How I'm fighting bias in algorithms | Joy Buolamwini

308,224 views ・ 2017-03-29

TED


請雙擊下方英文字幕播放視頻。

譯者: Suzie Tang 審譯者: Helen Chang
00:12
Hello, I'm Joy, a poet of code,
0
12861
3134
你好 我叫玖伊 是個寫媒體程式的詩人
00:16
on a mission to stop an unseen force that's rising,
1
16019
4993
我的使命是
終止一個隱形力量的崛起
00:21
a force that I called "the coded gaze,"
2
21036
2856
我稱這種力量為「數碼凝視」
00:23
my term for algorithmic bias.
3
23916
3309
是我替偏差演算法取的名稱
00:27
Algorithmic bias, like human bias, results in unfairness.
4
27249
4300
偏差的演算法跟人的偏見一樣
會導致不公平的結果
00:31
However, algorithms, like viruses, can spread bias on a massive scale
5
31573
6022
然而演算法更像病毒
它傳播的偏見
大量而迅速
00:37
at a rapid pace.
6
37619
1582
00:39
Algorithmic bias can also lead to exclusionary experiences
7
39763
4387
演算法偏差讓人 體驗到什麼叫做被排擠
00:44
and discriminatory practices.
8
44174
2128
也會導致差別對待
00:46
Let me show you what I mean.
9
46326
2061
讓我告訴你我的意思
00:48
(Video) Joy Buolamwini: Hi, camera. I've got a face.
10
48800
2436
嗨 相機 我有一張臉
00:51
Can you see my face?
11
51982
1864
你能看見我的臉嗎?
00:53
No-glasses face?
12
53871
1625
不戴眼鏡呢?
00:55
You can see her face.
13
55521
2214
你看得見她啊
00:58
What about my face?
14
58057
2245
那麼我的臉呢?
01:03
I've got a mask. Can you see my mask?
15
63710
3750
戴上面具 你看得見戴上面具嗎?
01:08
Joy Buolamwini: So how did this happen?
16
68294
2365
到底是怎麽回事?
01:10
Why am I sitting in front of a computer
17
70683
3141
我為什麽要坐在電腦前
01:13
in a white mask,
18
73848
1424
戴著白色面具
01:15
trying to be detected by a cheap webcam?
19
75296
3650
好讓這台廉價的攝影機能看得見我
01:18
Well, when I'm not fighting the coded gaze
20
78970
2291
如果我沒有忙著對抗數碼凝視
01:21
as a poet of code,
21
81285
1520
當個媒體程式詩人
01:22
I'm a graduate student at the MIT Media Lab,
22
82829
3272
我就是麻省理工媒體實驗室的研究生
01:26
and there I have the opportunity to work on all sorts of whimsical projects,
23
86125
4917
我在那裡從事一些稀奇古怪的計劃
01:31
including the Aspire Mirror,
24
91066
2027
包括照妖鏡
01:33
a project I did so I could project digital masks onto my reflection.
25
93117
5134
照妖鏡計劃
讓我能把數位面具投射在自己臉上
01:38
So in the morning, if I wanted to feel powerful,
26
98275
2350
早上起來如果我需要強大的力量
01:40
I could put on a lion.
27
100649
1434
我就投上一個獅子面具
01:42
If I wanted to be uplifted, I might have a quote.
28
102107
3496
如果我缺乏鬥志
我就放一段名人名言
01:45
So I used generic facial recognition software
29
105627
2989
因為我使用一般的臉部辨識軟體
01:48
to build the system,
30
108640
1351
來測試這個系統
01:50
but found it was really hard to test it unless I wore a white mask.
31
110015
5103
結果竟然發現
電腦無法偵測到我
除非我戴上白色面具
很不幸我之前就碰過這種問題
01:56
Unfortunately, I've run into this issue before.
32
116102
4346
02:00
When I was an undergraduate at Georgia Tech studying computer science,
33
120472
4303
先前我在喬治亞理工學院
攻讀電腦科學學士學位時
02:04
I used to work on social robots,
34
124799
2055
我研究社交機器人
02:06
and one of my tasks was to get a robot to play peek-a-boo,
35
126878
3777
其中的一個實驗
就是和機器人玩躲貓貓
02:10
a simple turn-taking game
36
130679
1683
這個簡單的互動遊戲
02:12
where partners cover their face and then uncover it saying, "Peek-a-boo!"
37
132386
4321
讓對手先遮住臉再放開
同時要說 peek-a-boo
02:16
The problem is, peek-a-boo doesn't really work if I can't see you,
38
136731
4429
問題是如果看不到對方
遊戲就玩不下去了
02:21
and my robot couldn't see me.
39
141184
2499
我的機器人就是看不到我
02:23
But I borrowed my roommate's face to get the project done,
40
143707
3950
最後我只好借我室友的臉來完成
02:27
submitted the assignment,
41
147681
1380
做完實驗時我想
02:29
and figured, you know what, somebody else will solve this problem.
42
149085
3753
總有一天會有別人解決這個問題
02:33
Not too long after,
43
153489
2003
不久之後
02:35
I was in Hong Kong for an entrepreneurship competition.
44
155516
4159
我去香港參加一個
業界舉辦的競技比賽
02:40
The organizers decided to take participants
45
160159
2694
主辦單位先帶每位參賽者
02:42
on a tour of local start-ups.
46
162877
2372
去參觀當地的新創市場
02:45
One of the start-ups had a social robot,
47
165273
2715
其中一項就是社交機器人
02:48
and they decided to do a demo.
48
168012
1912
當他們用社交機器人展示成果時
02:49
The demo worked on everybody until it got to me,
49
169948
2980
社交機器人對每個參賽者都有反應
02:52
and you can probably guess it.
50
172952
1923
直到遇到了我
接下來的情形你應該能想像
02:54
It couldn't detect my face.
51
174899
2965
社交機器人怎樣都偵測不到我的臉
02:57
I asked the developers what was going on,
52
177888
2511
我問軟體開發人員是怎麼一回事
03:00
and it turned out we had used the same generic facial recognition software.
53
180423
5533
才驚覺當年通用的
人臉辨識軟體
03:05
Halfway around the world,
54
185980
1650
竟然飄洋過海到了香港
03:07
I learned that algorithmic bias can travel as quickly
55
187654
3852
偏差的演算邏輯快速散播
03:11
as it takes to download some files off of the internet.
56
191530
3170
只要從網路下載幾個檔案就搞定了
03:15
So what's going on? Why isn't my face being detected?
57
195565
3076
為什麼機器人就是看不見我的臉?
03:18
Well, we have to look at how we give machines sight.
58
198665
3356
得先知道我們如何賦予機器視力
03:22
Computer vision uses machine learning techniques
59
202045
3409
電腦使用機器學習的技術
03:25
to do facial recognition.
60
205478
1880
來辨識人臉
03:27
So how this works is, you create a training set with examples of faces.
61
207382
3897
你必須用許多實作測試來訓練他們
這是人臉這是人臉這是人臉
03:31
This is a face. This is a face. This is not a face.
62
211303
2818
這不是人臉
03:34
And over time, you can teach a computer how to recognize other faces.
63
214145
4519
一而再再而三你就能教機器人
辨識其他的人臉
03:38
However, if the training sets aren't really that diverse,
64
218688
3989
但是如果實作測試不夠多樣化
03:42
any face that deviates too much from the established norm
65
222701
3349
當出現的人臉
與既定規範相去太遠時
03:46
will be harder to detect,
66
226074
1649
電腦就很難判斷了
03:47
which is what was happening to me.
67
227747
1963
我的親身經驗就是這樣
03:49
But don't worry -- there's some good news.
68
229734
2382
但別慌張 有好消息
03:52
Training sets don't just materialize out of nowhere.
69
232140
2771
實作測試並不是無中生有
03:54
We actually can create them.
70
234935
1788
事實上我們能夠建的
03:56
So there's an opportunity to create full-spectrum training sets
71
236747
4176
我們可以有一套更周詳的測試樣本
04:00
that reflect a richer portrait of humanity.
72
240947
3824
涵蓋人種的多樣性
04:04
Now you've seen in my examples
73
244795
2221
我的實驗說明了
04:07
how social robots
74
247040
1768
社交機器人
04:08
was how I found out about exclusion with algorithmic bias.
75
248832
4611
產生排他現象
因為偏差的演算邏輯
04:13
But algorithmic bias can also lead to discriminatory practices.
76
253467
4815
偏差的演算邏輯
也可能讓偏見成為一種習慣
04:19
Across the US,
77
259257
1453
美國各地的警方
04:20
police departments are starting to use facial recognition software
78
260734
4198
正開始使用這套人臉辨識軟體
04:24
in their crime-fighting arsenal.
79
264956
2459
來建立警方的打擊犯罪系統
04:27
Georgetown Law published a report
80
267439
2013
喬治城大學法律中心的報告指出
04:29
showing that one in two adults in the US -- that's 117 million people --
81
269476
6763
每兩個美國成年人就有一個人
也就是一億一千七百萬筆臉部資料
04:36
have their faces in facial recognition networks.
82
276263
3534
在美國警方這套系統裡
04:39
Police departments can currently look at these networks unregulated,
83
279821
4552
警方這套系統既缺乏規範
04:44
using algorithms that have not been audited for accuracy.
84
284397
4286
也缺乏正確合法的演算邏輯
04:48
Yet we know facial recognition is not fail proof,
85
288707
3864
你要知道人臉辨識並非萬無一失
04:52
and labeling faces consistently remains a challenge.
86
292595
4179
要一貫正確地標註人臉 往往不是那麼容易
04:56
You might have seen this on Facebook.
87
296798
1762
或許你在臉書上看過
04:58
My friends and I laugh all the time when we see other people
88
298584
2988
朋友和我常覺得很好笑
看見有人標註朋友卻標錯了
05:01
mislabeled in our photos.
89
301596
2458
05:04
But misidentifying a suspected criminal is no laughing matter,
90
304078
5591
如果標錯的是犯人的臉呢
那就讓人笑不出來了
05:09
nor is breaching civil liberties.
91
309693
2827
侵害公民自由也同樣讓人笑不出來
05:12
Machine learning is being used for facial recognition,
92
312544
3205
不僅辨識人臉倚賴機器學習的技術
05:15
but it's also extending beyond the realm of computer vision.
93
315773
4505
許多領域其實都要用到機器學習
05:21
In her book, "Weapons of Math Destruction,"
94
321086
4016
《大數據的傲慢與偏見》 這本書的作者
05:25
data scientist Cathy O'Neil talks about the rising new WMDs --
95
325126
6681
數據科學家凱西 歐尼爾
談到新 WMD 勢力的崛起
05:31
widespread, mysterious and destructive algorithms
96
331831
4353
WMD 是廣泛 神秘和具破壞性的算法
05:36
that are increasingly being used to make decisions
97
336208
2964
演算法漸漸取代我們做決定
05:39
that impact more aspects of our lives.
98
339196
3177
影響我們生活的更多層面
05:42
So who gets hired or fired?
99
342397
1870
例如誰升了官?誰丟了飯碗?
05:44
Do you get that loan? Do you get insurance?
100
344291
2112
你借到錢了嗎?你買保險了嗎?
05:46
Are you admitted into the college you wanted to get into?
101
346427
3503
你進入心目中理想的大學了嗎?
05:49
Do you and I pay the same price for the same product
102
349954
3509
我們花同樣多的錢在同樣的平台上
05:53
purchased on the same platform?
103
353487
2442
買到同樣的產品嗎?
05:55
Law enforcement is also starting to use machine learning
104
355953
3759
警方也開始使用機器學習
05:59
for predictive policing.
105
359736
2289
來防範犯罪
06:02
Some judges use machine-generated risk scores to determine
106
362049
3494
法官根據電腦顯示的危險因子數據
06:05
how long an individual is going to spend in prison.
107
365567
4402
來決定一個人要在監獄待幾年
06:09
So we really have to think about these decisions.
108
369993
2454
我們得仔細想想這些判定
06:12
Are they fair?
109
372471
1182
它們真的公平嗎?
06:13
And we've seen that algorithmic bias
110
373677
2890
我們親眼看見偏差的演算邏輯
06:16
doesn't necessarily always lead to fair outcomes.
111
376591
3374
未必做出正確的判斷
06:19
So what can we do about it?
112
379989
1964
我們該怎麽辦呢?
06:21
Well, we can start thinking about how we create more inclusive code
113
381977
3680
我們要先確定程式碼是否具多樣性
06:25
and employ inclusive coding practices.
114
385681
2990
以及寫程式的過程是否周詳
06:28
It really starts with people.
115
388695
2309
事實上全都始於人
06:31
So who codes matters.
116
391528
1961
程式是誰寫的有關係
06:33
Are we creating full-spectrum teams with diverse individuals
117
393513
4119
寫程式的團隊是否由 多元的個體組成呢?
06:37
who can check each other's blind spots?
118
397656
2411
這樣才能互補並找出彼此的盲點
06:40
On the technical side, how we code matters.
119
400091
3545
從技術面而言 我們如何寫程式很重要
06:43
Are we factoring in fairness as we're developing systems?
120
403660
3651
我們是否對公平這項要素
在系統開發階段就考量到呢?
06:47
And finally, why we code matters.
121
407335
2913
最後 我們為什麼寫程式也重要
06:50
We've used tools of computational creation to unlock immense wealth.
122
410605
5083
我們使用計算創造工具 開啟了巨額財富之門
06:55
We now have the opportunity to unlock even greater equality
123
415712
4447
我們現在有機會實現更大的平等
07:00
if we make social change a priority
124
420183
2930
如果我們將社會變革作為優先事項
07:03
and not an afterthought.
125
423137
2170
而不是事後的想法
07:05
And so these are the three tenets that will make up the "incoding" movement.
126
425828
4522
這裡有改革程式的三元素
07:10
Who codes matters,
127
430374
1652
程式是誰寫的重要
07:12
how we code matters
128
432050
1543
如何寫程式重要
07:13
and why we code matters.
129
433617
2023
以及為何寫程式重要
07:15
So to go towards incoding, we can start thinking about
130
435664
3099
要成功改革程式
我們可以先從建立能夠 找出偏差的分析平台開始
07:18
building platforms that can identify bias
131
438787
3164
07:21
by collecting people's experiences like the ones I shared,
132
441975
3078
作法是收集人們的親身經歷 像是我剛才分享的經歷
07:25
but also auditing existing software.
133
445077
3070
也檢視現存的軟體
07:28
We can also start to create more inclusive training sets.
134
448171
3765
我們可以著手建立 更具包容性的測試樣本
07:31
Imagine a "Selfies for Inclusion" campaign
135
451960
2803
想像「包容的自拍」活動
07:34
where you and I can help developers test and create
136
454787
3655
我們可以幫助開發人員測試和創建
07:38
more inclusive training sets.
137
458466
2093
更具包容性的測試樣本
07:41
And we can also start thinking more conscientiously
138
461122
2828
我們也要更自省
07:43
about the social impact of the technology that we're developing.
139
463974
5391
我們發展的科技帶給社會的衝擊
07:49
To get the incoding movement started,
140
469389
2393
為了著手程式改革
07:51
I've launched the Algorithmic Justice League,
141
471806
2847
我發起了「演算邏輯正義聯盟」
07:54
where anyone who cares about fairness can help fight the coded gaze.
142
474677
5872
只要你贊同公平
就可以加入打擊數碼凝視的行列
08:00
On codedgaze.com, you can report bias,
143
480573
3296
只要上 codedgaze.com 網路
可以舉報你發現的偏差演算邏輯
08:03
request audits, become a tester
144
483893
2445
可以申請測試
可以成為受測者
08:06
and join the ongoing conversation,
145
486362
2771
也可以加入論壇
08:09
#codedgaze.
146
489157
2287
只要搜尋 #codedgaze
08:12
So I invite you to join me
147
492562
2487
我在此邀請大家加入我的行列
08:15
in creating a world where technology works for all of us,
148
495073
3719
創造一個技術適用於 我們所有人的世界
08:18
not just some of us,
149
498816
1897
而不是只適用於某些人
08:20
a world where we value inclusion and center social change.
150
500737
4588
一個重視包容性 和以社會變革為中心的世界
08:25
Thank you.
151
505349
1175
謝謝
08:26
(Applause)
152
506548
4271
(掌聲)
08:32
But I have one question:
153
512693
2854
我還有一個問題
08:35
Will you join me in the fight?
154
515571
2059
你要和我並肩作戰嗎?
08:37
(Laughter)
155
517654
1285
(笑聲)
08:38
(Applause)
156
518963
3687
(掌聲)
關於本網站

本網站將向您介紹對學習英語有用的 YouTube 視頻。 您將看到來自世界各地的一流教師教授的英語課程。 雙擊每個視頻頁面上顯示的英文字幕,從那裡播放視頻。 字幕與視頻播放同步滾動。 如果您有任何意見或要求,請使用此聯繫表與我們聯繫。

https://forms.gle/WvT1wiN1qDtmnspy7