bees have one of the biggest jobs in the world.

蜜蜂的身形嬌小

Hi, I'm Katharina!

卻肩負世上最重要的工作之一

And I am Frederic.

- 大家好，我是 Katharina！

We use machine learning

- 我是 Frederic

to better understand how nature works.

- 我們運用機器學習技術

And we're doing that by observing bees.

探究自然界的運作機制

Bees have a symbiotic relationship with nature

方法是觀察蜜蜂的活動

Plants depend on the pollination

- 蜜蜂與大自然有共生關係

of bees, including the fruits and vegetables

- 蜜蜂與大自然有共生關係

that we eat every day.

許多植物都仰賴蜜蜂授粉

Bees and other insects

包括我們日常食用的水果和蔬菜

are dying around the world

包括我們日常食用的水果和蔬菜

and we do not know why.

- 全球的蜜蜂和其他昆蟲正在大量死亡

We don't want the lack of knowledge

- 全球的蜜蜂和其他昆蟲正在大量死亡

to be the reason that bees continue to disappear.

但我們卻對箇中原因一無所知

I thought, 'We have to do something about it'.

- 我們不想因為知識方面的匱乏

So we built a hive monitor

導致蜜蜂滅絕

that uses a camera to track the bees

- 我心想：「我們得做點什麼才行」

as they come and go.

因此，我們打造出一台蜂箱監控器

Bees always return to their queen.

這台監控器會運用攝影機追蹤記錄蜜蜂活動資料

If we observe 500 leaving the hive

這台監控器會運用攝影機追蹤記錄蜜蜂活動資料

and only 200 returning

蜜蜂一定會回到蜂后身邊

we can assume there is a problem

假如我們觀察到有 500 隻蜜蜂離開蜂巢

in the local ecosystem.

最後只有 200 隻回巢

We train the model using Tensorflow

就可以假設當地生態系統出了狀況

Google's open-source machine learning framework

就可以假設當地生態系統出了狀況

that examines the video from each hive.

我們訓練模型時採用了 TensorFlow 架構

But what do we do with all this data that we collect?

這是 Google 提供的 開放原始碼機器學習架構

We share it with experts

能夠檢查每個蜂巢的錄影片段

so they can make informed decisions

我們收集的這些資料有什麼用途呢？

on things like when to mow lawns

我們會將資料分享給專家

and where to plant trees and flowers

協助他們做出更明智的決策

so that bees can continue to power our ecosystem.

例如決定除草的時機

This tech is only the first step.

以及挑選種植樹木和花卉的地點

Us humans will have to take it from there.

讓蜜蜂能夠繼續協助維持生態系統的運作

To all those putting nature first

這項技術只是第一步而已

Search on.

想要翻轉現況還必須靠人類的努力