human-computer interaction

全新的人與電腦的互動科技

so that computers can understand

到人類互動的領域

not only what you direct it to do,

這麼一來電腦不只可以

but it can also respond

明白你指示它所做的事情

to your facial expressions

而且也會對面部表情

and emotional experiences.

和情緒經歷

And what better way to do this

作出反應

than by interpreting the signals

還有什麼比從大腦的

naturally produced by our brain,

情感控制中樞直接解譯

our center for control and experience.

大腦產生的電波

Well, it sounds like a pretty good idea,

來得更好呢？

but this task, as Bruno mentioned,

這聽起來好像是不錯的主意

isn't an easy one for two main reasons:

但這個任務，正如Bruno所說

First, the detection algorithms.

並不容易，原因有兩個

Our brain is made up of

第一是大腦的偵查演算法

billions of active neurons,

我們的腦是由

around 170,000 km

數十億個活躍的神經元所組成

of combined axon length.

如果把神經細胞的軸索連在一起

When these neurons interact,

大概有十七萬公里

the chemical reaction emits an electrical impulse,

這些神經元互動時

which can be measured.

產生的化學作用所發射出的電脈衝

The majority of our functional brain

能夠被測量到

is distributed over

大部分功能性腦

the outer surface layer of the brain,

是分佈在

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

大腦的表層

the brain surface is highly folded.

心智能力功能也位於此，為了增加表面積

Now this cortical folding

大腦皮質層有非常多的褶皺

presents a significant challenge

大腦皮質褶皺

for interpreting surface electrical impulses.

對分析電脈衝

Each individual's cortex

帶來一個很大的挑戰

is folded differently,

每個人大腦皮質層

very much like a fingerprint.

的褶皺都不同

So even though a signal

就像指紋一樣

may come from the same functional part of the brain,

因此電脈衝訊息

by the time the structure has been folded,

雖然來自功能腦同樣的區域

its physical location

但大腦皮質褶皺結構早已形成

is very different between individuals,

在不同的人的大腦裡

even identical twins.

即使是雙胞胎

There is no longer any consistency

訊息發生位置也不同

in the surface signals.

大腦皮質層電脈衝訊息

Our breakthrough was to create an algorithm

沒有一致性

that unfolds the cortex,

我們的突破是建立一個演算法

so that we can map the signals

攤開大腦皮質層

closer to its source,

去勘測這些

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

訊息的原點

The second challenge

繼而把它運用在大眾身上

is the actual device for observing brainwaves.

第二項挑戰是

EEG measurements typically involve

觀察腦電波的儀器

a hairnet with an array of sensors,

腦波測量基本上包括

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

一個有許多感應器的髮網

A technician will put the electrodes

就像現在圖中所看到的

onto the scalp

技術人員會把電極

using a conductive gel or paste

用導電的膠或漿糊

and usually after a procedure of preparing the scalp

固定在頭皮上

by light abrasion.

這個準備程序需要在頭皮製造

Now this is quite time consuming

輕微的擦傷

and isn't the most comfortable process.

這個程序既費時

And on top of that, these systems

又不舒服

actually cost in the tens of thousands of dollars.

再加上，這些系統

So with that, I'd like to invite onstage

非常昂貴，得花上數萬美金

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

現在，我邀請Evan Grant

who's kindly agreed

去年的演講者上台

to help me to demonstrate

他很樂意

what we've been able to develop.

幫忙示範

(Applause)

我們所設計的儀器

So the device that you see

(鼓掌)

is a 14-channel, high-fidelity

你們所看到的儀器是

EEG acquisition system.

有十四個頻道，高傳真的

It doesn't require any scalp preparation,

腦電波訊號擷取系統

no conductive gel or paste.

不需要任何頭皮準備程序

It only takes a few minutes to put on

沒有導電的膠或漿糊

and for the signals to settle.

戴上它，等訊號穩定

It's also wireless,

只要幾分鐘

so it gives you the freedom to move around.

而且是無線的

And compared to the tens of thousands of dollars

它讓你活動自如

for a traditional EEG system,

比起那些幾萬美元的

this headset only costs

傳統腦電波系統

a few hundred dollars.

這個頭戴式耳機

Now on to the detection algorithms.

只要幾百美金

So facial expressions --

現在來談談大腦感應演算法

as I mentioned before in emotional experiences --

好，面部表情--

are actually designed to work out of the box

如同之前講到的情緒經驗--

with some sensitivity adjustments

這套系統有令人意想不到的設計

available for personalization.

只要做一些敏感度調整

But with the limited time we have available,

就可以運用於個人化的使用

I'd like to show you the cognitive suite,

但因時間的關係

which is the ability for you

現在只示範認知的部份

to basically move virtual objects with your mind.

這套系統能夠讓您

Now, Evan is new to this system,

只用意念移動虛擬物件

so what we have to do first

Evan是第一次接觸這個系統

is create a new profile for him.

因此我們要先

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

建立一個新的檔案

Evan. Okay.

他當然不是Joanne， 所以要增加一個用戶

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

Evan，好了!

is to start with training

首先要做的是

a neutral signal.

練習發出一個

With neutral, there's nothing in particular

中立的訊號

that Evan needs to do.

Evan不需要做

He just hangs out. He's relaxed.

什麼特別的事

And the idea is to establish a baseline

就這樣放輕鬆

or normal state for his brain,

重點是建立一個基準線

because every brain is different.

或是大腦的正常狀態

It takes eight seconds to do this,

因為每個人的腦都不相同

and now that that's done,

這大概需要八秒的時間

we can choose a movement-based action.

完成了

So Evan, choose something

我們可以選擇一個有動作的活動

that you can visualize clearly in your mind.

Evan，你可選擇一個

Evan Grant: Let's do "pull."

在你腦海中可以清楚看到的事情

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

讓我們做一個"拉"的動作

So the idea here now

好，點選"拉"

is that Evan needs to

我們現在

imagine the object coming forward

需要Evan想像

into the screen,

一件物品在螢幕上

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

往前移動

while he's doing that.

他這樣做的時候

The first time, nothing will happen,

螢幕上會出現一個測量棒

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

第一次沒有任何事情發生

But maintain that thought

因為系統還不知道他怎麼想像"拉"的動作

for the entire duration of the eight seconds.

在這八秒中

So: one, two, three, go.

持續想著這個念頭

Okay.

一、二、三、開始

So once we accept this,

好了

the cube is live.

當我們按了接受

So let's see if Evan

這個方塊就活了起來

can actually try and imagine pulling.

讓我們看看Evan

Ah, good job!

能否真的嘗試想像"拉"的動作

(Applause)

哇! 非常好!

That's really amazing.

(鼓掌)

(Applause)

真是令人驚訝！

So we have a little bit of time available,

(鼓掌)

so I'm going to ask Evan

我們還有一些時間

to do a really difficult task.

我要請Evan

And this one is difficult

做一些比較困難的動作

because it's all about being able to visualize something

這個有點難

that doesn't exist in our physical world.

因為要想像

This is "disappear."

在物質界裡不存在的事物

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

就是 "消失"

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

就動作而言

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

因為經常做這些動作，所以能"看見"它

so Evan, what you want to do here

但"消失"沒有任何類似的動作

is to imagine the cube slowly fading out, okay.

Evan, 現在請你

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

想像這個方塊慢慢消失

Okay. Let's try that.

一樣的練習。 一、二、三、開始

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

可以了，我們試試吧

Let's try that again.

我的天啊！他真的是非常厲害

EG: Losing concentration.

再試一次

(Laughter)

（EG儀器:） 失去專注力

TL: But we can see that it actually works,

(笑聲)

even though you can only hold it

這套系統真的辦到了

for a little bit of time.

雖然只維持

As I said, it's a very difficult process

一段很短的時間

to imagine this.

我認為想像"消失"

And the great thing about it is that

真的是非常困難

we've only given the software one instance

這個系統了不起的是

of how he thinks about "disappear."

這套軟體只有一次機會

As there is a machine learning algorithm in this --

知道Evan是怎麼想像"消失"的

(Applause)

而這部機器便學會了演算它

Thank you.

(鼓掌)

Good job. Good job.

謝謝

(Applause)

很棒！很棒！

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

(鼓掌)

example of the technology.

謝謝，Evan你真的是這項科技

So, as you can see, before,

最佳的展示人員

there is a leveling system built into this software

正如你們所見

so that as Evan, or any user,

這個軟體有一個水準測量系統

becomes more familiar with the system,

Evan或其他使用者

they can continue to add more and more detections,

對這個系統越熟悉

so that the system begins to differentiate

就能不斷地增加更多，更多的檢測項目

between different distinct thoughts.

這個系統就能開始分辨

And once you've trained up the detections,

不同的明顯想法

these thoughts can be assigned or mapped

當你訓練做這些檢測項目

to any computing platform,

這些念頭、想法就能指定或聯繫到

application or device.

任何的電腦平台、

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

應用程式或儀器上

because there are many possible applications

讓我為你們展示幾個例子

for this new interface.

這個新界面有

In games and virtual worlds, for example,

很多可運用的應用程式

your facial expressions

例如在遊戲或虛擬世界

can naturally and intuitively be used

你可以用臉部表情

to control an avatar or virtual character.

自然、直覺地

Obviously, you can experience the fantasy of magic

操控遊戲角色或虛擬人物

and control the world with your mind.

無庸置疑，你將會親身體驗幻想的魔力

And also, colors, lighting,

和運用意念來控制世界

sound and effects

顏色，燈光

can dynamically respond to your emotional state

聲音和音效

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

也可以不斷地變化來反映你的情緒狀態

And moving on to some applications

即時強化你的感受

developed by developers and researchers around the world,

現在來看看應用程式

with robots and simple machines, for example --

全世界的研發人員發明了

in this case, flying a toy helicopter

不同的機械人和簡單的機器，例如

simply by thinking "lift" with your mind.

這個例子是操作玩具直昇機

The technology can also be applied

只要用意念就可以讓它飛起來

to real world applications --

這項科技也可以應用在

in this example, a smart home.

實際生活中

You know, from the user interface of the control system

看看智能家居的例子

to opening curtains

從使用者界面控制系統

or closing curtains.

來打開

And of course, also to the lighting --

或關上窗簾

turning them on

當然電燈也可以

or off.

開

And finally,

或關

to real life-changing applications,

最後

such as being able to control an electric wheelchair.

是應用在改善真實生活

In this example,

例如能夠控制電動輪椅

facial expressions are mapped to the movement commands.

這個例子裡

Man: Now blink right to go right.

面部表情對應於移動方向的指令

Now blink left to turn back left.

男聲: 現在眨右眼右轉

Now smile to go straight.

眨左眼左轉

TL: We really -- Thank you.

微笑往前

(Applause)

TL: 我們真的.... 多謝各位。

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

(鼓掌)

and with the community's input,

現今我們所做到的只是很小的一部分

and also with the involvement of developers

有研發團隊的投入

and researchers from around the world,

及全世界的研發和

we hope that you can help us to shape

研究人員的參與

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

我們希望這一項科技能夠