How AI could compose a personalized soundtrack to your life | Pierre Barreau

翻译人员: Ch'ng Tsu Pang 校对人员: Yolanda Zhang

About two and a half years ago, I watched this movie called "Her."

大概两年半前，我看了《她》这部电影。

And it features Samantha, a superintelligent form of AI

电影里的“女主角”沙曼塔 是极度聪明的人工智能，

that cannot take physical form.

但却并不拥有实体肉身。

And because she can't appear in photographs,

她没办法拥有自己的照片，

Samantha decides to write a piece of music

所以她决定写首曲子

that will capture a moment of her life just like a photograph would.

来捕捉她生命中的片段，正如拍照一般。

As a musician and an engineer, and someone raised in a family of artists,

我出生于艺术世家， 现在是一名音乐家兼工程师，

I thought that this idea of musical photographs was really powerful.

我觉得音乐摄影这个想法很强大。

And I decided to create an AI composer.

所以我决定制造一款会作曲的人工智能。

Her name is AIVA, and she's an artificial intelligence

这款人工智能的名字就叫做“AIVA”。

that has learned the art of music composition

通过阅读历史上最优秀的

by reading over 30,000 scores of history's greatest.

三万多首乐谱，她学会了作曲的艺术。

So here's what one score looks like to the algorithm

如果以这种矩阵似的形式来表现的话，

in a matrix-like representation.

在我们的算法看来， 乐谱就是长这样子的。

And here's what 30,000 scores,

而如果把莫扎特、贝多芬等人

written by the likes of Mozart and Beethoven,

写的三万首乐谱同置于一个框架之中，

look like in a single frame.

看起来就是这个样子了。

So, using deep neural networks, AIVA looks for patterns in the scores.

利用深层神经网络，AIVA 开始从这些乐谱中找出规律。

And from a couple of bars of existing music,

从现有音乐中的几个小节，

it actually tries to infer what notes should come next in those tracks.

她会推测乐曲中接下来该有什么音符。

And once AIVA gets good at those predictions,

一旦预测变得越来越准确，

it can actually build a set of mathematical rules

AIVA 就开始为该种曲风建立起

一套数学规则，

for that style of music

in order to create its own original compositions.

并以此写出她自己的原创歌曲。

And in a way, this is kind of how we, humans, compose music, too.

在某种程度上，这其实也是 我们人类作曲的方式。

It's a trial-and-error process,

这是一种反复试验的过程，

during which we may not get the right notes all the time.

我们未必每次都找得到最合适的音符，

But we can correct ourselves,

但我们可以用我们的

either with our musical ear or our musical knowledge.

音乐知识或耳朵来自我修正。

But for AIVA, this process is taken from years and years of learning,

这需要长时间的学习， 但 AIVA 把段漫长的学习时间，

decades of learning as an artist, as a musician and a composer,

也就是成为艺术家、音乐家、 作曲家所需的那几十年学习，

down to a couple of hours.

浓缩成只有几个小时。

But music is also a supersubjective art.

当然音乐也是个极为主观的艺术。

And we needed to teach AIVA

所以我们需要教导 AIVA

how to compose the right music for the right person,

如何为不同的人制作不同的音乐，

because people have different preferences.

毕竟每个人的喜好都有所不同。

And to do that, we show to the algorithm over 30 different category labels

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，

for each score in our database.

并展示给我们的算法。

So those category labels are like mood

这些类别标签包括音乐的气氛、

or note density or composer style of a piece

音符的密度、曲风

or the epoch during which it was written.

或作品的年代等等。

And by seeing all this data,

而通过这些数据，

AIVA can actually respond to very precise requirements.

AIVA 就可以满足一些很精确的要求。

Like the ones, for example, we had for a project recently,

比如，我们最近接了一个项目，

where we were commissioned to create a piece

要创作一首会让人想起

that would be reminiscent of a science-fiction film soundtrack.

科幻电影配乐的曲子。

And the piece that was created is called "Among the Stars"

于是我们就作了一首曲子， 叫《众星之中》，

and it was recorded with CMG Orchestra in Hollywood,

后来由指挥家约翰 · 比尔领衔的 好莱坞 CMG 管弦乐队

under great conductor John Beal,

录制了这首曲子。

and this is what they recorded, made by AIVA.

这就是他们录制好的，AIVA 的作品。

(Music)

（音乐）

(Music ends)

（音乐结束）

What do you think?

你们觉得怎样？

(Applause)

（掌声）

Thank you.

谢谢。

So, as you've seen, AI can create beautiful pieces of music,

你们都看到了，人工智能 可以写出漂亮的乐章，

and the best part of it

而最棒的是，

is that humans can actually bring them to life.

人类可以把这音乐演奏出来。

And it's not the first time in history

科技可以扩展人类的创造力，

that technology has augmented human creativity.

这其实并不新鲜。

Live music was almost always used in silent films

以前的默片就经常会有现场伴奏，

to augment the experience.

好带给观众更好的观影经验。

But the problem with live music is that it didn't scale.

但现场伴奏有个问题： 它没办法适应不同场景的需要。

It's really hard to cram a full symphony into a small theater,

要把整个交响乐团塞进一个 小剧院里面已经很不容易，

and it's really hard to do that for every theater in the world.

要让全世界所有的剧院 都有交响乐团，那就更难了。

So when music recording was actually invented,

所以当人类发明了录音，

it allowed content creators, like film creators,

所有的内容创作者，比如电影创作者，

to have prerecorded and original music

便可以为他们故事里的每一个画面

tailored to each and every frame of their stories.

量身定制原创音乐，并且预先录制好。

And that was really an enhancer of creativity.

这无疑强化了我们的创造力。

Two and a half years ago, when I watched this movie "Her,"

两年半前，我在看《她》这部电影的时候，

I thought to myself that personalized music

不禁心想到，个人化的音乐

would be the next single biggest change in how we consume and create music.

将会是我们未来创作和消费 音乐时最重大的一个改变。

Because nowadays, we have interactive content, like video games,

因为如今，我们有许多 互动性的内容，比如说电子游戏，

that have hundreds of hours of interactive game plays,

里面有上百个小时的互动式游戏内容，

but only two hours of music, on average.

平均却只配有两个小时的音乐。

And it means that the music loops and loops and loops

也就是说，游戏的音乐 是不停地循环重播，

over and over again, and it's not very immersive.

而且没法让人真正沉浸其中。

So what we're working on is to make sure that AI can compose

我们目前正在努力让人工智能创作出

hundreds of hours of personalized music

上百个小时的个性化音乐，

for those use cases where human creativity doesn't scale.

以填补那些人类创造力无法涵盖的领域。

And we don't just want to do that for games.

而且我们想做的绝不限于电子游戏。

Beethoven actually wrote a piece for his beloved, called "Für Elise,"

贝多芬曾经给他的挚爱 写过一首《致爱丽丝》，

and imagine if we could bring back Beethoven to life.

想像一下，如果我们可以让贝多芬重生，

And if he was sitting next to you, composing a music for your personality

让他坐在你身旁，依据 你的性情和人生故事，

专门谱写一首曲子。

and your life story.

Or imagine if someone like Martin Luther King, for example,

或者你也可以想象一下某个人， 比如说马丁 · 路德 · 金，

had a personalized AI composer.

如果他拥有一款个人化的、 会作曲的人工智能。

Maybe then we would remember

那《我有一个梦想》对我们而言，

"I Have a Dream" not only as a great speech,

除了是一次伟大的演讲，

but also as a great piece of music, part of our history,

或许也会是我们历史上一首可以体现

and capturing Dr. King's ideals.

金博士理想的伟大音乐作品。

And this is our vision at AIVA:

我们对 AIVA 的期许就是：

to personalize music so that each and every one of you

让她为你们在座的每一位， 甚至世界上的每一个人

and every individual in the world

量身创作音乐。

can have access to a personalized live soundtrack,

让每个人可以根据 自己的人生故事和性格特点，

based on their story and their personality.

找到专属于自己的现场配乐。

So this moment here together at TED is now part of our life story.

现在我们聚集在 TED 会场， 这也是我们人生经历的一部分，

So it only felt fitting that AIVA would compose music for this moment.

如果 AIVA 可以为此谱曲一首的话， 那就再适合不过了。

And that's exactly what we did.

我们还真的就这么做了。

So my team and I worked on biasing AIVA on the style of the TED jingle,

我和团队让 AIVA 模拟 TED 的广告曲风，

and on music that makes us feel a sense of awe and wonder.

以及那些会让我们惊叹 或心生敬畏的音乐风格。

And the result is called "The Age of Amazement."

于是我们就有了这首《惊奇的年代》。

Didn't take an AI to figure that one out.

当然这曲名不是人工智能想出来的。

(Laughter)

（笑声）

And I couldn't be more proud to show it to you,

今天可以给大家带来这首曲子， 我感到非常自豪，

so if you can, close your eyes and enjoy the music.

可以的话，请闭上双眼， 好好享受一下这音乐吧。

Thank you very much.

谢谢。

(Music)

（音乐）

[The Age of Amazement Composed by AIVA]

【《惊奇的年代》 作曲：AIVA】

(Music ends)

（音乐结束）

This was for all of you.

谨以此曲献给大家。

Thank you.

谢谢。

(Applause)

（掌声）