Can we learn to talk to sperm whales? | David Gruber | TED

譯者: Lilian Chiu 審譯者: Helen Chang

You are about to hear the sounds of the largest-toothed predator

你將要聽到的聲音，是來自 地球上牙齒最大的捕食性動物：

on the planet:

這種動物比校車還大，

an animal bigger than a school bus

with perhaps the most sophisticated form of communication

其溝通方式可能是 有史以來最複雜的。

that has ever existed.

(Video: whale clicking)

（影片：鯨魚發出卡嗒聲）

These are the sounds of the mighty sperm whale,

這些是巨型抹香鯨發出的聲音，

a fellow mammal that can dive almost a mile,

這種哺乳類夥伴可以下潛近一哩深，

hold its breath for more than an hour

閉氣超過一小時，

and lives in these amazingly complex, matriarchal societies.

生活在這種極複雜的母系社會中。

These clicks you heard,

你聽到的這些卡嗒聲 叫做「尾聲音節」，

called codas,

只是我們所了解的抹香鯨 溝通的其中一個面向。

are just a facet of what we know of their communication.

We know these animals are communicating,

我們知道這些動物正在溝通，

we just don't yet know what they're saying.

我們只是還不知道牠們在說什麼。

Project CETI aims to find out.

CETI 計畫的目標 就是解答這個問題。

Over the next five years,

在接下來的五年間，

our team of AI specialists,

我們的團隊，包括 人工智慧專家、機器人專家、

roboticists, linguists

語言學家，及海洋生物學家，

and marine biologists

aim to use the most cutting-edge technologies

打算要用最尖端的技術 與其他的物種接觸，

to make contact with another species,

and hopefully communicate back.

希望牠們會回應溝通。

We believe that by listening deeply to nature,

我們相信藉由深刻傾聽大自然，

we can change our perspective of ourselves

我們就能改變我們對自己的看法，

and reshape our relationship with all life on this planet.

並重新塑造

我們與地球上所有生命的關係。

This of course seems like an impossible goal.

當然，這似乎是個 不可能達成的目標。

People have been trying to make contact with other animals

數百年來，人類一直 在嘗試和動物溝通。

for hundreds of years.

How could we do what others could not,

我們要怎麼做到別人做不到的事？

especially given that I'm sitting here on my couch in New York City

尤其是，我還在紐約市裡， 坐在我家的沙發上，

in the middle of a pandemic and protests?

現在還正值疫情和抗議？

I've spent the last 20 years as a marine biologist and oceanographer,

以海洋生物學家和海洋學家的 身分，我在過去二十年間，

studying the ocean from all different perspectives,

從各種不同的角度研究海洋，

from microbes to sharks.

從微生物到鯊魚。

I've assembled interdisciplinary teams

我組織了跨領域團隊， 打造了第一個鯊魚眼攝影機，

that have built the first shark-eye camera

to see the world from a shark's perspective,

讓我們從鯊魚的視角看世界，

and have collaborated with engineers

我也與工程師合作，

to design robots so gentle that they don't even stress a jellyfish.

設計非常溫柔的機器人，

溫柔到甚至不會驚動水母。

But it wasn't until 2018

但直到 2018 年，

when I was on fellowship

我在拉德克利夫研究所 擔任研究員時，

at the Radcliffe Institute for Advanced Study

that I realized that perhaps the best way to understand the ocean

我才了解，也許，了解海洋 及其居民的最佳方式，

and its inhabitants

不是僅透過牠們的視角看世界，

wasn't just by seeing the world through their eyes,

but by listening --

而是要傾聽——

by really, deeply listening.

真正的、深刻的傾聽。

I became interested in sperm whales when I heard their sounds.

我聽見抹香鯨的聲音後， 便對牠們十分感興趣。

They sounded like they were coming from another universe;

牠們聽起來就像是來自另一個宇宙；

a siren song being broadcast from the darkest reaches of the sea.

從海洋最黑暗的深處 傳出來的美妙歌曲。

These weren't the typical harmonious whale songs

這些聲音並不是我熟悉的 典型和諧鯨魚歌聲。

that I had been accustomed to.

These sounded more like digital data transfer.

它們聽起來更像數位資料傳輸。

We assembled the future Project CETI team

我們為未來的計畫 CETI 組織了一個團隊，

and began discussing how to use the most advanced technologies

開始討論要如何運用 最先進的科技與鯨魚溝通。

to communicate with whales.

One of the principal conclusions

主要的結論之一是： 機器學習很有機會

was that machine learning had a really good chance

of understanding the patterns of sperm whale communication.

了解抹香鯨溝通的模式。

And the time to apply these technologies was now.

而現在該是使用這些技術的時候了。

Cracking the interspecies communication code

破解物種之間的溝通密碼

didn't just seem possible,

不僅是有可能的，

it almost seemed inevitable.

似乎還是無可避免的。

But how can analyzing patterns help us converse with whales

但，分析模式怎麼能協助我們 與鯨魚或其他動物交談？

and other animals?

Well, step one is to understand the elements of sperm whale communication.

這個嘛，第一步是要了解 抹香鯨溝通的元素。

你聽到這些的尾聲音節 似乎不是我們所知的句子，

These codas you heard don't appear to be sentences as we know them,

but there's clear structure in how these animals communicate.

但這些動物溝通的方式 有著清楚的架構。

Sperm whales send codas back and forth to each other

抹香鯨之間會來回傳送 一連串尾聲音節，

in sequences,

and there are regional dialects like British and Australian accents.

牠們也有區域性的方言， 就像英國和澳洲口音。

This is exactly why machine learning is such a powerful tool.

這正是為什麼機器學習 是如此強大的工具。

These approaches analyze patterns in relationship and map meaning to them.

這些方法可以用關聯性來分析模式，

並找到對應的意義。

Just a few years ago, scientists used machine learning

幾年前，科學家用機械學習

to translate between two totally unknown human languages.

將兩種完全未知的 人類語言互相翻譯。

Not by using a Rosetta Stone or a dictionary,

不是用羅塞塔語言學習軟體或字典，

but by mapping them on patterns in higher-dimensional space.

而是把它們投射到 更高維度空間中的模式上。

But for machine learning to work effectively,

但若要讓機械學習能發揮效果，

it needs data --

就需要資料——

it needs lots and lots of data.

需要很多很多資料。

In the past half-century,

在前五十年間，

marine researchers have painstakingly collected

海洋研究者煞費苦心地收集了

and hand annotated just a few thousand sperm whale vocalizations,

幾千筆抹香鯨的發聲資料 並手動加上註記，

but in order to learn sperm whale communication,

但，為了學習抹香鯨的溝通， 我們需要收集到

we'll need to collect millions,

數百萬筆，甚至數千萬筆，

if not tens of millions

of carefully annotated sperm whale vocalizations

加上仔細註記的抹香鯨發聲資料，

搭配和行為的關聯。

correlated with behaviors.

We'll do it with noninvasive, autonomous, free-swimming robots,

我們用非侵略性、自動化、 自由游動的機器人、

aerial-aquatic drones,

空中－水底無人機、

bottom-mounted hydrophone arrays

底部水下麥克風陣列 等工具來做這項工作。

and more.

We'll work with our close partners at the Dominica Sperm Whale Project

我們將會與多明尼克抹香鯨 計畫的密切夥伴合作，

以涵蓋二十平方公里的區域，

to cover a 20-square-kilometer area

that is frequented by over 25 well-known families of sperm whales.

有至少二十五個知名的抹香鯨 家庭經常出入這塊區域。

We're going to put specific focus on the interactions of mothers and calfs,

我們將把焦點明確放在 母鯨和幼鯨的互動上，

training our machine learning algorithms

訓練我們的機械學習演算法，

to learn whale language from the bottom up.

由下而上學習鯨魚的語言。

All this data we'll have sent through a pipeline

這些資料會透過一條管道線傳輸，

and analyzed by the Project CETI translation team.

交由 CETI 計畫的翻譯團隊來分析。

The raw audio and context data will go through our machine learning engine

原始聲音以及情境資料

則會交給我們的機械學習引擎，

where it's going to be first sorted by structure.

首先會依結構來分類。

The linguistics team will then search for things like syntax

語言學團隊接著會試著尋找 如語法及時間移位。

and time displacement.

For example,

比如，

if we find an event where a whale was talking about something yesterday,

如果哪一次我們發現鯨魚 在談昨天發生的事，

that alone would be a major finding,

光這一點就是重大的發現，

something that has thus far only been shown in humans.

因為到目前只有人類可以做到。

And once we've really mastered listening,

一旦我們真正精通了傾聽，

we're going to try really carefully to talk back

我們會試著非常謹慎地回話，

even on the most simplistic level.

即使是過度簡單的層級。

Finally, Project CETI will build an open-source platform

最後，CETI 計畫將會建造 一個開放原始碼的平台，

where we will make our data sets available to the public,

我們的把我們的資料集 公開給大眾使用，

encouraging the global community

鼓勵全球大眾

to come along on this journey for understanding.

一同參與這趟「了解」之旅。

These animals could be the most intelligent beings on this planet.

這些動物可能是地球上 最有智慧的生物。

They have a neocortex and spindle cells --

牠們有大腦新皮質以及梭狀細胞——

structure that in humans control our higher thoughts,

人類身上的這種結構 是用來控制更高階的思想、

emotions, memory, language and love.

情緒、記憶、語言，及愛。

And all the platforms that we develop can be cross-applied to other animals:

而我們開發的所有平台

都能交叉應用到其他 動物身上，如大象、鳥類、

to elephants, birds,

primates, dolphins --

靈長類、海豚——

essentially any animal.

基本上，任何動物皆可。

In the late 1960s,

1960 年代末，

我們的團隊成員羅傑佩恩

our team member, Roger Payne, discovered that whales sing.

發現鯨魚會唱歌。 （錄音：鯨魚唱歌。）

(Recording: whale singing)

這項發現促成了「拯救鯨魚」運動，

A finding that sparked the Save the Whales movement

led to the end of large-scale whaling

導致大規模捕鯨的終止，

and prevented several whale species from extinction

並預防數個鯨魚物種絕種，

just by showing that whales sing.

就只靠著展示出鯨魚會唱歌。

Imagine if we could understand what they're saying.

想像一下，如果我們 能了解牠們在說什麼。

Now is the time to open this larger dialogue.

現在該是開啟 更大規模對話的時候了。

Now is the time to listen deeply

現在該是深刻傾聽的時候了，

and show these magical animals,

讓這些魔法般的動物及彼此見識一下

and each other,

newfound respect.

新發現的尊重。

Thank you.

謝謝。