Next Up for AI? Dancing Robots | Catie Cuan | TED

52,061 views ・ 2024-07-02

TED


请双击下面的英文字幕来播放视频。

翻译人员: Fanglin ren 校对人员: Sue Lu
00:04
I have an unusual job.
0
4376
1918
我有一份不寻常的工作。
00:06
I teach robots how to dance.
1
6294
2878
我教机器人如何跳舞。
00:09
In 2017,
2
9214
1501
2017 年,
00:10
I transitioned from dancing on stages with humans
3
10757
4755
我从与人类在舞台上共舞过渡到
00:15
to dancing with robots.
4
15512
2294
与机器人共舞。
00:18
Big robots, small robots,
5
18640
2586
大型机器人、小型机器人、
00:21
single-arms and roving robots.
6
21268
2252
单臂和巡回机器人,
00:23
Robots that fly,
7
23562
1292
会飞的机器人,
00:24
vacuuming robots.
8
24854
1419
吸尘机器人,
00:26
A lot of robots that look like humans.
9
26273
2085
很多看起来像人类的机器人,
00:28
And a whole bunch that don't.
10
28400
2419
还有一大群不像的。
00:30
I'm pretty sure I've choreographed
11
30860
1627
我敢肯定,我编舞和
00:32
and danced with more different types of robots
12
32529
2252
跳舞时合作过的不同类型的机器人
00:34
than anyone on Earth.
13
34781
1710
比地球上任何人都多。
00:37
Why do I do this?
14
37784
1460
我为什么要这样做?
00:39
Because robots move
15
39619
2086
因为机器人会移动,
00:41
and choreographers understand how movement affects us.
16
41746
3254
编舞家知道运动如何影响我们。
00:45
If a robot slides politely out of a doorway to let you pass,
17
45333
4338
如果机器人礼貌地从门口滑开让你通过,
00:49
might make you feel seen and acknowledged.
18
49671
2836
可能会让你感到被看见和认可。
00:52
If a robot marches quickly towards you and avoids you at the last second,
19
52841
4212
如果机器人快速走向你, 只在最后一秒避开你,
00:57
it might cause revulsion and fear.
20
57095
2336
那可能会引起厌恶和恐惧。
01:00
Robots are beginning to show up in our everyday environments,
21
60390
2878
机器人开始出现在我们的日常环境中,
01:03
from sidewalks to offices, backyards to hospitals.
22
63310
3169
从人行道到办公室,从后院到医院。
01:06
And they will be threatening and confusing to us
23
66855
2919
他们将对我们构成威胁和困惑,
01:09
if we do not carefully examine how they move.
24
69816
4588
除非我们仔细研究他们的行动方式。
01:14
I'm also a roboticist and an engineer.
25
74446
3211
我也是机器人专家和工程师。
01:17
I study this connection between robot motion and emotion.
26
77657
3587
我研究机器人动作与情感间的联系。
01:21
I combine my expertise in robotics and dance
27
81244
3504
我结合自己在机器人技术 和舞蹈方面的专业知识
01:24
in the emerging field of choreo-robotics,
28
84748
3336
用于新兴的编舞机器人领域,
01:28
the intersection of dance and robotics.
29
88084
3087
即舞蹈和机器人技术的交汇处。
01:31
Right now, we're giving AI a robot body with physical intelligence.
30
91963
5673
现在,我们正在为人工智能(AI) 提供具有物理智能的机器人身体。
01:37
AI technologies like GPT and Gemini are becoming the robot's brains,
31
97677
4296
AI 技术如 GPT 和 Gemini 正在成为机器人的大脑,
01:42
which means that robots are getting smarter.
32
102015
3795
这意味着机器人变得越来越聪明,
01:45
Much smarter.
33
105810
1168
非常聪明。
01:47
Think of what AI has already done to text and video generation.
34
107562
3462
想想 AI 能在文本和视频生成方面做的。
01:51
We are just scratching the surface on seeing what AI can do
35
111024
4296
我们对 AI 的了解还仅仅停留在
01:55
in the physical world.
36
115362
1918
表面阶段。
01:57
We are going to be interacting
37
117739
3795
我们将越来越多地
02:01
with these new intelligent robots more and more,
38
121576
3295
与这些新的智能机器人互动,
02:04
interacting with them through movement.
39
124871
3420
通过动作与它们互动。
02:08
Roboticists already spend a lot of time thinking about movement,
40
128708
3003
机器人专家花了很多时间思考这些动作,
02:11
but they think about it in terms of efficiency,
41
131753
3253
但他们思考的角度是效率,
02:15
optimizing for speed or success.
42
135048
2919
优化速度或成功率。
02:18
That's not enough.
43
138593
1418
可那是不够的。
02:20
We need to teach robots how to dance.
44
140679
4379
我们需要教机器人如何跳舞。
02:26
If a robot can dance,
45
146101
1501
如果一个机器人能够跳舞,
02:27
make every movement with intention, agility, balance,
46
147644
4087
每个动作都带有企图、敏捷性、平衡性、
02:31
emotional expression and grace,
47
151773
2711
情感表达和优雅风度,
02:34
I believe they will also become better caregivers,
48
154484
3003
我相信他们就会成为更好的看护者、
02:37
teachers and companions.
49
157529
1668
老师和同伴。
02:39
It will help us build better general-purpose robots
50
159656
3545
这样我们就能建造更好的通用机器人,
02:43
that we actually want to live amongst us.
51
163201
2711
我们希望与之共存的机器人。
02:46
If we teach robots to dance,
52
166204
2586
如果我们教机器人跳舞,
02:48
we expand the future of intelligent machines.
53
168790
4421
我们就能拓展智能机器的未来。
02:54
Choreographers and engineers have made dances with robots since the 1990s,
54
174087
4546
自 20 世纪 90 年代以来, 编舞家和工程师一直在训练机器人跳舞,
02:58
and these highly scripted performances have captured the public's imagination.
55
178675
4796
这些高度编排的表演 俘获了公众的想象力。
03:03
They've made us feel wonder,
56
183888
2169
它们让我们感到惊奇,
03:06
made us laugh,
57
186057
1710
让我们开心大笑,
03:07
made us want to dance along,
58
187809
2252
让我们想随之起舞,
03:10
and sometimes made us never want to meet a robot in real life.
59
190103
3795
但有时也让我们永远不想再遇到机器人。
03:15
Dancing with robots is no longer just about art or entertainment.
60
195024
4630
与机器人共舞不再只是艺术或娱乐。
03:20
It is about transferring the lessons that one, all movement
61
200196
6215
它更是关于传授经验教训, 第一,所有动作都会
03:26
changes emotions,
62
206453
2168
改变情绪;
03:28
and two, contextual movement is critically important
63
208663
3754
第二,情境动作至关重要。
03:32
and applying those lessons to robots in our everyday lives.
64
212459
4337
这些经验教训应该用在机器人身上。
03:37
Robotics engineers are beginning to consider emotion and context
65
217422
5172
机器人工程师开始考虑情感和背景,
03:42
when designing intelligent machines that interact with us.
66
222594
3753
尤其在设计与我们互动的智能机器时。
03:46
The choices we make now
67
226389
2294
我们现在做出的选择
03:48
are going to dictate how we live our lives with robots.
68
228725
4296
将决定我们如何与机器人共存。
03:53
If we get this right,
69
233813
1669
如果我们做对了,
03:55
robots will be more welcomed, safe, and delightful.
70
235482
3545
机器人将更受欢迎、安全且令人愉快。
03:59
Humans will be more empowered and comfortable.
71
239694
4004
人类将变得更有能力,更舒适。
04:04
Before AI, programmers needed hours to script a simple
72
244407
6423
在 AI 出现之前,程序员 需要数小时来编写一个简单的
04:10
dance sequence for robot to perform.
73
250872
2753
舞蹈序列供机器人表演,
04:14
Just like they needed hours to script the robot to open a single door.
74
254125
4588
就像他们需花数小时 编写机器人打开一扇门的程序。
04:19
With AI, you can teach the robot to open just a few specific doors,
75
259088
5172
借助 AI,你可以教机器人 打开几扇特定的门,
04:24
and it will learn to open all of them,
76
264302
2252
然后它就能学会打开所有的门,
04:26
even ones it hasn't seen before.
77
266554
2169
即使是以前从未见过的门。
04:29
It's also true for dance.
78
269349
2210
舞蹈也是如此。
04:32
You can teach the robot to dance with a specific person,
79
272018
3962
你可以教机器人与某个人共舞,
04:35
and it will learn how to dance
80
275980
1627
它就能学习如何
04:37
and move with many others
81
277649
2335
与其他许多人跳舞,
04:39
in many different environments and circumstances.
82
279984
2753
即使是在许多不同的环境里。
04:43
This is what I did at Everyday Robots,
83
283363
2544
这就是我在 Everyday Robots 做的工作,
04:45
then a robot AI moonshot at Google.
84
285907
2920
谷歌的一个机器人 AI 探索项目。
04:49
But rather than teach one robot,
85
289327
2711
但不是教一个机器人,
04:52
I used AI to teach 15 robots how to move together as a flock.
86
292038
5339
而是用 AI 来教 15 个机器人 如何成群结队地一起移动。
04:58
We imagined a world where you could walk down a hallway filled with robots,
87
298586
5005
我们设想了一个世界,在那里, 你可以沿着满是机器人的走廊行走,
05:03
and they would part to make space for you,
88
303633
2419
它们会分散开来为你让路,
05:06
like a flock of doves or a crowd of people on a city street.
89
306094
4546
如同一群鸽子或城市街道上的人群。
05:10
Where a robot could navigate seamlessly and even beautifully through a busy,
90
310640
4713
在那里,机器人可以天衣无缝地 甚至优雅地穿行在繁忙、
05:15
chaotic Times Square.
91
315395
2043
混乱的纽约时代广场上。
05:17
The world we live in is complicated for robots to understand,
92
317981
4296
我们生活的世界很复杂, 机器人无法理解,
05:22
so it was impossible to script or choreograph interactions like these
93
322277
3670
因此不可能按部就班地逐步编排
05:25
step by step.
94
325947
1668
这样的互动。
05:27
We needed to use the magic of AI
95
327657
1835
我们需要使用 AI 的魔力
05:29
to teach the robots how to improvise and react.
96
329492
4046
来教机器人如何即兴创作和反应。
05:34
This was a massive challenge that had never been done before,
97
334539
3712
这是一项前所未有的巨大挑战,
05:38
and the result is an interactive performance and installation work
98
338293
3837
结果就是一个互动表演和装置作品,
05:42
called Project Starling.
99
342130
2043
名为 Project Starling。
05:44
This took us two years of hardcore engineering to complete.
100
344173
3504
我们花了两年才完成这项艰巨的工程。
05:48
We trained an AI agent on the preferences of a choreographer,
101
348428
3420
我们培训了一名 AI 代理, 是根据编舞家的喜好训练的,
05:51
me.
102
351848
1251
也就是我的喜好。
05:53
I chose the most engaging patterns and sequences.
103
353141
3837
我选择了引人入胜的舞蹈模式和序列。
05:57
We also taught the robots how to recognize
104
357437
3378
我们还教会机器人如何识别
06:00
and respond to human gestures.
105
360815
2628
和回应人类的示意动作。
06:04
While we were at it,
106
364193
1210
在那期间,
06:05
we turned the robots into musical instruments.
107
365445
2878
我们还把机器人变成了乐器。
06:08
Each joint on the robot was mapped to a different sound.
108
368948
4880
机器人的每个关节都映射不同的声响。
06:13
When the robot's torso moved, triggered a bass sound.
109
373870
4212
当它的躯干移动时,会触发低音声。
06:18
Its hand opening and closing triggered a bell sound.
110
378750
3795
它的双手开合会触发钟声。
06:23
The robot's actions created its own unique symphony.
111
383004
4755
机器人的移动创造出自己独特的交响曲。
06:28
This is what Project Starling looks like.
112
388259
2127
这就是 Project Starling。
06:30
Our AI generated movement and our robot generated music.
113
390428
4505
我们的 AI 生成舞蹈动作, 机器人生成音乐。
06:36
(Music)
114
396684
6131
(音乐)
06:54
(Music ends)
115
414118
1710
(音乐结束)
06:56
(Applause)
116
416829
3254
(掌声)
07:00
I wanted it to feel like stepping out of your normal reality
117
420124
5172
我希望你的感觉就像是走出现实,
07:05
and into a waking dream,
118
425338
2377
进入一个清醒的梦境一样。
07:07
one that was so far from where we stand today.
119
427757
3545
这个梦离我们的现状还很遥远,
07:11
Yet, all the same felt like a world
120
431636
2044
然而,你依然感觉那像一个安宁的世界,
07:13
where your kids could grow up and feel calm,
121
433680
3878
你的孩子可以在那里健康成长,
07:17
where you could be surrounded by machines and it felt natural and safe.
122
437558
5089
你成天被机器人围着却感觉自然而安全。
07:23
People were drawn to just wander amongst these robots, you know,
123
443523
4254
人们被吸引到这些机器人中闲逛,
07:27
some were a bit tentative at first,
124
447819
2335
有的起初有点犹豫,
07:30
but then they began to engage and explore
125
450196
2586
但慢慢就开始进行互动和探索了,
07:32
with curiosity and playfulness.
126
452782
2502
是带着好奇和好玩的心情的。
07:35
People kept coming back.
127
455910
1710
之后他们不断回来看这些机器人。
07:38
What we learned about gesture, navigation,
128
458746
2878
我们学到的关于动作示意、导航、
07:41
human interaction and expression can be applied to robots more generally.
129
461666
4880
人的互动和表情, 可以更广泛地应用于机器人。
07:47
Think of the following future scenario.
130
467171
2086
想想以下未来的情景。
07:49
A robot in a senior living facility,
131
469590
3170
养老院里的机器人,
07:52
waiting in a crowded cafe.
132
472802
2544
在拥挤的咖啡馆做服务员。
07:55
You sit at your table,
133
475805
1710
你坐在桌边,
07:57
you wave over to it, make a sipping gesture,
134
477557
2836
向它挥手,并做出啜饮的手势,
08:00
and the robot understands that you would like a glass of water.
135
480435
4212
机器人就明白你想要一杯水。
08:05
And when it drives over to you with its glass of water,
136
485106
2794
当它托着一杯水驶向你时,
08:07
it nods at the other people it passes
137
487942
3003
它会向遇到的人点头致意,
08:10
and pauses graciously to let them go by.
138
490987
3170
并友好地停下来让他们先行。
08:14
And when it arrives at your table, it places the glass gracefully down.
139
494615
4922
当它来到你的餐桌时, 它会优雅地将水杯放下。
08:19
And when you thank it, it nods in acknowledgment.
140
499537
3170
当你感谢它时,它会点头致意,
08:23
Maybe it even performs a small celebratory dance in place.
141
503124
5964
也许还会跳一支小型的庆祝舞。
08:29
We can choose to build and program robots that we welcome.
142
509130
5297
我们可以选择构建 和编程我们喜欢的机器人,
08:34
Robots that make us feel understood.
143
514886
2752
让我们感到被理解的机器人,
08:38
Robots that we want to spend time with.
144
518097
2586
我们想与之共度时光的机器人。
08:41
The promise of robots that will live and work amongst us
145
521684
3170
机器人将在我们中间 生活和工作的承诺
08:44
has existed for a very long time.
146
524896
2335
已经存在了很长时间。
08:47
From Karel Capek's 1920 play "Rossum's Universal Robots"
147
527648
5881
从卡雷尔·卡佩克 1920 年的 戏剧 《罗森的万能机器人》
08:53
to The Jetsons' Rosey the Robot
148
533571
2002
到《杰森一家》的机器人罗西,
08:55
to C-3PO of "Star Wars" and even Terminator,
149
535573
4463
再到《星球大战》中的 C-3PO, 甚至是《终结者》,
09:00
these stories have been deeply impactful in what we imagine and build.
150
540036
5297
这些故事对我们想象 和构建机器人产生了深远的影响。
09:06
Yes, we can have robots that do our chores.
151
546167
2878
是的,我们可以让机器人做家务,
09:10
But we can also have robots be our dance partners.
152
550046
4212
我们也可以让机器人作我们的舞伴。
09:14
Humans do a range of things,
153
554967
2336
人类从事各种各样的活动,
09:17
ceramics to diplomacy,
154
557303
2294
从陶艺到外交,
09:19
singing in choirs to vertical farming.
155
559597
2586
从合唱团到垂直种植。
09:22
Imagine a world where robots can do a myriad of things
156
562767
4588
想象一个世界,机器人可以做各种事情,
09:27
that are functional or creative,
157
567396
2753
无论是功能性或创造性的,
09:30
and everything in between.
158
570191
2461
或者介于两者之间的。
09:33
General-purpose robots are coming,
159
573736
3045
通用机器人即将问世,
09:36
and they will help us to address big demographic challenges
160
576823
3545
它们将帮助我们应对重大人口挑战,
09:40
like aging populations and massive labor shortages.
161
580368
3628
如人口老龄化和大规模劳动力短缺等。
09:44
And how they move will unlock
162
584747
4004
他们如何移动将决定
09:48
whether we let them into our homes
163
588751
2294
我们是否允许他们进入我们的家中,
09:51
and welcome them in our workplaces.
164
591045
1919
是否欢迎他们进入我们的工作场所。
09:53
Because we want to feel like we're still marvelously human
165
593714
4880
因为我们希望感觉到 自己仍然属于奇妙的人类,
09:58
in this increasingly AI world.
166
598594
2294
尤其在这个日益人工智能化的世界里。
10:01
And we can,
167
601848
1668
我们能做到的,
10:03
because we have the power to choreograph the future we want.
168
603558
5213
因为我们有能力编排我们想要的未来。
10:09
Thank you.
169
609230
1209
谢谢。
10:10
(Applause)
170
610481
3879
(掌声)
关于本网站

这个网站将向你介绍对学习英语有用的YouTube视频。你将看到来自世界各地的一流教师教授的英语课程。双击每个视频页面上显示的英文字幕,即可从那里播放视频。字幕会随着视频的播放而同步滚动。如果你有任何意见或要求,请使用此联系表与我们联系。

https://forms.gle/WvT1wiN1qDtmnspy7