Rana el Kaliouby: This app knows how you feel — from the look on your face

譯者: Regina Chu 審譯者: Marssi Draw

Our emotions influence every aspect of our lives,

我們的情緒會影響 日常生活的各個層面，

from our health and how we learn, to how we do business and make decisions,

從我們的健康到如何學習、 如何做事、做決定，

big ones and small.

無論事情大小都受此影響。

Our emotions also influence how we connect with one another.

我們的情緒也會影響 我們如何與他人交流。

We've evolved to live in a world like this,

我們已經進化到生活在 一個像這樣的世界，

but instead, we're living more and more of our lives like this --

然而我們的生活卻愈來愈像這樣──

this is the text message from my daughter last night --

這是我女兒昨晚傳來的簡訊──

in a world that's devoid of emotion.

一個缺乏情感的世界。

So I'm on a mission to change that.

所以我帶著使命要改變這種狀況。

I want to bring emotions back into our digital experiences.

我想將情感重新注入數位體驗中。

I started on this path 15 years ago.

我在 15 年前走上這條路。

I was a computer scientist in Egypt,

當時我在埃及是電腦科學家，

and I had just gotten accepted to a Ph.D. program at Cambridge University.

而且我才拿到劍橋大學 博士班的入學許可。

So I did something quite unusual

所以我做了一件 對身為年輕新婚的埃及回教婦女來說

for a young newlywed Muslim Egyptian wife:

相當不尋常的事：

With the support of my husband, who had to stay in Egypt,

在我先生的支持下， 他留在埃及，

I packed my bags and I moved to England.

我整理行囊搬到英格蘭。

At Cambridge, thousands of miles away from home,

在劍橋，離家千里遠的地方，

I realized I was spending more hours with my laptop

我發現我與筆電相處的時間，

than I did with any other human.

遠超過與人交流的時間。

Yet despite this intimacy, my laptop had absolutely no idea how I was feeling.

儘管與筆電相處如此親密， 它卻完全不了解我的感受，

It had no idea if I was happy,

它不知道我是否開心，

having a bad day, or stressed, confused,

今天順不順，是否緊張或困惑，

and so that got frustrating.

所以那令我沮喪。

Even worse, as I communicated online with my family back home,

更糟的是，在我上線 與遠方的家人聯絡時，

I felt that all my emotions disappeared in cyberspace.

我覺得我的情感 在這虛擬空間裡消失無蹤。

I was homesick, I was lonely, and on some days I was actually crying,

我好想家，我好孤單， 有些日子我真的哭了，

but all I had to communicate these emotions was this.

但我所能傳達的只有這個。

(Laughter)

（笑聲）

Today's technology has lots of I.Q., but no E.Q.;

今天的科技有很多智商， 卻沒有情緒智商；

lots of cognitive intelligence, but no emotional intelligence.

有很多認知智商， 卻沒有情緒智商。

So that got me thinking,

所以這讓我思考，

what if our technology could sense our emotions?

如果我們的科技可以 感受我們的情緒會怎樣？

What if our devices could sense how we felt and reacted accordingly,

如果我們的電子裝置可以 感受我們的感覺並做出相對回應，

just the way an emotionally intelligent friend would?

就像一位高情商的朋友一樣， 會是怎樣？

Those questions led me and my team

這些問題讓我及我的團隊

to create technologies that can read and respond to our emotions,

創造出可以讀懂情緒 並做出回應的科技，

and our starting point was the human face.

我們的起始點是人的臉。

So our human face happens to be one of the most powerful channels

人類的臉恰好就是有力的管道，

that we all use to communicate social and emotional states,

能用來傳遞社交及情緒狀態，

everything from enjoyment, surprise,

從愉快、驚訝，

empathy and curiosity.

到同情、好奇都可以。

In emotion science, we call each facial muscle movement an action unit.

情緒科學中，我們稱每一種 顏面肌肉運動為一個動作單位。

So for example, action unit 12,

舉例來說，動作單位 12，

it's not a Hollywood blockbuster,

這可不是好萊塢的動作巨片，

it is actually a lip corner pull, which is the main component of a smile.

這其實是拉嘴角， 這是微笑的主要部分。

Try it everybody. Let's get some smiles going on.

大家都試一下吧！ 讓會場有點笑容。

Another example is action unit 4. It's the brow furrow.

另一個例子是動作單位 4。 這是蹙額。

It's when you draw your eyebrows together

就是你把眉頭皺在一起

and you create all these textures and wrinkles.

所產生的紋理和皺紋。

We don't like them, but it's a strong indicator of a negative emotion.

我們都不喜歡皺紋， 但那是負面情緒的重要指標。

So we have about 45 of these action units,

我們有約 45 種動作單位，

and they combine to express hundreds of emotions.

排列組合後可以表現出數百種情緒。

Teaching a computer to read these facial emotions is hard,

要教電腦讀懂這些顏面表情很難，

because these action units, they can be fast, they're subtle,

因為這些動作單位很快、很細微，

and they combine in many different ways.

而且還有各種不同的組合法。

So take, for example, the smile and the smirk.

所以再舉個例子，微笑和假笑。

They look somewhat similar, but they mean very different things.

兩者看起來有點像， 但是意義大不相同。

(Laughter)

（笑聲）

So the smile is positive,

微笑是正面的，

a smirk is often negative.

假笑往往是負面的。

Sometimes a smirk can make you become famous.

有時候一個假笑可以讓你成名。

But seriously, it's important for a computer to be able

但是說真的，要讓電腦能夠

to tell the difference between the two expressions.

辨認出這兩種表情的不同很重要。

So how do we do that?

所以我們怎麼做呢？

We give our algorithms

我們給我們的演算法

tens of thousands of examples of people we know to be smiling,

成千上萬筆我們知道在微笑的例子，

from different ethnicities, ages, genders,

各式人種、年齡、性別都有，

and we do the same for smirks.

假笑也如法泡製。

And then, using deep learning,

然後，機器用深度學習法，

the algorithm looks for all these textures and wrinkles

讓演算法找出臉上 所有的紋理、皺紋，

and shape changes on our face,

及臉型的改變，

and basically learns that all smiles have common characteristics,

基本上學得所有的微笑 都有共同的特點，

all smirks have subtly different characteristics.

所有的假笑也有 稍稍不同的特點，

And the next time it sees a new face,

所以下一次電腦看到新的面孔，

it essentially learns that

它基本上會得知

this face has the same characteristics of a smile,

這張臉與微笑有相同的特點，

and it says, "Aha, I recognize this. This is a smile expression."

然後它會說，「啊哈！ 我認得這個，這是微笑的表情。」

So the best way to demonstrate how this technology works

要展示怎麼用 這項科技的最佳方法，

is to try a live demo,

就是來一個現場示範，

so I need a volunteer, preferably somebody with a face.

所以我需要一名志願者， 最好是有臉的。

(Laughter)

（笑聲）

Cloe's going to be our volunteer today.

我們今天的志願者是克蘿伊。

So over the past five years, we've moved from being a research project at MIT

過去五年，我們從 麻省理工的一項研究計畫

to a company,

發展成一家公司，

where my team has worked really hard to make this technology work,

我的團隊很努力 讓這項科技能快速傳播，

as we like to say, in the wild.

好像我們常說的，（病毒）擴散中。

And we've also shrunk it so that the core emotion engine

我們也把它縮小， 讓核心情緒引擎能用在

works on any mobile device with a camera, like this iPad.

任何有照相機的行動裝置上， 像是這台 iPad。

So let's give this a try.

現在來試一下。

As you can see, the algorithm has essentially found Cloe's face,

正如你們所見，基本上 演算法已經找到了克蘿伊的臉，

so it's this white bounding box,

就是這個白色的框框，

and it's tracking the main feature points on her face,

它正在找她臉上的 幾個主要特徵點，

so her eyebrows, her eyes, her mouth and her nose.

像是她的眉毛、 眼睛、嘴巴和鼻子。

The question is, can it recognize her expression?

問題是，它能辨識她的表情嗎？

So we're going to test the machine.

我們來考一下機器。

So first of all, give me your poker face. Yep, awesome. (Laughter)

首先，來一張撲克臉。 對，好極了！（笑聲）

And then as she smiles, this is a genuine smile, it's great.

然後她微笑的時後， 這是真誠的微笑，很棒，

So you can see the green bar go up as she smiles.

你們可以看到她微笑的時候， 綠色的信號格增加。

Now that was a big smile.

那可是個好大的微笑。

Can you try a subtle smile to see if the computer can recognize?

你可以試一下淺淺的微笑嗎？ 看看電腦能不能辨識？

It does recognize subtle smiles as well.

它的確也能辨識淺淺的微笑。

We've worked really hard to make that happen.

我們真的很努力要做到這一點。

And then eyebrow raised, indicator of surprise.

然後抬眉毛，表示驚訝。

Brow furrow, which is an indicator of confusion.

蹙額，表示困惑。

Frown. Yes, perfect.

皺眉，很好，很完美。

So these are all the different action units. There's many more of them.

這些就是不同的動作單位。 還有更多。

This is just a slimmed-down demo.

這只是瘦身版示範。

But we call each reading an emotion data point,

我們稱每一個讀取 為一個情緒資料點，

and then they can fire together to portray different emotions.

然後它們一起發動 就能描繪出不同的情緒。

So on the right side of the demo -- look like you're happy.

右邊的這張示範── 表現你很開心。

So that's joy. Joy fires up.

所以那是高興。高興出現了。

And then give me a disgust face.

然後給我一張噁心的臉。

Try to remember what it was like when Zayn left One Direction.

試著回想贊恩退出 男團一世代的那種感覺。

(Laughter)

（笑聲）

Yeah, wrinkle your nose. Awesome.

沒錯，皺鼻子。太棒了！

And the valence is actually quite negative, so you must have been a big fan.

效價呈現高負值， 所以你一定是大粉絲。

So valence is how positive or negative an experience is,

效價指的是感受的好壞程度，

and engagement is how expressive she is as well.

而投入程度指的是 她的表情有多大。

So imagine if Cloe had access to this real-time emotion stream,

想像一下如果克羅伊 能使用這套即時情緒串流，

and she could share it with anybody she wanted to.

而且她還可以跟任何人分享。

Thank you.

謝謝妳！

(Applause)

（掌聲）

So, so far, we have amassed 12 billion of these emotion data points.

到目前為止我們已經 累積了 120 億筆情緒數據點。

It's the largest emotion database in the world.

這是世界上最大的情緒資料庫。

We've collected it from 2.9 million face videos,

我們從 290 萬筆臉孔短片 收集資料，

people who have agreed to share their emotions with us,

由同意與我們分享他們情緒的人提供，

and from 75 countries around the world.

來源遍及全球 75 個國家。

It's growing every day.

資料每天都在增加。

It blows my mind away

這真令我驚異萬分，

that we can now quantify something as personal as our emotions,

我們能量化像情緒 這麼個人的東西，

and we can do it at this scale.

還能做到這個地步。

So what have we learned to date?

所以至今我們學到什麼？

Gender.

性別。

Our data confirms something that you might suspect.

我們的數據證實了一些 你們大概已經料到的事。

Women are more expressive than men.

女人的表情比男人的更豐富。

Not only do they smile more, their smiles last longer,

她們不但更常微笑， 微笑的時間還更久，

and we can now really quantify what it is that men and women

而且我們現在真的能量化

造成男女不同反應的東西。

respond to differently.

Let's do culture: So in the United States,

來看文化：在美國，

women are 40 percent more expressive than men,

女性比男性多 40% 更願意表達情感，

but curiously, we don't see any difference in the U.K. between men and women.

但奇怪的是， 在英國看不到這樣的差距。

(Laughter)

（笑聲）

Age: People who are 50 years and older

再看年齡：50 歲以上的人

are 25 percent more emotive than younger people.

比年輕人多 25% 更願意表現情感。

Women in their 20s smile a lot more than men the same age,

20 多歲的女性 比同年齡的男性更常微笑，

perhaps a necessity for dating.

大概是因為這是約會必殺技。

But perhaps what surprised us the most about this data

但是這筆數據最讓我們訝異的，

is that we happen to be expressive all the time,

大概是我們隨時都有表情，

even when we are sitting in front of our devices alone,

即使我們獨自坐在裝置前也是如此，

and it's not just when we're watching cat videos on Facebook.

而且不只是在我們看 臉書上貓短片的的時候。

We are expressive when we're emailing, texting, shopping online,

我們在寫信、傳簡訊、網購，

or even doing our taxes.

甚至在報稅時都表情豐富。

Where is this data used today?

今天這筆數據用在哪裡呢？

In understanding how we engage with media,

用在瞭解我們如何與媒體互動，

so understanding virality and voting behavior;

所以能瞭解影片爆紅及投票行為，

and also empowering or emotion-enabling technology,

也用在情緒辨識科技，

and I want to share some examples that are especially close to my heart.

我想分享幾個 讓我特別感動的例子。

Emotion-enabled wearable glasses can help individuals

情緒辨識眼鏡能幫助

who are visually impaired read the faces of others,

視障者讀取別人臉上的表情，

and it can help individuals on the autism spectrum interpret emotion,

也能幫助各種程度的 自閉症患者解讀情緒，

something that they really struggle with.

這是他們的最大難題。

In education, imagine if your learning apps

在教育上，想像一下 如果你的學習應用程式

sense that you're confused and slow down,

感受到你的困惑並放慢速度，

or that you're bored, so it's sped up,

或是知道你覺得無聊了 所以加快速度，

just like a great teacher would in a classroom.

就像一位好老師 在課堂上做的一樣。

What if your wristwatch tracked your mood,

如果你的手錶能追蹤你的心情，

or your car sensed that you're tired,

或是你的車能感受到 你現在很疲倦，

or perhaps your fridge knows that you're stressed,

或是你的冰箱能知道 你現在壓力很大，

so it auto-locks to prevent you from binge eating. (Laughter)

所以它會自動鎖住， 你就不能拿東西來吃。（笑聲）

我會喜歡那個，真的。

I would like that, yeah.

What if, when I was in Cambridge,

當我在劍橋的時候，

I had access to my real-time emotion stream,

如果我能用這套 即時情緒串流工具，

and I could share that with my family back home in a very natural way,

我就能用非常自然的方法 與遠在家鄉的家人分享，

just like I would've if we were all in the same room together?

就好像我們都在 同一間房間一樣，那有多好？

I think five years down the line,

我想五年後，

all our devices are going to have an emotion chip,

我們所有的裝置 都會有一個情緒晶片，

and we won't remember what it was like when we couldn't just frown at our device

我們就會忘記當年 裝置還不會回應我們皺眉的時候說出：

and our device would say, "Hmm, you didn't like that, did you?"

「嗯，你不喜歡這個，是吧？」 是什麼樣子。

Our biggest challenge is that there are so many applications of this technology,

我們最大的挑戰是 這種科技有許多應用程式，

my team and I realize that we can't build them all ourselves,

我和我的團隊瞭解 我們不可能只靠自己發展全部，

so we've made this technology available so that other developers

所以我們開放這項科技 讓其他開發者

can get building and get creative.

能繼續開發並激發創意。

We recognize that there are potential risks

我們知道會有潛在風險，

and potential for abuse,

也可能遭到濫用，

but personally, having spent many years doing this,

但是個人認為， 在花了這麼多年做這個之後，

I believe that the benefits to humanity

我相信這對人類的益處，

from having emotionally intelligent technology

就是開發情緒智能科技的益處，

far outweigh the potential for misuse.

遠超過誤用的潛在危險。

And I invite you all to be part of the conversation.

我請大家口耳相傳。

The more people who know about this technology,

愈多人知道這項科技，

the more we can all have a voice in how it's being used.

我們就愈能發聲說明 這該如何使用。

So as more and more of our lives become digital,

隨著我們的生活愈來愈數位化，

we are fighting a losing battle trying to curb our usage of devices

試圖以遏止使用裝置來重拾情緒

in order to reclaim our emotions.

是一場必敗的仗。

So what I'm trying to do instead is to bring emotions into our technology

與其如此， 我寧可把情感帶進科技，

and make our technologies more responsive.

讓我們的科技更有回應。

So I want those devices that have separated us

所以我想用這些 原本使我們疏遠的裝置，

to bring us back together.

讓我們重新結合在一起。

And by humanizing technology, we have this golden opportunity

藉著把科技人性化， 我們擁有這個黃金時機

to reimagine how we connect with machines,

來重新想像我們如何 與機器連結，

and therefore, how we, as human beings,

進而想像我們身為人類

connect with one another.

如何能重新連結彼此。

Thank you.

謝謝。

(Applause)

（掌聲）