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

譯者: Lilian Chiu 審譯者: Helen Chang

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

大約兩年半前，我看了 《雲端情人》這部電影。

And it features Samantha, a superintelligent form of AI

片中的莎曼珊 是一種超級人工智慧，

that cannot take physical form.

她無法有實體。

And because she can't appear in photographs,

因為莎曼珊無法出現在照片中，

Samantha decides to write a piece of music

她決定要寫一首音樂，

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

像照片一樣捕捉她人生的片刻。

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

我是音樂家也是工程師， 且在藝術家的家庭中長大，

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

我認為這音樂照片的點子強而有力。

And I decided to create an AI composer.

我決定創造一個人工智慧作曲家。

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

她的名字叫艾娃（AIVA）， 她是人工智慧，

that has learned the art of music composition

她已經讀過了歷史上 最偉大的三萬首曲子，

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

從中學會了音樂作曲的藝術。

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

從演算法的角度來看樂譜 就是這個樣子，

in a matrix-like representation.

用類似矩陣的方式來表示。

And here's what 30,000 scores,

如果把莫札特和貝多芬 等級的作曲家

written by the likes of Mozart and Beethoven,

所寫的三萬首曲子 放到一個畫面中，

look like in a single frame.

看起來就是這樣子。

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

艾娃會用深度學習的類神經網路 在樂譜中找模式。

And from a couple of bars of existing music,

從既有音樂的一些小節當中，

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

她能夠試著推論出下一個 出現的音符應該是什麼。

And once AIVA gets good at those predictions,

等艾娃很擅長預測之後，

it can actually build a set of mathematical rules

她就可以為那種音樂風格

建立一組數學規則，

for that style of music

in order to create its own original compositions.

用來創作她自己的原創曲。

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

就某種意義上來說， 我們人類也是這樣作曲的。

It's a trial-and-error process,

它是一個試誤的過程，

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

在這個過程中，我們不見得 總是都能讓所有音符完美。

But we can correct ourselves,

但我們能做自我修正，

either with our musical ear or our musical knowledge.

用的是用我們聽音樂的耳朵， 或是用我們的音樂知識。

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

但對艾娃來說，藝術家、 音樂家，和作曲家會需要

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

花數年甚至數十年時間 來學習的這個過程，

down to a couple of hours.

可縮短為幾小時。

But music is also a supersubjective art.

但音樂是非常主觀的藝術。

And we needed to teach AIVA

我們得要教導艾娃

how to compose the right music for the right person,

如何譜寫出對的音樂給對的人，

because people have different preferences.

因為大家的偏好都不相同。

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

為了達到這個目標，我們讓演算法 看過我們資料庫中每一首曲子的

for each score in our database.

至少三十種不同的類別標籤。

So those category labels are like mood

這些標籤包括像是調性、

or note density or composer style of a piece

音符密度、作曲家風格、

or the epoch during which it was written.

曲子被譜寫的時期。

And by seeing all this data,

看過這些資料之後，

AIVA can actually respond to very precise requirements.

艾娃就可以對非常精確的 需求來做出反應。

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

比如，最近我們接了一個專案計畫，

where we were commissioned to create a piece

我們受委託要創作一首作品，

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

它要能讓聽眾憶起科幻電影配樂。

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

創作出來的作品叫做《星星之中》，

and it was recorded with CMG Orchestra in Hollywood,

且錄音時請到好萊塢的 CMG 交響樂團來演奏，

under great conductor John Beal,

指揮是偉大的約翰貝爾，

and this is what they recorded, made by AIVA.

以下是錄音的結果，由艾娃創作。

(Music)

（音樂）

(Music ends)

（音樂結束）

What do you think?

你們覺得如何？

(Applause)

（掌聲）

Thank you.

謝謝。

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

所以，你們都看到了， 人工智慧可以創作出美麗的音樂，

and the best part of it

最棒的部分是，

is that humans can actually bring them to life.

人類可以賦予這些音樂生命。

And it's not the first time in history

這並不是史上第一次

that technology has augmented human creativity.

科技可以增強人類創意。

Live music was almost always used in silent films

默片通常都是用現場音樂，

to augment the experience.

目的就是要增強體驗。

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

但現場音樂的問題是 它的規模有限。

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

實在很難把整個 交響樂團塞到小戲院中，

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

且很難在世界上每個戲院 都提供現場音樂。

So when music recording was actually invented,

所以，當音樂錄音發明出來之後，

it allowed content creators, like film creators,

內容創作者，比如電影製作人，

to have prerecorded and original music

就可以用事前錄製好的原創音樂

tailored to each and every frame of their stories.

根據他們故事的每一個畫格 來客製化打造。

And that was really an enhancer of creativity.

那其實就是創意的強化器。

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

兩年半前，我看《雲端情人》時，

I thought to myself that personalized music

我心想，個人化的音樂

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

會為消費與創作音樂的方式 帶來下一個大改變。

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

因為現今我們有互動式的 內容，比如電玩遊戲，

that have hundreds of hours of interactive game plays,

雖然能播放數百小時的互動遊戲，

but only two hours of music, on average.

但平均只有兩小時的音樂。

And it means that the music loops and loops and loops

那就意味著音樂不斷重覆，

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

一而再再而三地播放， 很難非常身歷其境。

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

所以我們正在努力做的， 是要讓人工智慧能夠創作出

hundreds of hours of personalized music

數百小時的個人化音樂，

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

用在人類創意規模 受限的那些情況中。

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

且我們的目標不只是遊戲。

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

貝多芬為他的愛人寫過 一首作品《給愛麗絲》，

and imagine if we could bring back Beethoven to life.

想像一下，如果我們 能讓貝多芬復活。

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

如果他就坐在你旁邊， 為你的個性及人生故事

and your life story.

譜寫音樂。

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

或是，想像一下， 比如像是金恩博士這樣的人，

had a personalized AI composer.

有一個個人化的人工智慧作曲家，

Maybe then we would remember

也許我們對於 「我有一個夢」的記憶

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

就不只是偉大的演說，也是 很棒的音樂，我們歷史的一部分，

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

and capturing Dr. King's ideals.

捕捉到金恩博士的理念。

And this is our vision at AIVA:

這就是我們對於艾娃的遠景：

to personalize music so that each and every one of you

將音樂個人化，讓在座每一個人

and every individual in the world

以及世界上的每一個人

can have access to a personalized live soundtrack,

都能取得個人化的現場配樂，

based on their story and their personality.

依據他們的故事和個性所譜寫。

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

現在，我們一起在 TED 的這一刻， 就是我們人生故事的一部分。

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

所以，讓艾娃為這一刻 譜寫音樂是再適合不過了。

And that's exactly what we did.

我們也的確這麼做了。

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

我和我的團隊讓艾娃 採用偏向 TED 的風格，

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

再加上讓我們感到 敬畏感和奇妙感的音樂。

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

產出的成果叫做《驚奇年代》。

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

這不需要人工智慧也能想得出來。

(Laughter)

（笑聲）

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

接下來就要放給各位聽， 我感到非常驕傲，

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

如果可以的話， 請閉上眼睛，享受音樂。

Thank you very much.

非常謝謝。

(Music)

（音樂）

[The Age of Amazement Composed by AIVA]

【驚奇時代，艾娃作品】

(Music ends)

（音樂結束）

This was for all of you.

這是給在座所有人的。

Thank you.

謝謝。

(Applause)

（掌聲）