Daphne Koller: What we're learning from online education

Translator: Morton Bast Reviewer: Thu-Huong Ha

譯者: 瑞文Eleven 林Lim 審譯者: Adrienne Lin

Like many of you, I'm one of the lucky people.

就像許多人一樣，我是很幸運的。

I was born to a family where education was pervasive.

我出生在書香世家，

I'm a third-generation PhD, a daughter of two academics.

我家三代出博士， 父母都是高知識份子。

In my childhood, I played around in my father's university lab.

我的童年都是在父親的 大學實驗室玩大的，

So it was taken for granted that I attend some of the best universities,

之後也順理成章的進了頂尖大學，

which in turn opened the door to a world of opportunity.

為我開啓了機會的大門。

Unfortunately, most of the people in the world are not so lucky.

不幸的是，世界上 大部分的人並非如此幸運。

In some parts of the world, for example, South Africa,

在這世界的某些角落，例如南非，

education is just not readily accessible.

教育並非人人可及，

In South Africa, the educational system was constructed

南非的教育系統

是在種族隔離制度的時代 為少數白人建立的。

in the days of apartheid for the white minority.

And as a consequence, today there is just not enough spots

其結果是，現今卻沒有足夠的名額

for the many more people who want and deserve a high quality education.

提供給很多想要也值得 擁有高品質教育的人。

That scarcity led to a crisis in January of this year

今年一月在約翰尼斯堡大學，

at the University of Johannesburg.

情況更趨惡化。

There were a handful of positions left open

有一些正常招生 所剩餘的名額開放註冊。

from the standard admissions process, and the night before

在登記註冊的前一晚，

they were supposed to open that for registration,

thousands of people lined up outside the gate in a line a mile long,

成千上萬人在大門外大排長龍，

hoping to be first in line to get one of those positions.

希望拔得頭籌搶到這些名額。

When the gates opened, there was a stampede,

大門一開，群眾相互踩踏，

and 20 people were injured and one woman died.

20 個人受傷，一名婦女死亡。

She was a mother who gave her life

她是一位為子犧牲的母親，

trying to get her son a chance at a better life.

只為下一代有更好的將來。

But even in parts of the world like the United States

即使在其他國家，像是美國，

where education is available, it might not be within reach.

教育雖然存在，但不一定能得到。

There has been much discussed in the last few years

這幾年有個很熱門的議題，

about the rising cost of health care.

就是不斷飆高的醫療支出。

What might not be quite as obvious to people

但民眾不易查覺的是，

is that during that same period the cost of higher education tuition

同一時期高等教育的費用，

has been increasing at almost twice the rate,

以將近兩倍的速度飆漲，

for a total of 559 percent since 1985.

比 1985 年漲了 5.59 倍。

This makes education unaffordable for many people.

對許多人來說， 高等教育因而變得遙不可及。

Finally, even for those who do manage to get the higher education,

即使設法接受高等教育的人，

the doors of opportunity might not open.

機會之門也沒有為他打開。

Only a little over half of recent college graduates

近期只有略為超過一半的

in the United States who get a higher education

美國高等教育大學畢業生，

actually are working in jobs that require that education.

從事需受高等教育的工作。

This, of course, is not true for the students

對於那些畢業於頂尖學府的學生， 這不成問題，

who graduate from the top institutions,

but for many others, they do not get the value

但對其他的許多人，

他們付出時間和努力後， 沒有獲得適當回報。

for their time and their effort.

Tom Friedman, in his recent New York Times article,

湯姆 • 佛里曼最近 在紐約時報的文章中，

captured, in the way that no one else could, the spirit behind our effort.

以人所未及的方式， 掌握住我們努力的精髓

He said the big breakthroughs are what happen

他指出所謂「重大的突破」，即是：

when what is suddenly possible meets what is desperately necessary.

「當有迫切需求時， 有些事突然變得可能。」

I've talked about what's desperately necessary.

我已經說過什麼是 「迫切需求(普遍的教育)」，

Let's talk about what's suddenly possible.

接著讓我們談談 有那些事「突然變得可能」。

What's suddenly possible was demonstrated by

在史丹佛大學的三堂大型課程，

three big Stanford classes,

展示了什麼是「突然變得可能」，

each of which had an enrollment of 100,000 people or more.

每堂課超過10 萬人修課。

So to understand this, let's look at one of those classes,

要理解這點，讓我們看看其中一堂課，

the Machine Learning class offered by my colleague

我的同事與合夥人 Andrew Ng，

and cofounder Andrew Ng.

所開設的機器學習課 。

Andrew teaches one of the bigger Stanford classes.

機器學習課是 Andrew 在史丹佛大學

It's a Machine Learning class,

學生較多的課程，

and it has 400 people enrolled every time it's offered.

每次有 400 人註冊。

When Andrew taught the Machine Learning class to the general public,

當 Andrew 對一般大眾 線上講授機器學習時，

it had 100,000 people registered.

有 10 萬人註冊。

So to put that number in perspective,

從數字上來看，

for Andrew to reach that same size audience

以 Andrew 在史丹佛大學的課堂規模，

by teaching a Stanford class,

he would have to do that for 250 years.

他得花 250 年 才會教到這麼多學生，

Of course, he'd get really bored.

當然，他也會覺得單調乏味。

So, having seen the impact of this,

了解到這樣的影響力，

Andrew and I decided that we needed to really try and scale this up,

Andrew 與我認為 需要嘗試將規模擴大，

to bring the best quality education to as many people as we could.

儘量給更多的人 帶來最好品質的教育。

So we formed Coursera,

所以我們打造了線上學習網站 Coursera。

whose goal is to take the best courses

其目標是把最好的課程、

from the best instructors at the best universities

由頂尖大學裡的最佳講師來上，

and provide it to everyone around the world for free.

免費提供給在全球的人們。

We currently have 43 courses on the platform

目前我們的平臺上有 43 堂課程，

from four universities across a range of disciplines,

來自四所大學，包含各種學科，

and let me show you a little bit of an overview

讓我給各位看看

of what that looks like.

它會是像什麼樣子：

(Video) Robert Ghrist: Welcome to Calculus.

（影片）Robert Ghrist： 歡迎來到微積分課

Ezekiel Emanuel: Fifty million people are uninsured.

Ezekiel Emanuel： 5000 萬人沒有醫療保險。

Scott Page: Models help us design more effective institutions and policies.

Scott Page：模型幫助我們 設計更有效的體制和政策，

We get unbelievable segregation.

我們獲得令人難以置信的分隔...

Scott Klemmer: So Bush imagined that in the future,

Scott Klemmer： 所以布希想像，在未來

you'd wear a camera right in the center of your head.

你會在頭上戴上相機。

Mitchell Duneier: Mills wants the student of sociology to develop the quality of mind ...

Mitchell Duneier：米爾斯希望 研習社會學的學生增強心理素質...

RG: Hanging cable takes on the form of a hyperbolic cosine.

RG：懸掛的電纜表現出雙曲餘弦函數

Nick Parlante: For each pixel in the image, set the red to zero.

Nick Parlante：圖像中的每個像素， 將紅色設置為零。

Paul Offit: ... Vaccine allowed us to eliminate polio virus.

Paul Offit： ...疫苗讓我們 得以消滅脊髓灰質炎病毒

Dan Jurafsky: Does Lufthansa serve breakfast and San Jose? Well, that sounds funny.

Dan Jurafsky： 漢莎航空 提供早餐和聖荷西嗎？這聽上去很滑稽。

Daphne Koller: So this is which coin you pick, and this is the two tosses.

Daphne Koller：這是你挑的硬幣， 這是兩次拋擲...

Andrew Ng: So in large-scale machine learning, we'd like to come up with computational ...

Andrew Ng：在大規模機器學習， 我們想計算...

(Applause)

（掌聲）

DK: It turns out, maybe not surprisingly,

或許不令人驚訝，事實證明

that students like getting the best content

學生們喜歡從最好的大學

from the best universities for free.

免費得到最好的內容。

Since we opened the website in February,

自今年二月我們開放網站以來，

we now have 640,000 students from 190 countries.

至今我們有 64 萬名學生， 來自 190 個國家。

We have 1.5 million enrollments,

我們有 150 萬人次的登記註冊。

6 million quizzes in the 15 classes that have launched

在已推出的 15 堂課中， 共完成了 600 萬次小測驗。

so far have been submitted, and 14 million videos have been viewed.

影片已有 1400 萬次的點閱。

But it's not just about the numbers,

但不只是數字方面，

it's also about the people.

還有人的方面，

Whether it's Akash, who comes from a small town in India

像是來自印度小鎮的阿喀許，

and would never have access in this case

永遠沒機會也負擔不起

to a Stanford-quality course

接受史丹佛這樣品質的課程。

and would never be able to afford it.

Or Jenny, who is a single mother of two

或是像詹妮，兩個小孩的單身母親，

and wants to hone her skills

想要磨練自己的技能，

so that she can go back and complete her master's degree.

這樣她才可以回去完成她碩士學位。

Or Ryan, who can't go to school,

還有瑞恩，他無法到學校，

because his immune deficient daughter

因為他女兒有免疫缺陷，

can't be risked to have germs come into the house,

為了不能冒險讓細菌進到房子，

so he couldn't leave the house.

所以他不能離家去學校。

I'm really glad to say --

我很高興地宣布，

recently, we've been in correspondence with Ryan --

最近，我們一直在與瑞恩通信，

that this story had a happy ending.

這個故事有一個快樂的結局，

Baby Shannon -- you can see her on the left --

在左邊的就是小寶貝 Shannon ，

is doing much better now,

現在過得更好，

and Ryan got a job by taking some of our courses.

瑞恩藉由修習我們的課程 找到一份工作，

So what made these courses so different?

是什麼讓這些課程如此不同？

After all, online course content has been available for a while.

畢竟，線上課程已經開始了好一陣子

What made it different was that this was real course experience.

使它不同的是，這是真正的課程體驗。

It started on a given day,

從開課開始那天起 ，

and then the students would watch videos on a weekly basis

學生們觀看每週的影片，

and do homework assignments.

並且完成家庭作業。

And these would be real homework assignments

這些是真正的作業，

for a real grade, with a real deadline.

有真正的評分、真正的繳交期限，

You can see the deadlines and the usage graph.

你可以看到繳交期限 和網站使用率圖表，

These are the spikes showing

這些峰值顯示

that procrastination is global phenomenon.

拖延是全球性的現象。

(Laughter)

（笑聲）

At the end of the course,

在課程結束時，

the students got a certificate.

學生們獲得了證書，

They could present that certificate

他們可以提出該證書，

to a prospective employer and get a better job,

給未來的雇主並謀得更好的工作。

and we know many students who did.

而我們知道很多學生做到了，

Some students took their certificate

一些學生拿著他們的證書，

and presented this to an educational institution at which they were enrolled

給他們申請入學的教育機構，

for actual college credit.

做為實際大學學分。

So these students were really getting something meaningful

所以這些學生從投資的時間和精力，

for their investment of time and effort.

獲得了真正有意義的東西。

Let's talk a little bit about some of the components

讓我們再談一些

that go into these courses.

在這些課程中的成份：

The first component is that when you move away

第一個成份是，

from the constraints of a physical classroom

當你消除實體課堂的約束

and design content explicitly for an online format,

並設計明確的線上內容，

you can break away from, for example,

例如，你可以將獨立一小時的講課

the monolithic one-hour lecture.

You can break up the material, for example,

分成為到 8 至 12 分鐘 的模組化單元，

into these short, modular units of eight to 12 minutes,

each of which represents a coherent concept.

其中每個單元都代表一個連貫的概念，

Students can traverse this material in different ways,

學生能用不同的方式學習這些題材，

depending on their background, their skills or their interests.

取決於他們的背景、技能或興趣。

So, for example, some students might benefit

例如，有些學生可以從其他學生

from a little bit of preparatory material

可能已經學會的先備題材中得到益處。

that other students might already have.

Other students might be interested in a particular

有些學生可能會特別感興趣於

enrichment topic that they want to pursue individually.

他們個別想要的豐富主題。

So this format allows us to break away

所以這種模式使我們能夠

突破統一制式的教育模式，

from the one-size-fits-all model of education,

and allows students to follow a much more personalized curriculum.

並允許學生有更多個人化的課程。

Of course, we all know as educators

當然，身為教育工作者，我們都知道

that students don't learn by sitting and passively watching videos.

學生無法光坐著看影片就能學習。

Perhaps one of the biggest components of this effort

也許最需要下工夫努力的部分

is that we need to have students

是讓學生藉練習題

who practice with the material

in order to really understand it.

去真正理解課程。

There's been a range of studies that demonstrate the importance of this.

已有各種研究證明此點的重要性，

This one that appeared in Science last year, for example,

例如，去年"科學"期刊中有一篇文章，

demonstrates that even simple retrieval practice,

談到即使只是簡單的復習，

where students are just supposed to repeat

只需要學生回顧他們剛學過的東西，

what they already learned

gives considerably improved results

在之後的成就測驗，

on various achievement tests down the line

都會有顯著的進步。

than many other educational interventions.

分數比其他的教育干預行動都多。

We've tried to build in retrieval practice into the platform,

我們嘗試在平臺建立回顧性的練習，

as well as other forms of practice in many ways.

還有其他形式的題目。

For example, even our videos are not just videos.

例如，我們的影片不光是教學，

Every few minutes, the video pauses

每隔幾分鐘，影片會暫停並對學生提問：

and the students get asked a question.

(Video) SP: ... These four things. Prospect theory, hyperbolic discounting,

（影片）SP： ...這四件事。 前景理論、雙曲貼現、

status quo bias, base rate bias. They're all well documented.

現狀偏誤、基準利率偏差。 這些都被完整的記錄下來，

So they're all well documented deviations from rational behavior.

因此它們都是有案可稽的理性行為偏差。

DK: So here the video pauses,

在此影片會暫停，

and the student types in the answer into the box

在方框中打上答案後提交。

and submits. Obviously they weren't paying attention.

很明顯他們沒有專心。

(Laughter)

（笑聲）

So they get to try again,

所以他們得再試一次，

and this time they got it right.

這次就答對了。

There's an optional explanation if they want.

這裡有題目的解說 供有需求的人參考。

And now the video moves on to the next part of the lecture.

接著來影片才會接續下去。

This is a kind of simple question

這類題目是我在課堂上 會提到的簡單問題，

that I as an instructor might ask in class,

but when I ask that kind of a question in class,

但是當我提出這種問題時，

80 percent of the students

80% 的學生會努力 琢磨我前面所教的內容 ，

are still scribbling the last thing I said,

15 percent are zoned out on Facebook,

15% 的學生則沉浸在 Facebook 裡，

and then there's the smarty pants in the front row

然後坐在前排的聰明蛋 ,

who blurts out the answer

在其他人還沒機會思考之前， 一口氣就報出正確答案.

before anyone else has had a chance to think about it,

and I as the instructor am terribly gratified

為師的我當然很得意，

that somebody actually knew the answer.

竟然有人知道答案 ，

And so the lecture moves on before, really,

所以甚至在大多數學生 尚未注意到了我提出的問題之前，

most of the students have even noticed that a question had been asked.

我就繼續往下講了。

Here, every single student

但是，這裡每個問題

has to engage with the material.

應該是每個學生都要思考回答的。

And of course these simple retrieval questions

當然不光是這些簡單的回顧性問題，

are not the end of the story.

One needs to build in much more meaningful practice questions,

一個課程需要有更多有用的練習題，

and one also needs to provide the students with feedback

也需要針對學生的回答提供回饋 。

on those questions.

Now, how do you grade the work of 100,000 students

問題是，如果你沒有足夠的助教，

if you do not have 10,000 TAs?

該如何批改十萬名學生的作業？

The answer is, you need to use technology

答案是，善用技術。

to do it for you.

Now, fortunately, technology has come a long way,

幸運的是，這類技術已經很成熟了，

and we can now grade a range of interesting types of homework.

我們現在可以批改各式各樣的作業，

In addition to multiple choice

不光是選擇題，

and the kinds of short answer questions that you saw in the video,

和你在影片中所見的簡答題，

we can also grade math, mathematical expressions

我們可以批改數學、 運算式及數學推導。

as well as mathematical derivations.

We can grade models, whether it's

我們還可以批改模型，

無論是在商業課程的財務模型，

financial models in a business class

or physical models in a science or engineering class

或在科學或工程課中的物理模型。

and we can grade some pretty sophisticated programming assignments.

我們還可以批改一些 相當複雜的程式設計作業。

Let me show you one that's actually pretty simple

讓我展示一個其實很簡單 、

but fairly visual.

但相當視覺化的例子，

This is from Stanford's Computer Science 101 class,

這是來自史丹佛大學的 電腦科學基礎課，

and the students are supposed to color-correct

學生須校正這張

that blurry red image.

暗紅色圖片的顏色。

They're typing their program into the browser,

他們在瀏覽器中鍵入程式，

and you can see they didn't get it quite right, Lady Liberty is still seasick.

你可以看到顏色不對， 自由女神像在暈船，

And so, the student tries again, and now they got it right, and they're told that,

所以，現在他們搞對了，

於是學生被告知可以繼續下一個任務。

and they can move on to the next assignment.

This ability to interact actively with the material

與教材即時互動，

and be told when you're right or wrong

並被告知回答正確或錯誤，

is really essential to student learning.

對學生的學習至關重要。

Now, of course we cannot yet grade

不過，當然我們還無法批改

the range of work that one needs for all courses.

所有種類的作業。

Specifically, what's lacking is the kind of critical thinking work

具體而言，我們缺少 批改批判性思維作業的能力，

that is so essential in such disciplines

這種思維在人文社會科學、

as the humanities, the social sciences, business and others.

商業和其他領域都極其重要。

So we tried to convince, for example,

所以我們嘗試說服，例如

some of our humanities faculty

人文科學教師，

that multiple choice was not such a bad strategy.

選擇題並不是那麼差勁的題型。

That didn't go over really well.

不過效果不佳。

So we had to come up with a different solution.

所以我們不得不 提出另一種解決方案。

And the solution we ended up using is peer grading.

我們最終的解決方案是同儕互評。

It turns out that previous studies show,

比如說 Saddler 和 Good

like this one by Saddler and Good,

之前所做的研究顯示，

that peer grading is a surprisingly effective strategy

同儕互評是十分有效的策略，

for providing reproducible grades.

結果與教師評分雷同。

It was tried only in small classes,

雖然這只曾經在小規模課堂中用過，

but there it showed, for example,

不過這裡顯示了

that these student-assigned grades on the y-axis

學生互評的分數在 Y 軸，

are actually very well correlated

老師批改的分數在 X 軸，

with the teacher-assigned grade on the x-axis.

結果相當吻合。

What's even more surprising is that self-grades,

更神的是自我評分，

where the students grade their own work critically --

就是讓學生給自己打分數，

so long as you incentivize them properly

只要你稍微激勵他們，

so they can't give themselves a perfect score --

確保他們不會 隨隨便便就給自己打滿分，

are actually even better correlated with the teacher grades.

自評結果與老師批改的成績

甚至比互評還要吻合。

And so this is an effective strategy

因此需大規模評分時， 這是很有效的評分策略。

that can be used for grading at scale,

and is also a useful learning strategy for the students,

對於學生來說， 也是一個有用的學習策略，

because they actually learn from the experience.

因為他們能從中學到了東西。

So we now have the largest peer-grading pipeline ever devised,

因此我們有現今最大規模的互評系統，

where tens of thousands of students

成千上萬的學生互相批改作業，

are grading each other's work,

and quite successfully, I have to say.

而且我必須說，這很成功。

But this is not just about students

學生不單只是

sitting alone in their living room working through problems.

獨自坐在客廳解答問題。

Around each one of our courses,

我們的每個課程

a community of students had formed,

都有學生社群，

a global community of people

共同為知識努力的全球性社群。

around a shared intellectual endeavor.

What you see here is a self-generated map

這是一張系統自動產生的地圖，

from students in our Princeton Sociology 101 course,

你看到的是普林斯頓大學

where they have put themselves on a world map,

社會學入門課程學生的來源，

and you can really see the global reach of this kind of effort.

參與修課的人遍及全球。

Students collaborated in these courses in a variety of different ways.

學生在課程中 以不同的方式進行合作，

First of all, there was a question and answer forum,

首先，我們有一個問答平臺，

where students would pose questions,

學生可以提出問題，

and other students would answer those questions.

然後其他學生可以回答。

And the really amazing thing is,

最令人驚奇的是，

because there were so many students,

因為學生太多了，

it means that even if a student posed a question

這意味著即使有人在淩晨三點

at 3 o'clock in the morning,

提出一個問題，

somewhere around the world,

世界的某個角落，

there would be somebody who was awake

總有人會醒著，

and working on the same problem.

並正研究同一問題。

And so, in many of our courses,

因此，在許多課程中，

the median response time for a question

問答平臺上提問與被解答的

on the question and answer forum was 22 minutes.

間隔時間中間值只有22分鐘，

Which is not a level of service I have ever offered to my Stanford students.

這比我給史丹佛學生 的服務要好得多。

(Laughter)

（笑聲）

And you can see from the student testimonials

你可以從學生感言看到，

that students actually find

他們發現，

that because of this large online community,

在這個大型的線上社群，

they got to interact with each other in many ways

可用許多方式互動，

that were deeper than they did in the context of the physical classroom.

比在實體教室上課更密切。

Students also self-assembled,

學生還會自組學習小組，

without any kind of intervention from us,

我們從不干預。

into small study groups.

Some of these were physical study groups

其中有些是實體小組，

along geographical constraints

受限於地域，

and met on a weekly basis to work through problem sets.

他們每週聚會一次， 主要是進行解題。

This is the San Francisco study group,

這是舊金山學習小組，

but there were ones all over the world.

成員來自世界各地。

Others were virtual study groups,

還有人組成虛擬學習小組，

sometimes along language lines or along cultural lines,

有時會依語言或文化分組，

and on the bottom left there,

左下角你可以看到

you see our multicultural universal study group

我們的多元文化學習小組，

where people explicitly wanted to connect

成員們熱切地想要

with people from other cultures.

與其他文化的人交流。

There are some tremendous opportunities

這種架構帶來

to be had from this kind of framework.

絕佳的機會

The first is that it has the potential of giving us

首先，讓我們極有可能地

a completely unprecedented look

能用前所未有的視野

into understanding human learning.

來瞭解人類的學習方式。

Because the data that we can collect here is unique.

因為我們這裡 能收集到獨一無二的資料，

You can collect every click, every homework submission,

你可以收集到每一次 點擊與提交的作業，

every forum post from tens of thousands of students.

及成千上萬學生的貼文。

So you can turn the study of human learning

因此，你可以將學習研究的模式

from the hypothesis-driven mode

由假設驅動模式 (先提出假設，再設計實驗驗證)

to the data-driven mode, a transformation that,

轉變為以資料驅動 (用線上系統收集數據來研究)，

for example, has revolutionized biology.

這類的轉變已經 徹底改變了生物科學。

You can use these data to understand fundamental questions

你可以透過這些資料 去瞭解最基本的問題，

like, what are good learning strategies

比如說，哪些是好的學習策略？

that are effective versus ones that are not?

哪些是有效？哪些是無效的策略?

And in the context of particular courses,

而在特定的幾個課程背景中，

you can ask questions

你能提出像這樣的問題：

like, what are some of the misconceptions that are more common

有哪些是常見的錯誤概念，

and how do we help students fix them?

我們該如何幫助學生修正它們？

So here's an example of that,

這裡有一個例子，

also from Andrew's Machine Learning class.

一樣來自 Andrew 的機器學習課。

This is a distribution of wrong answers

這是 Andrew 一題作業的

to one of Andrew's assignments.

錯誤答案分佈圖。

The answers happen to be pairs of numbers,

答案剛好是成對的數字，

so you can draw them on this two-dimensional plot.

因此，你可以繪製它們的二維圖像。

Each of the little crosses that you see is a different wrong answer.

你看到的每個小叉叉 是不同的錯誤答案。

The big cross at the top left

左上角的大叉叉，

is where 2,000 students

顯示有 2,000 個學生 都寫了相同的錯誤答案。

gave the exact same wrong answer.

Now, if two students in a class of 100

如果在 100 人的課堂中

give the same wrong answer,

有兩個學生給出了 同樣的錯誤答案 ，

you would never notice.

你肯定不會發現，

But when 2,000 students give the same wrong answer,

不過當2000人都寫了 同樣的錯誤答案時，

it's kind of hard to miss.

你勢必會注意到。

So Andrew and his students went in,

因此 Andrew 和他的學生

looked at some of those assignments,

分析錯誤的答案，

understood the root cause of the misconception,

找到了錯誤概念的根源，

and then they produced a targeted error message

然後他們創建了一則特別資訊，

that would be provided to every student

專門提供給答錯的同學，

whose answer fell into that bucket,

which means that students who made that same mistake

這意味著那些寫出 同樣錯誤答案的學生，

would now get personalized feedback

都可以得到個別化的回饋，

telling them how to fix their misconception much more effectively.

告訴他們如何更有效地修正迷失概念。

So this personalization is something that one can then build

這種個別化的東西，

by having the virtue of large numbers.

是利用大規模數據的優勢所創建。

Personalization is perhaps

個別化也許是

one of the biggest opportunities here as well,

幫助我們解決

because it provides us with the potential

一個 30 年未決問題的契機。

of solving a 30-year-old problem.

Educational researcher Benjamin Bloom, in 1984,

教育學者 Benjamin Bloom 在 1984 年

posed what's called the 2 sigma problem,

提出了一個稱為 2 個標準差(sigma)問題。

which he observed by studying three populations.

他研究三種學習類型的學生。

The first is the population that studied in a lecture-based classroom.

第一種在課堂中 以講述法為主來學習，

The second is a population of students that studied

第二種學生也是以講述法學習，

using a standard lecture-based classroom,

but with a mastery-based approach,

但是有一個精熟的門檻，

so the students couldn't move on to the next topic

學生在繼續學習下一個主題之前，

before demonstrating mastery of the previous one.

必須熟練掌握前一個主題。

And finally, there was a population of students

最後一種就是

that were taught in a one-on-one instruction using a tutor.

以一對一方式教學的學生。

The mastery-based population was a full standard deviation,

結果在成就測驗分數上，

or sigma, in achievement scores better

第二類學生比 單以講述法受教的第一類學生，

than the standard lecture-based class,

高出 1 個標準差，

and the individual tutoring gives you 2 sigma

而接受一對一教學的學生比第一類學生

improvement in performance.

成績高出 2 個標準差。

To understand what that means,

要理解這其中的意義，

let's look at the lecture-based classroom,

讓我們看看以講述為主的課堂吧。

and let's pick the median performance as a threshold.

我們把中間值當做門檻，

So in a lecture-based class,

因此在以講述法為主的課堂中，

half the students are above that level and half are below.

是一半對一半，

In the individual tutoring instruction,

當教師一對一授課時，

98 percent of the students are going to be above that threshold.

98% 的學生都高於這個門檻。

Imagine if we could teach so that 98 percent of our students

想像如果我們教出的學生

能有 98% 在平均標準之上，

would be above average.

Hence, the 2 sigma problem.

這就帶到 2 個標準差的問題。

Because we cannot afford, as a society,

因為我們的社會不可能

to provide every student with an individual human tutor.

提供每位學生一對一的教學，

But maybe we can afford to provide each student

但是我們可以每人給一部電腦

with a computer or a smartphone.

或者一台智慧型手機。

So the question is, how can we use technology

因此問題是， 我們如何利用科技去推動

to push from the left side of the graph, from the blue curve,

我們擺脫左邊藍色曲線的圖像，

to the right side with the green curve?

最後到達右邊綠色曲線的圖像呢？

Mastery is easy to achieve using a computer,

透過電腦來達到精熟學習是很簡單的，

because a computer doesn't get tired

因為電腦就算播放同樣的影片 5 遍，

of showing you the same video five times.

也從來不嫌累，

And it doesn't even get tired of grading the same work multiple times,

而且多次批改同樣的作業也不會煩。

we've seen that in many of the examples that I've shown you.

在很多我舉的例子當中， 我們也看到了。

And even personalization

就連個別化的設計

is something that we're starting to see the beginnings of,

也正開始逐漸發展。

whether it's via the personalized trajectory through the curriculum

不管是在課程中個別化的安排，

or some of the personalized feedback that we've shown you.

或是一些我們展示 給大家看的個別化的回饋。

So the goal here is to try and push,

因此我們的目標是嘗試看看，

and see how far we can get towards the green curve.

能夠多接近那個綠色的圖像。

So, if this is so great, are universities now obsolete?

好吧，如果線上學習系統這麼好， 大學是不是過時了？

Well, Mark Twain certainly thought so.

馬克吐溫顯然是如此認為，

He said that, "College is a place where a professor's lecture notes

他說：「大學是一個教授的筆記

go straight to the students' lecture notes,

直接成為學生的筆記，

without passing through the brains of either."

並且都沒有經過兩者大腦的地方。」

(Laughter)

（笑聲）

I beg to differ with Mark Twain, though.

不過我不敢與馬克吐溫苟同。

I think what he was complaining about is not

我認為他在抱怨的不是大學，

universities but rather the lecture-based format

而是以講述法為主的教學模式，

that so many universities spend so much time on.

因為這種模式在大學太普遍了。

So let's go back even further, to Plutarch,

那讓我們再往回看看， 羅馬時代的希臘作家 Plutarch 說：

who said that, "The mind is not a vessel that needs filling,

「心智不是一個需要填充的容器，

but wood that needs igniting."

而是需要點燃的木頭。」

And maybe we should spend less time at universities

因此或許我們應該

filling our students' minds with content

少花點時間填充學生的腦袋，

by lecturing at them, and more time igniting their creativity,

多花些時間激發他們的創造力、

their imagination and their problem-solving skills

想像力和解決問題的能力，

by actually talking with them.

並透過積極與他們交流來實現。

So how do we do that?

那我們怎麼做呢？

We do that by doing active learning in the classroom.

我們透過主動學習來實現。

So there's been many studies, including this one,

有很多研究表明，包括這個，

that show that if you use active learning,

如果你使用積極學習法，

interacting with your students in the classroom,

與你的學生在課堂上互動，

performance improves on every single metric --

他們在每個項目的表現 都會不斷提升，

on attendance, on engagement and on learning

在出勤率、參與度

及標準化測驗的成績。

as measured by a standardized test.

You can see, for example, that the achievement score

比如說，你可以看到學習成就分數

almost doubles in this particular experiment.

在這個實驗中幾乎都翻倍了。

So maybe this is how we should spend our time at universities.

或許這就是我們 該度過大學時光的方式。

So to summarize, if we could offer a top quality education

最後總結一下，如果我們能免費

向世上的每個人提供高品質的教育，

to everyone around the world for free,

what would that do? Three things.

結果會怎麼樣呢？會是這樣：

First it would establish education as a fundamental human right,

首先，教育變成了人類最基本的權利。

where anyone around the world

世界上每個有能力和積極性的人，

with the ability and the motivation

could get the skills that they need

都可以得到他們需要的技能，

to make a better life for themselves,

為他們自己、家人 及社區追求更好的生活。

their families and their communities.

Second, it would enable lifelong learning.

第二、實現終生學習的理念。

It's a shame that for so many people,

很可惜的是，許多人

learning stops when we finish high school or when we finish college.

在高中、大學念完後 就再也不學習了。

By having this amazing content be available,

有了內容如此豐富的系統，

we would be able to learn something new

每當我們想要的時候，

我們可以學一些新知識，

every time we wanted,

whether it's just to expand our minds

不管是用來拓展思維，

or it's to change our lives.

或者是要改變我們的生活。

And finally, this would enable a wave of innovation,

最後，這個系統會帶來一股創新潮，

because amazing talent can be found anywhere.

因為到處都可能 找到有驚人天賦的人，

Maybe the next Albert Einstein or the next Steve Jobs

或許下一個愛因斯坦或賈伯斯

is living somewhere in a remote village in Africa.

就住在一個偏遠的非洲小鎮上。

And if we could offer that person an education,

如果我們能為那個人提供教育，

they would be able to come up with the next big idea

他們可能會想出下一個極好的主意，

and make the world a better place for all of us.

讓我們的世界變得更美好。

Thank you very much.

謝謝！

(Applause)

（掌聲）