TED2018
Pierre Barreau: How AI could compose a personalized soundtrack to your life
皮埃尔 · 巴罗: 为您的人生量身配乐的人工智能
Filmed:
Readability: 4.1
1,460,825 views
AIVA 读过三万余首的乐谱，是一款可以作曲的人工智能。在这场精彩的演讲兼演示会上，皮埃尔 · 巴罗分享了AIVA的作品，也和大家分享了他的梦想：根据每个人的心情脾性制作出属于我们自己的原创现场配乐。
Pierre Barreau - AI and music entrepreneur
Pierre Barreau is an entrepreneur, computer scientist, composer and director who loves to create impactful things. Full bio
Pierre Barreau is an entrepreneur, computer scientist, composer and director who loves to create impactful things. Full bio
Double-click the English transcript below to play the video.
00:12
About two and a half years ago,
I watched this movie called "Her."
I watched this movie called "Her."
0
913
3428
大概两年半前，我看了《她》这部电影。
电影里的“女主角”沙曼塔
是极度聪明的人工智能，
是极度聪明的人工智能，
00:16
And it features Samantha,
a superintelligent form of AI
a superintelligent form of AI
1
4365
5127
但却并不拥有实体肉身。
00:21
that cannot take physical form.
2
9516
1760
00:23
And because she can't
appear in photographs,
appear in photographs,
3
11858
2230
她没办法拥有自己的照片，
00:26
Samantha decides to write a piece of music
4
14112
2016
所以她决定写首曲子
来捕捉她生命中的片段，正如拍照一般。
00:28
that will capture a moment of her life
just like a photograph would.
just like a photograph would.
5
16152
3466
00:32
As a musician and an engineer,
and someone raised in a family of artists,
and someone raised in a family of artists,
6
20953
4408
我出生于艺术世家，
现在是一名音乐家兼工程师，
现在是一名音乐家兼工程师，
我觉得音乐摄影这个想法很强大。
00:37
I thought that this idea of musical
photographs was really powerful.
photographs was really powerful.
7
25385
4176
00:41
And I decided to create an AI composer.
8
29955
2868
所以我决定制造一款会作曲的人工智能。
00:44
Her name is AIVA,
and she's an artificial intelligence
and she's an artificial intelligence
9
32847
3488
这款人工智能的名字就叫做“AIVA”。
00:48
that has learned the art
of music composition
of music composition
10
36359
2393
通过阅读历史上最优秀的
三万多首乐谱，她学会了作曲的艺术。
00:50
by reading over 30,000 scores
of history's greatest.
of history's greatest.
11
38776
2785
00:54
So here's what one score
looks like to the algorithm
looks like to the algorithm
12
42165
2456
如果以这种矩阵似的形式来表现的话，
00:56
in a matrix-like representation.
13
44645
2202
在我们的算法看来，
乐谱就是长这样子的。
乐谱就是长这样子的。
00:58
And here's what 30,000 scores,
14
46871
2246
而如果把莫扎特、贝多芬等人
写的三万首乐谱同置于一个框架之中，
01:01
written by the likes
of Mozart and Beethoven,
of Mozart and Beethoven,
15
49141
2096
看起来就是这个样子了。
01:03
look like in a single frame.
16
51261
2031
利用深层神经网络，AIVA
开始从这些乐谱中找出规律。
开始从这些乐谱中找出规律。
01:07
So, using deep neural networks,
AIVA looks for patterns in the scores.
AIVA looks for patterns in the scores.
17
55609
4340
从现有音乐中的几个小节，
01:12
And from a couple of bars
of existing music,
of existing music,
18
60260
3563
01:15
it actually tries to infer what notes
should come next in those tracks.
should come next in those tracks.
19
63847
3708
她会推测乐曲中接下来该有什么音符。
一旦预测变得越来越准确，
01:19
And once AIVA gets good
at those predictions,
at those predictions,
20
67887
2498
01:22
it can actually build a set
of mathematical rules
of mathematical rules
21
70409
3613
AIVA 就开始为该种曲风建立起
一套数学规则，
01:26
for that style of music
22
74046
1166
01:27
in order to create
its own original compositions.
its own original compositions.
23
75236
2788
并以此写出她自己的原创歌曲。
01:30
And in a way, this is kind of
how we, humans, compose music, too.
how we, humans, compose music, too.
24
78490
3709
在某种程度上，这其实也是
我们人类作曲的方式。
我们人类作曲的方式。
这是一种反复试验的过程，
01:34
It's a trial-and-error process,
25
82561
1492
01:36
during which we may not
get the right notes all the time.
get the right notes all the time.
26
84077
3008
我们未必每次都找得到最合适的音符，
但我们可以用我们的
01:39
But we can correct ourselves,
27
87109
1389
01:40
either with our musical ear
or our musical knowledge.
or our musical knowledge.
28
88522
3293
音乐知识或耳朵来自我修正。
01:45
But for AIVA, this process
is taken from years and years of learning,
is taken from years and years of learning,
29
93495
4143
这需要长时间的学习，
但 AIVA 把段漫长的学习时间，
但 AIVA 把段漫长的学习时间，
01:49
decades of learning as an artist,
as a musician and a composer,
as a musician and a composer,
30
97662
3262
也就是成为艺术家、音乐家、
作曲家所需的那几十年学习，
作曲家所需的那几十年学习，
01:52
down to a couple of hours.
31
100948
1733
浓缩成只有几个小时。
01:55
But music is also a supersubjective art.
32
103030
2880
当然音乐也是个极为主观的艺术。
01:57
And we needed to teach AIVA
33
105934
1564
所以我们需要教导 AIVA
01:59
how to compose the right music
for the right person,
for the right person,
34
107522
2445
如何为不同的人制作不同的音乐，
毕竟每个人的喜好都有所不同。
02:01
because people have different preferences.
35
109991
2141
02:04
And to do that, we show to the algorithm
over 30 different category labels
over 30 different category labels
36
112156
4190
为此，我们为数据库里的每一首乐谱
标上了三十余种不同的类别标签，
标上了三十余种不同的类别标签，
02:08
for each score in our database.
37
116370
1945
并展示给我们的算法。
02:10
So those category labels are like mood
38
118339
2516
这些类别标签包括音乐的气氛、
02:12
or note density or composer
style of a piece
style of a piece
39
120879
2920
音符的密度、曲风
02:15
or the epoch during which it was written.
40
123823
2524
或作品的年代等等。
02:18
And by seeing all this data,
41
126371
1935
而通过这些数据，
02:20
AIVA can actually respond
to very precise requirements.
to very precise requirements.
42
128330
3056
AIVA 就可以满足一些很精确的要求。
02:23
Like the ones, for example,
we had for a project recently,
we had for a project recently,
43
131966
3103
比如，我们最近接了一个项目，
02:27
where we were commissioned
to create a piece
to create a piece
44
135093
3035
要创作一首会让人想起
02:30
that would be reminiscent
of a science-fiction film soundtrack.
of a science-fiction film soundtrack.
45
138152
3027
科幻电影配乐的曲子。
02:33
And the piece that was created
is called "Among the Stars"
is called "Among the Stars"
46
141839
4500
于是我们就作了一首曲子，
叫《众星之中》，
叫《众星之中》，
后来由指挥家约翰 · 比尔领衔的
好莱坞 CMG 管弦乐队
好莱坞 CMG 管弦乐队
02:38
and it was recorded
with CMG Orchestra in Hollywood,
with CMG Orchestra in Hollywood,
47
146363
3317
02:41
under great conductor John Beal,
48
149704
1714
录制了这首曲子。
这就是他们录制好的，AIVA 的作品。
02:43
and this is what they
recorded, made by AIVA.
recorded, made by AIVA.
49
151442
3067
02:47
(Music)
50
155657
7000
（音乐）
03:30
(Music ends)
51
198196
2190
（音乐结束）
03:34
What do you think?
52
202719
1207
你们觉得怎样？
03:35
(Applause)
53
203950
4380
（掌声）
03:40
Thank you.
54
208354
1150
谢谢。
03:42
So, as you've seen, AI can create
beautiful pieces of music,
beautiful pieces of music,
55
210001
4055
你们都看到了，人工智能
可以写出漂亮的乐章，
可以写出漂亮的乐章，
03:46
and the best part of it
56
214080
1803
而最棒的是，
人类可以把这音乐演奏出来。
03:47
is that humans can actually
bring them to life.
bring them to life.
57
215907
2581
03:51
And it's not the first time in history
58
219506
1838
科技可以扩展人类的创造力，
03:53
that technology has augmented
human creativity.
human creativity.
59
221368
3344
这其实并不新鲜。
03:56
Live music was almost always
used in silent films
used in silent films
60
224736
3072
以前的默片就经常会有现场伴奏，
03:59
to augment the experience.
61
227832
1666
好带给观众更好的观影经验。
04:01
But the problem with live music
is that it didn't scale.
is that it didn't scale.
62
229522
3182
但现场伴奏有个问题：
它没办法适应不同场景的需要。
它没办法适应不同场景的需要。
04:04
It's really hard to cram a full symphony
into a small theater,
into a small theater,
63
232728
3429
要把整个交响乐团塞进一个
小剧院里面已经很不容易，
小剧院里面已经很不容易，
04:08
and it's really hard to do that
for every theater in the world.
for every theater in the world.
64
236181
2991
要让全世界所有的剧院
都有交响乐团，那就更难了。
都有交响乐团，那就更难了。
04:11
So when music recording
was actually invented,
was actually invented,
65
239196
2849
所以当人类发明了录音，
04:14
it allowed content creators,
like film creators,
like film creators,
66
242069
2349
所有的内容创作者，比如电影创作者，
04:16
to have prerecorded and original music
67
244442
2786
便可以为他们故事里的每一个画面
04:19
tailored to each and every frame
of their stories.
of their stories.
68
247252
2873
量身定制原创音乐，并且预先录制好。
这无疑强化了我们的创造力。
04:22
And that was really
an enhancer of creativity.
an enhancer of creativity.
69
250149
2626
04:26
Two and a half years ago,
when I watched this movie "Her,"
when I watched this movie "Her,"
70
254617
3286
两年半前，我在看《她》这部电影的时候，
04:29
I thought to myself
that personalized music
that personalized music
71
257927
2907
不禁心想到，个人化的音乐
将会是我们未来创作和消费
音乐时最重大的一个改变。
音乐时最重大的一个改变。
04:32
would be the next single biggest change
in how we consume and create music.
in how we consume and create music.
72
260858
4417
04:38
Because nowadays, we have
interactive content, like video games,
interactive content, like video games,
73
266187
4222
因为如今，我们有许多
互动性的内容，比如说电子游戏，
互动性的内容，比如说电子游戏，
04:42
that have hundreds of hours
of interactive game plays,
of interactive game plays,
74
270433
3015
里面有上百个小时的互动式游戏内容，
04:45
but only two hours of music, on average.
75
273472
2113
平均却只配有两个小时的音乐。
04:47
And it means that the music
loops and loops and loops
loops and loops and loops
76
275609
2492
也就是说，游戏的音乐
是不停地循环重播，
是不停地循环重播，
而且没法让人真正沉浸其中。
04:50
over and over again,
and it's not very immersive.
and it's not very immersive.
77
278125
2332
04:52
So what we're working on
is to make sure that AI can compose
is to make sure that AI can compose
78
280467
3951
我们目前正在努力让人工智能创作出
04:56
hundreds of hours of personalized music
79
284442
2302
上百个小时的个性化音乐，
04:58
for those use cases
where human creativity doesn't scale.
where human creativity doesn't scale.
80
286768
3866
以填补那些人类创造力无法涵盖的领域。
05:03
And we don't just want
to do that for games.
to do that for games.
81
291363
2206
而且我们想做的绝不限于电子游戏。
05:06
Beethoven actually wrote a piece
for his beloved, called "Für Elise,"
for his beloved, called "Für Elise,"
82
294657
4428
贝多芬曾经给他的挚爱
写过一首《致爱丽丝》，
写过一首《致爱丽丝》，
05:11
and imagine if we could
bring back Beethoven to life.
bring back Beethoven to life.
83
299109
3651
想像一下，如果我们可以让贝多芬重生，
05:14
And if he was sitting next to you,
composing a music for your personality
composing a music for your personality
84
302784
5230
让他坐在你身旁，依据
你的性情和人生故事，
你的性情和人生故事，
专门谱写一首曲子。
05:20
and your life story.
85
308038
1333
05:22
Or imagine if someone like
Martin Luther King, for example,
Martin Luther King, for example,
86
310632
2779
或者你也可以想象一下某个人，
比如说马丁 · 路德 · 金，
比如说马丁 · 路德 · 金，
如果他拥有一款个人化的、
会作曲的人工智能。
会作曲的人工智能。
05:25
had a personalized AI composer.
87
313435
2039
05:27
Maybe then we would remember
88
315498
1373
那《我有一个梦想》对我们而言，
05:28
"I Have a Dream" not only
as a great speech,
as a great speech,
89
316895
2056
除了是一次伟大的演讲，
或许也会是我们历史上一首可以体现
05:30
but also as a great piece of music,
part of our history,
part of our history,
90
318975
2626
金博士理想的伟大音乐作品。
05:33
and capturing Dr. King's ideals.
91
321625
2133
05:36
And this is our vision at AIVA:
92
324069
1889
我们对 AIVA 的期许就是：
05:37
to personalize music
so that each and every one of you
so that each and every one of you
93
325982
2578
让她为你们在座的每一位，
甚至世界上的每一个人
甚至世界上的每一个人
量身创作音乐。
05:40
and every individual in the world
94
328584
1649
让每个人可以根据
自己的人生故事和性格特点，
自己的人生故事和性格特点，
05:42
can have access to a personalized
live soundtrack,
live soundtrack,
95
330257
3074
05:45
based on their story
and their personality.
and their personality.
96
333355
2505
找到专属于自己的现场配乐。
05:49
So this moment here together at TED
is now part of our life story.
is now part of our life story.
97
337915
4079
现在我们聚集在 TED 会场，
这也是我们人生经历的一部分，
这也是我们人生经历的一部分，
05:54
So it only felt fitting that AIVA
would compose music for this moment.
would compose music for this moment.
98
342018
4169
如果 AIVA 可以为此谱曲一首的话，
那就再适合不过了。
那就再适合不过了。
我们还真的就这么做了。
05:58
And that's exactly what we did.
99
346674
2444
06:01
So my team and I worked on biasing AIVA
on the style of the TED jingle,
on the style of the TED jingle,
100
349436
5022
我和团队让 AIVA
模拟 TED 的广告曲风，
模拟 TED 的广告曲风，
06:06
and on music that makes us feel
a sense of awe and wonder.
a sense of awe and wonder.
101
354482
3444
以及那些会让我们惊叹
或心生敬畏的音乐风格。
或心生敬畏的音乐风格。
06:09
And the result is called
"The Age of Amazement."
"The Age of Amazement."
102
357950
3913
于是我们就有了这首《惊奇的年代》。
06:13
Didn't take an AI to figure that one out.
103
361887
2396
当然这曲名不是人工智能想出来的。
06:16
(Laughter)
104
364307
1150
（笑声）
06:18
And I couldn't be more proud
to show it to you,
to show it to you,
105
366152
2402
今天可以给大家带来这首曲子，
我感到非常自豪，
我感到非常自豪，
06:20
so if you can, close your eyes
and enjoy the music.
and enjoy the music.
106
368578
2428
可以的话，请闭上双眼，
好好享受一下这音乐吧。
好好享受一下这音乐吧。
06:23
Thank you very much.
107
371030
1333
谢谢。
06:25
(Music)
108
373537
4195
（音乐）
06:35
[The Age of Amazement
Composed by AIVA]
Composed by AIVA]
109
383176
2667
【《惊奇的年代》
作曲：AIVA】
作曲：AIVA】
08:19
(Music ends)
110
487649
1151
（音乐结束）
谨以此曲献给大家。
08:20
This was for all of you.
111
488824
1373
08:22
Thank you.
112
490221
1166
谢谢。
08:23
(Applause)
113
491411
4573
（掌声）
Translated by Ch'ng Tsu Pang
ABOUT THE SPEAKER
Pierre Barreau - AI and music entrepreneurPierre Barreau is an entrepreneur, computer scientist, composer and director who loves to create impactful things.
Why you should listen
Pierre Barreau is the CEO of AIVA, an artificial intelligence that composes music. As Barreau writes: "Ever since I was a child, I was always passionate about building things. This drive to make the imaginary real is what pushed me to create a company as soon as I got out of university. The seed of the idea for AIVA was planted once I saw the science fiction movie Her, where an AI composes a beautiful piece of piano that captures the essence of the moment she is currently living. This simple concept of creating an artificial intelligence capable of composing personalized music is how I believe that I can make the world a slightly more magical place and meaningfully support the storytelling of content creators through music."
Pierre was nominated for a Gold Panda Award in the "Best Director" category when he was 15 years old for a four-film documentary series, One Night in the Cosmos.
Pierre Barreau | Speaker | TED.com