TED2015
Fei-Fei Li: How we're teaching computers to understand pictures
李飛飛: 我們如何教導電腦看懂圖像
Filmed:
Readability: 4.5
2,702,344 views
當孩童看著一個圖像，她能辨識出簡單的東西，例如貓、書、椅子等。如今，電腦也慢慢聰明得可以做到一樣的事情。那麼接下來呢？在這場動人的演說中，電腦視覺專家李飛飛與我們分享最先進的成果，包含了她的研究小組所建造，用來「教導」電腦識別的一千五百萬幅圖像，以及對未來世界的想像。
Fei-Fei Li - Computer scientist
As Director of Stanford’s Artificial Intelligence Lab and Vision Lab, Fei-Fei Li is working to solve AI’s trickiest problems -- including image recognition, learning and language processing. Full bio
As Director of Stanford’s Artificial Intelligence Lab and Vision Lab, Fei-Fei Li is working to solve AI’s trickiest problems -- including image recognition, learning and language processing. Full bio
Double-click the English transcript below to play the video.
00:14
Let me show you something.
0
2366
3738
容我為各位呈現一些照片
00:18
(Video) Girl: Okay, that's a cat
sitting in a bed.
sitting in a bed.
1
6104
4156
(影片)女孩：嗯，這是一隻貓，坐在床上。
00:22
The boy is petting the elephant.
2
10260
4040
這男孩在拍撫一隻象。
00:26
Those are people
that are going on an airplane.
that are going on an airplane.
3
14300
4354
這些人要去搭飛機。
00:30
That's a big airplane.
4
18654
2810
好大的飛機。
主講人：這是由一位三歲的小孩
00:33
Fei-Fei Li: This is
a three-year-old child
a three-year-old child
5
21464
2206
00:35
describing what she sees
in a series of photos.
in a series of photos.
6
23670
3679
所描述她看到的一系列照片
00:39
She might still have a lot
to learn about this world,
to learn about this world,
7
27349
2845
雖然對於這世界她還有更多要學習的地方，
00:42
but she's already an expert
at one very important task:
at one very important task:
8
30194
4549
但是她已經是其中一項重要技能的專家--
00:46
to make sense of what she sees.
9
34743
2846
為所見之聞賦予意義。
科技在我們的社會已進展到前所未有的程度：
00:50
Our society is more
technologically advanced than ever.
technologically advanced than ever.
10
38229
4226
00:54
We send people to the moon,
we make phones that talk to us
we make phones that talk to us
11
42455
3629
我們把人送上月球、發明可以與人交談的電話，
00:58
or customize radio stations
that can play only music we like.
that can play only music we like.
12
46084
4946
或是客製一個電台，只播放個人喜歡的音樂。
01:03
Yet, our most advanced
machines and computers
machines and computers
13
51030
4055
然而這台無比聰明的機器和電腦
01:07
still struggle at this task.
14
55085
2903
仍然無法發展這項技能，
01:09
So I'm here today
to give you a progress report
to give you a progress report
15
57988
3459
因此今天我來到這裡向各位報告
01:13
on the latest advances
in our research in computer vision,
in our research in computer vision,
16
61447
4047
我們在電腦視覺的最新研究進展，
01:17
one of the most frontier
and potentially revolutionary
and potentially revolutionary
17
65494
4161
這是現階段在資訊業領域中，
最先進、最具潛力的革命性技術。
01:21
technologies in computer science.
18
69655
3206
01:24
Yes, we have prototyped cars
that can drive by themselves,
that can drive by themselves,
19
72861
4551
是的，目前我們已經有自動駕駛的原型車，
01:29
but without smart vision,
they cannot really tell the difference
they cannot really tell the difference
20
77412
3853
但若不具備視覺辨識技術，
它將無法分辨同樣出現在馬路中，
它將無法分辨同樣出現在馬路中，
01:33
between a crumpled paper bag
on the road, which can be run over,
on the road, which can be run over,
21
81265
3970
一團它其實輾過也無妨的破紙袋，
01:37
and a rock that size,
which should be avoided.
which should be avoided.
22
85235
3340
以及一個大到它必須閃避的石塊，
兩者有何不同。
兩者有何不同。
我們製造出畫素極高的相機，
01:41
We have made fabulous megapixel cameras,
23
89415
3390
01:44
but we have not delivered
sight to the blind.
sight to the blind.
24
92805
3135
但我們卻無法賦予盲人視覺；
無人機可以翻山越嶺，
01:48
Drones can fly over massive land,
25
96420
3305
01:51
but don't have enough vision technology
26
99725
2134
卻沒有足夠的視覺技術可以
讓我們追蹤雨林的變化；
01:53
to help us to track
the changes of the rainforests.
the changes of the rainforests.
27
101859
3461
01:57
Security cameras are everywhere,
28
105320
2950
監視器滿佈在各個角落，
02:00
but they do not alert us when a child
is drowning in a swimming pool.
is drowning in a swimming pool.
29
108270
5067
卻無法在看到一個孩子將溺斃在泳池之際，
對我們發出警訊。
對我們發出警訊。
靜態及動態影像已逐漸與全世界的生活密不可分，
02:06
Photos and videos are becoming
an integral part of global life.
an integral part of global life.
30
114167
5595
02:11
They're being generated at a pace
that's far beyond what any human,
that's far beyond what any human,
31
119762
4087
它們發展的步伐已經遠遠超越人類
及其群體所相信的，
02:15
or teams of humans, could hope to view,
32
123849
2783
在座各位以及我自己
都是TED這個活動裡頭的推手。
都是TED這個活動裡頭的推手。
02:18
and you and I are contributing
to that at this TED.
to that at this TED.
33
126632
3921
02:22
Yet our most advanced software
is still struggling at understanding
is still struggling at understanding
34
130553
5232
然而，目前最先進的軟體卻仍在其中苦苦掙扎，
無法理解與應用這龐大的資料體。
02:27
and managing this enormous content.
35
135785
3876
02:31
So in other words,
collectively as a society,
collectively as a society,
36
139661
5272
換而言之，在這整個社會裡，
大家都有如盲人在運作，
02:36
we're very much blind,
37
144933
1746
02:38
because our smartest
machines are still blind.
machines are still blind.
38
146679
3387
因為連我們最聰明的機器都還看不見。
02:43
"Why is this so hard?" you may ask.
39
151526
2926
或許有人會問：這到底有什麼困難？
02:46
Cameras can take pictures like this one
40
154452
2693
任何相機都可以產生像這樣的照片，
02:49
by converting lights into
a two-dimensional array of numbers
a two-dimensional array of numbers
41
157145
3994
它是藉由將有色光轉換成2D的數字陣列，
02:53
known as pixels,
42
161139
1650
也就是大家熟知的像素。
02:54
but these are just lifeless numbers.
43
162789
2251
但這些數字是死的，
02:57
They do not carry meaning in themselves.
44
165040
3111
並沒有被賦予意義。
03:00
Just like to hear is not
the same as to listen,
the same as to listen,
45
168151
4343
就好像有「聽」，不代表有「到」。
03:04
to take pictures is not
the same as to see,
the same as to see,
46
172494
4040
同樣地，攝取到影像不等於看見，
03:08
and by seeing,
we really mean understanding.
we really mean understanding.
47
176534
3829
我們所認知的看到，應包含著了解其中的意義。
03:13
In fact, it took Mother Nature
540 million years of hard work
540 million years of hard work
48
181293
6177
事實上，這樣的成果，
是大自然花了五億四千萬年的光陰
是大自然花了五億四千萬年的光陰
03:19
to do this task,
49
187470
1973
才得到的。
03:21
and much of that effort
50
189443
1881
這其中的努力，
03:23
went into developing the visual
processing apparatus of our brains,
processing apparatus of our brains,
51
191324
5271
泰半是耗費在發展腦部的視覺處理這個區塊，
03:28
not the eyes themselves.
52
196595
2647
而不是眼睛的部分。
03:31
So vision begins with the eyes,
53
199242
2747
也就是說，視覺始於眼睛，
03:33
but it truly takes place in the brain.
54
201989
3518
但真正使它有用的，卻是大腦。
03:38
So for 15 years now, starting
from my Ph.D. at Caltech
from my Ph.D. at Caltech
55
206287
5060
十五年來，從在加州理工學院攻讀博士開始，
03:43
and then leading Stanford's Vision Lab,
56
211347
2926
到領導史丹佛的視覺實驗室，
03:46
I've been working with my mentors,
collaborators and students
collaborators and students
57
214273
4396
我和指導教授、同事及學生們，
03:50
to teach computers to see.
58
218669
2889
試圖讓電腦擁有智能之眼，
03:54
Our research field is called
computer vision and machine learning.
computer vision and machine learning.
59
222658
3294
我們研究的領域稱之為電腦視覺與機器學習，
03:57
It's part of the general field
of artificial intelligence.
of artificial intelligence.
60
225952
3878
這是人工智慧其中一環。
04:03
So ultimately, we want to teach
the machines to see just like we do:
the machines to see just like we do:
61
231000
5493
我們的終極目標就是教導機器能夠像人一樣理解所見之物，
04:08
naming objects, identifying people,
inferring 3D geometry of things,
inferring 3D geometry of things,
62
236493
5387
像是識別物品、辨認人臉、
推論物體的幾何形態，
推論物體的幾何形態，
04:13
understanding relations, emotions,
actions and intentions.
actions and intentions.
63
241880
5688
進而理解其中的關聯、情緒、動作及意圖。
04:19
You and I weave together entire stories
of people, places and things
of people, places and things
64
247568
6153
在座每一位和我，都可以在匆匆一瞥的瞬間，
理解到人事、地、物所交織而成的網絡，
04:25
the moment we lay our gaze on them.
65
253721
2164
04:28
The first step towards this goal
is to teach a computer to see objects,
is to teach a computer to see objects,
66
256955
5583
要電腦達成這個目標的第一步，就是教導它辨別物品，
04:34
the building block of the visual world.
67
262538
3368
這是視覺的基石。
04:37
In its simplest terms,
imagine this teaching process
imagine this teaching process
68
265906
4434
簡單來說，我們教導的方法就是
04:42
as showing the computers
some training images
some training images
69
270340
2995
給電腦看一些特定物體的影像，
04:45
of a particular object, let's say cats,
70
273335
3321
例如貓咪。
04:48
and designing a model that learns
from these training images.
from these training images.
71
276656
4737
我們設計了一個程式讓電腦利用這些影像來學習
04:53
How hard can this be?
72
281393
2044
這有啥困難？
04:55
After all, a cat is just
a collection of shapes and colors,
a collection of shapes and colors,
73
283437
4052
貓咪不就是由一些幾何圖形和顏色所組成的嘛，
04:59
and this is what we did
in the early days of object modeling.
in the early days of object modeling.
74
287489
4086
這就是我們初期所做的物體模型。
05:03
We'd tell the computer algorithm
in a mathematical language
in a mathematical language
75
291575
3622
我們用數學語言來告知電腦演繹方法，
05:07
that a cat has a round face,
a chubby body,
a chubby body,
76
295197
3343
貓就是有圓圓的臉、胖胖的身體，
05:10
two pointy ears, and a long tail,
77
298540
2299
兩個尖尖的耳朵和一條長尾巴。
05:12
and that looked all fine.
78
300839
1410
看起來很好啊，
05:14
But what about this cat?
79
302859
2113
但如果貓咪長這樣呢？
05:16
(Laughter)
80
304972
1091
(觀眾笑)
05:18
It's all curled up.
81
306063
1626
全身都捲起來了。
05:19
Now you have to add another shape
and viewpoint to the object model.
and viewpoint to the object model.
82
307689
4719
這下子我們又得在原來的模型
加上新的形狀和不同的視野角度。
加上新的形狀和不同的視野角度。
05:24
But what if cats are hidden?
83
312408
1715
又，如果貓咪是躲著的呢？
05:27
What about these silly cats?
84
315143
2219
像這群傻貓？
05:31
Now you get my point.
85
319112
2417
這樣各位了解我的意思嗎？
05:33
Even something as simple
as a household pet
as a household pet
86
321529
3367
即使簡單如貓這樣的家庭寵物，
05:36
can present an infinite number
of variations to the object model,
of variations to the object model,
87
324896
4504
也會有相對於原型以外，無數的其他形態表徵，
05:41
and that's just one object.
88
329400
2233
而這只是其中一樣。
05:44
So about eight years ago,
89
332573
2492
因此八年前，
05:47
a very simple and profound observation
changed my thinking.
changed my thinking.
90
335065
5030
一項極其簡單和深刻的觀察，改變了我的想法，
05:53
No one tells a child how to see,
91
341425
2685
沒有人教導孩子如何去「看」，
05:56
especially in the early years.
92
344110
2261
特別是在早期發育階段，
05:58
They learn this through
real-world experiences and examples.
real-world experiences and examples.
93
346371
5000
他們是從真實世界的經驗中學習。
06:03
If you consider a child's eyes
94
351371
2740
如果你把孩童的眼睛
06:06
as a pair of biological cameras,
95
354111
2554
當成生物相機的概念，
06:08
they take one picture
about every 200 milliseconds,
about every 200 milliseconds,
96
356665
4180
就如同每200毫秒就拍一張照片一樣，
06:12
the average time an eye movement is made.
97
360845
3134
這是眼球移動的平均時間。
06:15
So by age three, a child would have seen
hundreds of millions of pictures
hundreds of millions of pictures
98
363979
5550
年紀到了三歲時，
孩子們已經看過了真實世界中
孩子們已經看過了真實世界中
數以百萬計的照片，
06:21
of the real world.
99
369529
1834
06:23
That's a lot of training examples.
100
371363
2280
這樣的訓練範例是很大量的。
06:26
So instead of focusing solely
on better and better algorithms,
on better and better algorithms,
101
374383
5989
因此，我的直覺告訴我
應該以孩童的學習經驗法則，
應該以孩童的學習經驗法則，
06:32
my insight was to give the algorithms
the kind of training data
the kind of training data
102
380372
5272
並兼以質與量，
提供訓練的資料給電腦，
06:37
that a child was given through experiences
103
385644
3319
06:40
in both quantity and quality.
104
388963
3878
而非一昧追求更好的程式演算。
06:44
Once we know this,
105
392841
1858
有了上述的洞見，
06:46
we knew we needed to collect a data set
106
394699
2971
我們接下來必須要收集
06:49
that has far more images
than we have ever had before,
than we have ever had before,
107
397670
4459
前所未有的大量資料群，
06:54
perhaps thousands of times more,
108
402129
2577
甚至於是千倍以上的。
06:56
and together with Professor
Kai Li at Princeton University,
Kai Li at Princeton University,
109
404706
4111
於是我與普林斯頓大學的李凱教授
07:00
we launched the ImageNet project in 2007.
110
408817
4752
共同於2007年開始了
我們稱之為 ImageNet 的專案。
我們稱之為 ImageNet 的專案。
07:05
Luckily, we didn't have to mount
a camera on our head
a camera on our head
111
413569
3838
很幸運地，我們不必在頭上綁一個相機，
07:09
and wait for many years.
112
417407
1764
然後花費數年收集影像，
07:11
We went to the Internet,
113
419171
1463
而是轉而由網際網路，
07:12
the biggest treasure trove of pictures
that humans have ever created.
that humans have ever created.
114
420634
4436
這個由人類所創造出來 龐大的影像寶窟，
07:17
We downloaded nearly a billion images
115
425070
3041
我們下載了數以百萬計的影像，
07:20
and used crowdsourcing technology
like the Amazon Mechanical Turk platform
like the Amazon Mechanical Turk platform
116
428111
5880
並且使用如Amazon Mechanical Turk
這樣的群眾外包平台，
這樣的群眾外包平台，
07:25
to help us to label these images.
117
433991
2339
來協助我們處理及分類這些照片。
07:28
At its peak, ImageNet was one of
the biggest employers
the biggest employers
118
436330
4900
在高峰期，ImageNet 甚至是整個亞馬遜平台
07:33
of the Amazon Mechanical Turk workers:
119
441230
2996
最大的雇主之一，
07:36
together, almost 50,000 workers
120
444226
3854
我們一共聘請了來自167個國家，
07:40
from 167 countries around the world
121
448080
4040
約5萬個工作者，
07:44
helped us to clean, sort and label
122
452120
3947
來協助我們分類處理並標示
07:48
nearly a billion candidate images.
123
456067
3575
將近10億幅影像，
07:52
That was how much effort it took
124
460612
2653
花費了這麼多的資源，
07:55
to capture even a fraction
of the imagery
of the imagery
125
463265
3900
就是為了捕捉那一絲絲
07:59
a child's mind takes in
in the early developmental years.
in the early developmental years.
126
467165
4171
孩童在早期心智發展的浮光掠影。
08:04
In hindsight, this idea of using big data
127
472148
3902
用現在眼光看來，使用大量的資料
08:08
to train computer algorithms
may seem obvious now,
may seem obvious now,
128
476050
4550
來訓練電腦演算是明顯合理的，
08:12
but back in 2007, it was not so obvious.
129
480600
4110
然而在2007年的世界卻非如此。
08:16
We were fairly alone on this journey
for quite a while.
for quite a while.
130
484710
3878
有好長一段時間，
我們在這個旅途中孤獨地踽踽而行，
我們在這個旅途中孤獨地踽踽而行，
08:20
Some very friendly colleagues advised me
to do something more useful for my tenure,
to do something more useful for my tenure,
131
488588
5003
有些同事好心地建議我，
與其苦苦掙扎於研究經費的募集，
與其苦苦掙扎於研究經費的募集，
08:25
and we were constantly struggling
for research funding.
for research funding.
132
493591
4342
還不如轉而先做些比較好拿到終身聘的研究，
08:29
Once, I even joked to my graduate students
133
497933
2485
我還曾跟我的研究生開玩笑說
08:32
that I would just reopen
my dry cleaner's shop to fund ImageNet.
my dry cleaner's shop to fund ImageNet.
134
500418
4063
我乾脆再開一間乾洗店來資助ImageNet 好了，
08:36
After all, that's how I funded
my college years.
my college years.
135
504481
4761
畢竟那就是我用以支付大學學費的方法。
08:41
So we carried on.
136
509242
1856
就這樣我們還是繼續往前走，
08:43
In 2009, the ImageNet project delivered
137
511098
3715
2009年起，ImageNet 已經是個擁有
08:46
a database of 15 million images
138
514813
4042
涵蓋了兩萬兩千種不同類別，
08:50
across 22,000 classes
of objects and things
of objects and things
139
518855
4805
多達150億幅圖像的資料庫，
08:55
organized by everyday English words.
140
523660
3320
並組織以英語日常生活用字為主，
08:58
In both quantity and quality,
141
526980
2926
這樣的規模，不論是「質」或「量」
09:01
this was an unprecedented scale.
142
529906
2972
都是史無前例的。
09:04
As an example, in the case of cats,
143
532878
3461
用貓來舉個例子說明，
09:08
we have more than 62,000 cats
144
536339
2809
我們有超過六萬兩千種
09:11
of all kinds of looks and poses
145
539148
4110
不同外觀和姿勢的貓咪，
09:15
and across all species
of domestic and wild cats.
of domestic and wild cats.
146
543258
5223
橫跨不同的種類，有家貓，也有野貓。
09:20
We were thrilled
to have put together ImageNet,
to have put together ImageNet,
147
548481
3344
ImageNet 的成果讓我們非常激動，
09:23
and we wanted the whole research world
to benefit from it,
to benefit from it,
148
551825
3738
我們希望它有助於全世界的研究，
09:27
so in the TED fashion,
we opened up the entire data set
we opened up the entire data set
149
555563
4041
就如同 TED 的貢獻，我們免費提供整個資料庫
09:31
to the worldwide
research community for free.
research community for free.
150
559604
3592
給全世界的研究單位。
(觀眾鼓掌)
09:36
(Applause)
151
564636
4000
09:41
Now that we have the data
to nourish our computer brain,
to nourish our computer brain,
152
569416
4538
有了這些資料，我們可以教育我們的電腦，
09:45
we're ready to come back
to the algorithms themselves.
to the algorithms themselves.
153
573954
3737
下一步就是回到程式演算的部分了。
09:49
As it turned out, the wealth
of information provided by ImageNet
of information provided by ImageNet
154
577691
5178
結果我們發現，ImageNet 所提供的豐富資訊
09:54
was a perfect match to a particular class
of machine learning algorithms
of machine learning algorithms
155
582869
4806
恰巧與機器學習演算的其中一門特定領域
不謀而合，
不謀而合，
09:59
called convolutional neural network,
156
587675
2415
我們稱它為「卷積神經網絡」，
10:02
pioneered by Kunihiko Fukushima,
Geoff Hinton, and Yann LeCun
Geoff Hinton, and Yann LeCun
157
590090
5248
在七零及八零年代，福島邦彥、Geoff Hinton
10:07
back in the 1970s and '80s.
158
595338
3645
和 Yann LeCun 等學者為該領域的先驅。
10:10
Just like the brain consists
of billions of highly connected neurons,
of billions of highly connected neurons,
159
598983
5619
正如同大腦是由無數個緊密連結的神經元所組成，
10:16
a basic operating unit in a neural network
160
604602
3854
神經網絡的基本運作單位
10:20
is a neuron-like node.
161
608456
2415
也是一個類神經元的節點。
10:22
It takes input from other nodes
162
610871
2554
它的運作方式是從別的節點得到資料，
10:25
and sends output to others.
163
613425
2718
然後再傳給其他的節點。
10:28
Moreover, these hundreds of thousands
or even millions of nodes
or even millions of nodes
164
616143
4713
而且這些數不清的節點
10:32
are organized in hierarchical layers,
165
620856
3227
擁有層層的組織架構，
10:36
also similar to the brain.
166
624083
2554
就好像我們的大腦一樣。
10:38
In a typical neural network we use
to train our object recognition model,
to train our object recognition model,
167
626637
4783
在一般的神經網絡中，
我們用作訓練的物品辨識模型
我們用作訓練的物品辨識模型
10:43
it has 24 million nodes,
168
631420
3181
就有兩千四百萬個節點、
10:46
140 million parameters,
169
634601
3297
一億四千萬個參數，
10:49
and 15 billion connections.
170
637898
2763
以及一百五十億個連結。
10:52
That's an enormous model.
171
640661
2415
這是一個大的不得了的模型。
10:55
Powered by the massive data from ImageNet
172
643076
3901
由ImageNet 提供巨大的資料群、
10:58
and the modern CPUs and GPUs
to train such a humongous model,
to train such a humongous model,
173
646977
5433
並使用先進的核心處理器及圖型處理器來訓練
這個龐然大物，
這個龐然大物，
11:04
the convolutional neural network
174
652410
2369
卷積神經網絡就在眾人的意料外
11:06
blossomed in a way that no one expected.
175
654779
3436
開花結果了。
11:10
It became the winning architecture
176
658215
2508
在物品辨識領域中，這樣的架構
11:12
to generate exciting new results
in object recognition.
in object recognition.
177
660723
5340
以令人興奮的嶄新成果，傲視群雄。
11:18
This is a computer telling us
178
666063
2810
電腦告訴我們
11:20
this picture contains a cat
179
668873
2300
這張圖中有隻貓，
11:23
and where the cat is.
180
671173
1903
還告訴我們貓在哪裡。
11:25
Of course there are more things than cats,
181
673076
2112
當然，這世界不會只有貓，
11:27
so here's a computer algorithm telling us
182
675188
2438
電腦的演算告訴我們
11:29
the picture contains
a boy and a teddy bear;
a boy and a teddy bear;
183
677626
3274
這張圖中有一個男孩和一隻泰迪熊；
11:32
a dog, a person, and a small kite
in the background;
in the background;
184
680900
4366
有狗，一個人，以及背景中的一支小風箏；
11:37
or a picture of very busy things
185
685266
3135
或這一張令人眼花撩亂的圖，
11:40
like a man, a skateboard,
railings, a lampost, and so on.
railings, a lampost, and so on.
186
688401
4644
有人、滑板、欄杆、路燈，等等。
11:45
Sometimes, when the computer
is not so confident about what it sees,
is not so confident about what it sees,
187
693045
5293
有時候，如果電腦不確定自己所見到的東西時，
11:51
we have taught it to be smart enough
188
699498
2276
我們已經將它教到可以聰明地
11:53
to give us a safe answer
instead of committing too much,
instead of committing too much,
189
701774
3878
給一個安全的答案，而非莽撞地回答，
11:57
just like we would do,
190
705652
2811
就像一般人會做的。
12:00
but other times our computer algorithm
is remarkable at telling us
is remarkable at telling us
191
708463
4666
更有些時候，電腦的運算竟能夠
12:05
what exactly the objects are,
192
713129
2253
精準地辨別物體品項
12:07
like the make, model, year of the cars.
193
715382
3436
例如製造商、型號、車子的年份。
12:10
We applied this algorithm to millions
of Google Street View images
of Google Street View images
194
718818
5386
Google 將這個演算程式廣泛地運用在
12:16
across hundreds of American cities,
195
724204
3135
數百個美國城市的街景裡，
12:19
and we have learned something
really interesting:
really interesting:
196
727339
2926
也因此我們從中得到了一些有趣的概念。
12:22
first, it confirmed our common wisdom
197
730265
3320
首先，它證實了一項廣為人知的說法，
12:25
that car prices correlate very well
198
733585
3290
也就是汽車價格和家庭收入
12:28
with household incomes.
199
736875
2345
是息息相關的。
12:31
But surprisingly, car prices
also correlate well
also correlate well
200
739220
4527
然而令人驚訝的是，汽車價格也和
12:35
with crime rates in cities,
201
743747
2300
城市中的犯罪率
12:39
or voting patterns by zip codes.
202
747007
3963
以及區域選舉模式，有相當的關係。
12:44
So wait a minute. Is that it?
203
752060
2206
等等，難道說我今天
12:46
Has the computer already matched
or even surpassed human capabilities?
or even surpassed human capabilities?
204
754266
5153
就是來告訴各位電腦已經趕上
甚至超越人類了嗎？
甚至超越人類了嗎？
12:51
Not so fast.
205
759419
2138
還早得很呢。
12:53
So far, we have just taught
the computer to see objects.
the computer to see objects.
206
761557
4923
到目前為止，我們只是教導電腦識別物品，
12:58
This is like a small child
learning to utter a few nouns.
learning to utter a few nouns.
207
766480
4644
就像小孩子牙牙學語一樣，
13:03
It's an incredible accomplishment,
208
771124
2670
雖然這是個傲人的進展，
13:05
but it's only the first step.
209
773794
2460
但它不過是第一步而已，
13:08
Soon, another developmental
milestone will be hit,
milestone will be hit,
210
776254
3762
很快地，下一波具指標性的後浪就會打上來了，
13:12
and children begin
to communicate in sentences.
to communicate in sentences.
211
780016
3461
小孩子開始進展到用句子來溝通。
13:15
So instead of saying
this is a cat in the picture,
this is a cat in the picture,
212
783477
4224
因此，他已經不會用「這是貓」
來描述圖片，
來描述圖片，
13:19
you already heard the little girl
telling us this is a cat lying on a bed.
telling us this is a cat lying on a bed.
213
787701
5202
而是會聽到這個小女孩說「這是躺在床上的貓」。
13:24
So to teach a computer
to see a picture and generate sentences,
to see a picture and generate sentences,
214
792903
5595
因此，要教導電腦看到圖並說出句子，
13:30
the marriage between big data
and machine learning algorithm
and machine learning algorithm
215
798498
3948
必須進一步地仰賴龐大資料群
13:34
has to take another step.
216
802446
2275
以及機器的學習演算。
13:36
Now, the computer has to learn
from both pictures
from both pictures
217
804721
4156
現在，電腦不僅要學習圖片識別，
13:40
as well as natural language sentences
218
808877
2856
還要學習人類自然的
13:43
generated by humans.
219
811733
3322
說話方式。
13:47
Just like the brain integrates
vision and language,
vision and language,
220
815055
3853
就如同大腦要結合視覺和語言一樣，
13:50
we developed a model
that connects parts of visual things
that connects parts of visual things
221
818908
5201
我們做出了一個模型，
它可以連結不同的可視物體，
它可以連結不同的可視物體，
13:56
like visual snippets
222
824109
1904
就像視覺片段一樣，
13:58
with words and phrases in sentences.
223
826013
4203
並附上句子用的字詞和片語。
14:02
About four months ago,
224
830216
2763
約四個月前，
14:04
we finally tied all this together
225
832979
2647
我們終於把所有的元素全部兜起來了，
14:07
and produced one of the first
computer vision models
computer vision models
226
835626
3784
做出了第一個電腦版的模型，
14:11
that is capable of generating
a human-like sentence
a human-like sentence
227
839410
3994
它有辦法在初次看到照片時
14:15
when it sees a picture for the first time.
228
843404
3506
說出像人類般自然的句子，
14:18
Now, I'm ready to show you
what the computer says
what the computer says
229
846910
4644
好，現在我要給各位看看電腦
14:23
when it sees the picture
230
851554
1975
對於演講一開頭
14:25
that the little girl saw
at the beginning of this talk.
at the beginning of this talk.
231
853529
3830
那位小女孩所看到的影像，
它又是如何理解的。
它又是如何理解的。
14:31
(Video) Computer: A man is standing
next to an elephant.
next to an elephant.
232
859519
3344
(電腦) 有個人站在大象旁邊。
14:36
A large airplane sitting on top
of an airport runway.
of an airport runway.
233
864393
3634
一架大飛機停在機場跑道上。
14:41
FFL: Of course, we're still working hard
to improve our algorithms,
to improve our algorithms,
234
869057
4212
(主講人) 當然，我們仍戮力於改善這電腦程式，
14:45
and it still has a lot to learn.
235
873269
2596
它還有很多要學。
14:47
(Applause)
236
875865
2291
(觀眾鼓掌)
14:51
And the computer still makes mistakes.
237
879556
3321
電腦還是會犯錯。
14:54
(Video) Computer: A cat lying
on a bed in a blanket.
on a bed in a blanket.
238
882877
3391
(電腦) 一隻貓包著毯子躺在床上。
14:58
FFL: So of course, when it sees
too many cats,
too many cats,
239
886268
2553
(主講人) 因為它看了太多貓了，
15:00
it thinks everything
might look like a cat.
might look like a cat.
240
888821
2926
以至於它見到了什麼都像貓咪。
15:05
(Video) Computer: A young boy
is holding a baseball bat.
is holding a baseball bat.
241
893317
2864
(電腦) 一位小男孩握著一支球棒。
15:08
(Laughter)
242
896181
1765
(觀眾笑)
(主講人) 或者，如果電腦是第一次看到牙刷，
會把它與球棒混淆。
會把它與球棒混淆。
15:09
FFL: Or, if it hasn't seen a toothbrush,
it confuses it with a baseball bat.
it confuses it with a baseball bat.
243
897946
4583
15:15
(Video) Computer: A man riding a horse
down a street next to a building.
down a street next to a building.
244
903309
3434
(電腦) 一個人在建築物旁的街道上騎馬。
15:18
(Laughter)
245
906743
2023
(觀眾笑)
15:20
FFL: We haven't taught Art 101
to the computers.
to the computers.
246
908766
3552
(主講人) 我們還沒讓電腦上基礎美術課。
15:25
(Video) Computer: A zebra standing
in a field of grass.
in a field of grass.
247
913768
2884
(電腦) 一匹斑馬站在原野中。
15:28
FFL: And it hasn't learned to appreciate
the stunning beauty of nature
the stunning beauty of nature
248
916652
3367
(主講人) 電腦還沒辦法像人類一樣，
15:32
like you and I do.
249
920019
2438
學會欣賞大自然的美景。
15:34
So it has been a long journey.
250
922457
2832
這是條漫漫長路，
15:37
To get from age zero to three was hard.
251
925289
4226
要從零歲發展到三歲是很難的，
15:41
The real challenge is to go
from three to 13 and far beyond.
from three to 13 and far beyond.
252
929515
5596
更艱深的挑戰在於從三歲發展到十三歲，
甚至到更遠的階段。
甚至到更遠的階段。
15:47
Let me remind you with this picture
of the boy and the cake again.
of the boy and the cake again.
253
935111
4365
讓我用這張男孩與蛋糕的圖片來進一步說明，
15:51
So far, we have taught
the computer to see objects
the computer to see objects
254
939476
4064
直到今日，我們已經教會了電腦識別物品，
15:55
or even tell us a simple story
when seeing a picture.
when seeing a picture.
255
943540
4458
甚至於在看到一張圖後，可以簡單地敘述。
15:59
(Video) Computer: A person sitting
at a table with a cake.
at a table with a cake.
256
947998
3576
(電腦) 一個人和蛋糕坐在桌旁。
16:03
FFL: But there's so much more
to this picture
to this picture
257
951574
2630
(主講人) 這張照片其實蘊涵著更多的東西，
16:06
than just a person and a cake.
258
954204
2270
不僅只有人和蛋糕。
16:08
What the computer doesn't see
is that this is a special Italian cake
is that this is a special Italian cake
259
956474
4467
電腦看不出這是種特別的義式蛋糕，
16:12
that's only served during Easter time.
260
960941
3217
人們只有在復活節時才會做。
16:16
The boy is wearing his favorite t-shirt
261
964158
3205
這個男孩穿著他最心愛的T恤，
16:19
given to him as a gift by his father
after a trip to Sydney,
after a trip to Sydney,
262
967363
3970
是去雪梨玩的時候，他的父親送的，
16:23
and you and I can all tell how happy he is
263
971333
3808
各位和我都可以看得出他有多快樂，
16:27
and what's exactly on his mind
at that moment.
at that moment.
264
975141
3203
以及當時他的心裡在想什麼。
16:31
This is my son Leo.
265
979214
3125
這是我兒子，李奧。
16:34
On my quest for visual intelligence,
266
982339
2624
在探索智能視覺的旅途上，
16:36
I think of Leo constantly
267
984963
2391
我不斷地想到他，
16:39
and the future world he will live in.
268
987354
2903
以及他在將來生活的世界，
16:42
When machines can see,
269
990257
2021
當未來，機器有了視覺，
16:44
doctors and nurses will have
extra pairs of tireless eyes
extra pairs of tireless eyes
270
992278
4712
醫生和護士就多了雙永不倦怠的眼睛，
16:48
to help them to diagnose
and take care of patients.
and take care of patients.
271
996990
4092
幫助他們診斷及照顧病人；
16:53
Cars will run smarter
and safer on the road.
and safer on the road.
272
1001082
4383
行駛在路上的車子可以更聰明、更安全；
16:57
Robots, not just humans,
273
1005465
2694
人類與機器人能一起
17:00
will help us to brave the disaster zones
to save the trapped and wounded.
to save the trapped and wounded.
274
1008159
4849
共同投入災區的救援工作，拯救受困人員及傷者；
17:05
We will discover new species,
better materials,
better materials,
275
1013798
3796
我們還可以發現新品種
與更好的材料，
與更好的材料，
17:09
and explore unseen frontiers
with the help of the machines.
with the help of the machines.
276
1017594
4509
探索未知的疆界，
這一切都可仰賴機器的協助。
這一切都可仰賴機器的協助。
17:15
Little by little, we're giving sight
to the machines.
to the machines.
277
1023113
4167
一步一步地，我們賦予機器視覺，
17:19
First, we teach them to see.
278
1027280
2798
先教他們識別物品，
17:22
Then, they help us to see better.
279
1030078
2763
然後它們也讓我們看得更清楚，
17:24
For the first time, human eyes
won't be the only ones
won't be the only ones
280
1032841
4165
這是第一次人類的眼睛不是唯一
17:29
pondering and exploring our world.
281
1037006
2934
可以用來思考和探索世界的工具，
17:31
We will not only use the machines
for their intelligence,
for their intelligence,
282
1039940
3460
我們不僅可以利用機器的智能，
17:35
we will also collaborate with them
in ways that we cannot even imagine.
in ways that we cannot even imagine.
283
1043400
6179
更可以運用更多你想像不到的方式攜手合作。
17:41
This is my quest:
284
1049579
2161
這是我想追求的目標：
17:43
to give computers visual intelligence
285
1051740
2712
給予機器智慧之眼，
17:46
and to create a better future
for Leo and for the world.
for Leo and for the world.
286
1054452
5131
為李奧和整個世界創造更美好的未來。
17:51
Thank you.
287
1059583
1811
謝謝各位。
17:53
(Applause)
288
1061394
3785
(觀眾鼓掌)
ABOUT THE SPEAKER
Fei-Fei Li - Computer scientistAs Director of Stanford’s Artificial Intelligence Lab and Vision Lab, Fei-Fei Li is working to solve AI’s trickiest problems -- including image recognition, learning and language processing.
Why you should listen
Using algorithms built on machine learning methods such as neural network models, the Stanford Artificial Intelligence Lab led by Fei-Fei Li has created software capable of recognizing scenes in still photographs -- and accurately describe them using natural language.
Li’s work with neural networks and computer vision (with Stanford’s Vision Lab) marks a significant step forward for AI research, and could lead to applications ranging from more intuitive image searches to robots able to make autonomous decisions in unfamiliar situations.
Fei-Fei was honored as one of Foreign Policy's 2015 Global Thinkers.
Fei-Fei Li | Speaker | TED.com