TEDxBeaconStreet
Katie Bouman: How to take a picture of a black hole
Katie Bouman: 如何拍攝黑洞的照片
Filmed:
Readability: 4.5
5,455,257 views
在銀河系中心，有個超大質量黑洞。壯觀的它，有個熱氣旋的轉盤，會吸引敢於近距離接近它的所有東西，連光線也不例外。其實我們看不見它，由於它的表面投下了陰影，而這個陰影的影像有助於解釋關於宇宙的一些重大的問題。科學家之前認為，要製作出這樣的影像需要地球般大小的望遠鏡。然而 Katie Bouman 和天文學家團隊想出一個聰明的解決方案。和我們一起來探索，如何看清這無盡的黑暗吧。
Katie Bouman - Imaging scientist
Katie Bouman is part of an international team of astronomers that's creating the world's largest telescope to take the very first picture of a black hole. Full bio
Katie Bouman is part of an international team of astronomers that's creating the world's largest telescope to take the very first picture of a black hole. Full bio
Double-click the English transcript below to play the video.
00:13
In the movie "Interstellar,"
0
1436
1860
在「星際效應」這部電影中，
00:15
we get an up-close look
at a supermassive black hole.
at a supermassive black hole.
1
3320
3327
我們更近距離地看到了
超質量黑洞。
超質量黑洞。
00:18
Set against a backdrop of bright gas,
2
6671
2143
在明亮氣體的背景下，
00:20
the black hole's massive
gravitational pull
gravitational pull
3
8838
2118
黑洞的巨大引力
00:22
bends light into a ring.
4
10980
1435
使光線形成戒指般的環狀。
00:24
However, this isn't a real photograph,
5
12439
2109
但是，這不是張真實的照片，
00:26
but a computer graphic rendering --
6
14572
1786
而是電腦圖像的呈現，
00:28
an artistic interpretation
of what a black hole might look like.
of what a black hole might look like.
7
16382
3390
是對黑洞可能的呈像的
藝術化的演繹。
藝術化的演繹。
00:32
A hundred years ago,
8
20401
1166
一百年前，
00:33
Albert Einstein first published
his theory of general relativity.
his theory of general relativity.
9
21591
3601
愛因斯坦首先發表了他的相對論。
00:37
In the years since then,
10
25216
1439
那之後的幾年，
00:38
scientists have provided
a lot of evidence in support of it.
a lot of evidence in support of it.
11
26679
2973
科學家提供了很多證據支持他的理論。
00:41
But one thing predicted
from this theory, black holes,
from this theory, black holes,
12
29676
3084
但是從他的理論中預測到的黑洞
00:44
still have not been directly observed.
13
32784
2350
仍然無法有直接證據證實。
00:47
Although we have some idea
as to what a black hole might look like,
as to what a black hole might look like,
14
35158
3206
雖然我們對於黑洞的呈像有一些想法，
00:50
we've never actually taken
a picture of one before.
a picture of one before.
15
38388
2779
但是我們從來沒有真正
拍攝過一張黑洞的相片。
拍攝過一張黑洞的相片。
00:53
However, you might be surprised to know
that that may soon change.
that that may soon change.
16
41191
4279
也許你會驚訝於這種困境即將改變。
00:57
We may be seeing our first picture
of a black hole in the next couple years.
of a black hole in the next couple years.
17
45494
4164
我們在未來幾年內也許
可以得到第一張黑洞的相片。
可以得到第一張黑洞的相片。
01:01
Getting this first picture will come down
to an international team of scientists,
to an international team of scientists,
18
49682
3958
國際的科學家團隊
將會獲得這第一張圖片，
將會獲得這第一張圖片，
01:05
an Earth-sized telescope
19
53664
1567
透過地球大小般的望遠鏡，
01:07
and an algorithm that puts together
the final picture.
the final picture.
20
55255
2832
和一個演算方法，獲得最後這張圖片。
01:10
Although I won't be able to show you
a real picture of a black hole today,
a real picture of a black hole today,
21
58111
3528
雖然，我今天無法讓大家看到
黑洞真正的照片，
黑洞真正的照片，
01:13
I'd like to give you a brief glimpse
into the effort involved
into the effort involved
22
61663
2911
但是，我想要簡單地向各位說明一下
01:16
in getting that first picture.
23
64598
1613
獲得這首張照片所付出的努力。
01:19
My name is Katie Bouman,
24
67477
1437
我是 Katie Bouman ，
01:20
and I'm a PhD student at MIT.
25
68938
2566
一名麻省理工學院博士生。
01:23
I do research in a computer science lab
26
71528
2027
我在電腦科學實驗室做研究，
01:25
that works on making computers
see through images and video.
see through images and video.
27
73579
3298
讓電腦透過影像及影片，
能夠「看見」、識別。
能夠「看見」、識別。
01:28
But although I'm not an astronomer,
28
76901
2162
雖然我不是天文學家，
01:31
today I'd like to show you
29
79087
1285
但是，我現在要讓大家看的是
01:32
how I've been able to contribute
to this exciting project.
to this exciting project.
30
80396
2903
我如何投入這令人興奮的專案。
01:35
If you go out past
the bright city lights tonight,
the bright city lights tonight,
31
83323
2831
如果今晚你們離開了城市明亮的燈光，
01:38
you may just be lucky enough
to see a stunning view
to see a stunning view
32
86178
2436
可能運氣夠好，
可以看到銀河系美麗的影像。
01:40
of the Milky Way Galaxy.
33
88638
1493
01:42
And if you could zoom past
millions of stars,
millions of stars,
34
90155
2462
如果你的視野能夠穿越數百萬顆星星，
01:44
26,000 light-years toward the heart
of the spiraling Milky Way,
of the spiraling Milky Way,
35
92641
3755
向着銀河的螺旋中心前進 26,000 光年，
01:48
we'd eventually reach
a cluster of stars right at the center.
a cluster of stars right at the center.
36
96420
3521
最後會在中心點遇到一群星星。
01:51
Peering past all the galactic dust
with infrared telescopes,
with infrared telescopes,
37
99965
3206
天文學家使用紅外線望遠鏡
透過銀河系塵埃
透過銀河系塵埃
01:55
astronomers have watched these stars
for over 16 years.
for over 16 years.
38
103195
3867
觀察這些星星，
已經超過了 16 年。
已經超過了 16 年。
01:59
But it's what they don't see
that is the most spectacular.
that is the most spectacular.
39
107086
3589
但是，最為壯觀的東西，
卻是他們無法看見的。
卻是他們無法看見的。
02:02
These stars seem to orbit
an invisible object.
an invisible object.
40
110699
3066
這些星星似乎繞著一個
隱形的物體運轉著。
隱形的物體運轉著。
02:05
By tracking the paths of these stars,
41
113789
2323
藉由追蹤這些星星的軌跡，
02:08
astronomers have concluded
42
116136
1294
天文學家得到一個結論：
02:09
that the only thing small and heavy
enough to cause this motion
enough to cause this motion
43
117454
3129
只有一個又小又重的物體
才能夠造成這樣的運動軌跡，
才能夠造成這樣的運動軌跡，
02:12
is a supermassive black hole --
44
120607
1968
那就是超質量黑洞，
02:14
an object so dense that it sucks up
anything that ventures too close --
anything that ventures too close --
45
122599
4178
它的密度高到能夠吸收
所有敢於近距離靠近它的東西，
所有敢於近距離靠近它的東西，
02:18
even light.
46
126801
1494
連光線也不例外。
02:20
But what happens if we were
to zoom in even further?
to zoom in even further?
47
128319
3061
但是，如果我們將影像放大，
會發生什麼事呢？
會發生什麼事呢？
02:23
Is it possible to see something
that, by definition, is impossible to see?
that, by definition, is impossible to see?
48
131404
4733
有沒有可能看到那些
原本被定義為看不見的東西呢？
原本被定義為看不見的東西呢？
02:28
Well, it turns out that if we were
to zoom in at radio wavelengths,
to zoom in at radio wavelengths,
49
136719
3244
事實顯示，如果我們
以無線電波長的尺度放大，
以無線電波長的尺度放大，
02:31
we'd expect to see a ring of light
50
139987
1682
我們預期可以看到一個光環，
02:33
caused by the gravitational
lensing of hot plasma
lensing of hot plasma
51
141693
2411
它是由黑洞旁
高速移動的熱離子體的
高速移動的熱離子體的
02:36
zipping around the black hole.
52
144128
1829
「引力透鏡」效應形成。
02:37
In other words,
53
145981
1160
換句話說，
02:39
the black hole casts a shadow
on this backdrop of bright material,
on this backdrop of bright material,
54
147165
3171
黑洞在明亮物質的背景下投射出陰影，
02:42
carving out a sphere of darkness.
55
150360
1842
刻畫出黑色的球體。
02:44
This bright ring reveals
the black hole's event horizon,
the black hole's event horizon,
56
152226
3339
這個光環揭露了黑洞的表面界限，
02:47
where the gravitational pull
becomes so great
becomes so great
57
155589
2400
在那個地方，引力拉扯的力量很大，
02:50
that not even light can escape.
58
158013
1626
連光線都無法逃脫。
02:51
Einstein's equations predict
the size and shape of this ring,
the size and shape of this ring,
59
159663
2859
愛因斯坦方程式預測了
這個光環的大小與形狀，
這個光環的大小與形狀，
02:54
so taking a picture of it
wouldn't only be really cool,
wouldn't only be really cool,
60
162546
3208
所以拍攝黑洞的相片不只是很酷，
02:57
it would also help to verify
that these equations hold
that these equations hold
61
165778
2618
它也有助於驗證這些方程式
03:00
in the extreme conditions
around the black hole.
around the black hole.
62
168420
2466
能在黑洞附近這樣的
極端環境下成立。
極端環境下成立。
03:02
However, this black hole
is so far away from us,
is so far away from us,
63
170910
2558
但是，這個黑洞距離我們非常遙遠，
03:05
that from Earth, this ring appears
incredibly small --
incredibly small --
64
173492
3098
從地球看過去，
這個光環是不可思議的小，
這個光環是不可思議的小，
03:08
the same size to us as an orange
on the surface of the moon.
on the surface of the moon.
65
176614
3590
就像是月球表面的
一個橘子那樣的小。
一個橘子那樣的小。
03:12
That makes taking a picture of it
extremely difficult.
extremely difficult.
66
180758
2824
所以拍攝黑洞的相片是極其困難的。
03:16
Why is that?
67
184645
1302
為什麼呢？
03:18
Well, it all comes down
to a simple equation.
to a simple equation.
68
186512
3188
因為，這所有的一切
可以歸結於一個簡單的方程式。
可以歸結於一個簡單的方程式。
03:21
Due to a phenomenon called diffraction,
69
189724
2416
由於「衍射」現象，
03:24
there are fundamental limits
70
192164
1355
我們所能觀察到的最小物體，
03:25
to the smallest objects
that we can possibly see.
that we can possibly see.
71
193543
2670
是有限小的，
我們無法洞察更小的結構。
我們無法洞察更小的結構。
03:28
This governing equation says
that in order to see smaller and smaller,
that in order to see smaller and smaller,
72
196789
3672
這個方程式說，
為了要看到越來越小的物體，
為了要看到越來越小的物體，
03:32
we need to make our telescope
bigger and bigger.
bigger and bigger.
73
200485
2587
我們必須製作越來越大的望遠鏡。
03:35
But even with the most powerful
optical telescopes here on Earth,
optical telescopes here on Earth,
74
203096
3069
但是，即使透過地球上
最強大的光學望遠鏡，
最強大的光學望遠鏡，
03:38
we can't even get close
to the resolution necessary
to the resolution necessary
75
206189
2419
我們還是遠遠達不到
03:40
to image on the surface of the moon.
76
208632
2198
拍攝月球表面的影像所需要的解析度。
03:42
In fact, here I show one of the highest
resolution images ever taken
resolution images ever taken
77
210854
3617
事實上，請大家看看這張從地球拍攝的
解析度最高的月球照片之一，
03:46
of the moon from Earth.
78
214495
1397
03:47
It contains roughly 13,000 pixels,
79
215916
2557
這張相片大約有一萬三千像素，
03:50
and yet each pixel would contain
over 1.5 million oranges.
over 1.5 million oranges.
80
218497
4050
而每一個像素可包含
超過 150 萬個橘子。
超過 150 萬個橘子。
03:55
So how big of a telescope do we need
81
223396
1972
那麼，為了要看到月球表面的橘子，
03:57
in order to see an orange
on the surface of the moon
on the surface of the moon
82
225392
2765
我們需要多大的望遠鏡呢？
04:00
and, by extension, our black hole?
83
228181
2214
再者，為了要看到黑洞？
04:02
Well, it turns out
that by crunching the numbers,
that by crunching the numbers,
84
230419
2340
事實證明，透過大量運算，
04:04
you can easily calculate
that we would need a telescope
that we would need a telescope
85
232783
2610
我們可以很容易地計算出
我們所需要的望遠鏡
我們所需要的望遠鏡
04:07
the size of the entire Earth.
86
235417
1393
必須是整個地球那麼大。
04:08
(Laughter)
87
236834
1024
(笑聲)
04:09
If we could build
this Earth-sized telescope,
this Earth-sized telescope,
88
237882
2119
如果我們建造出地球般大小的望遠鏡，
04:12
we could just start to make out
that distinctive ring of light
that distinctive ring of light
89
240025
2925
我們馬上就可以探測出一個獨特光環，
04:14
indicative of the black
hole's event horizon.
hole's event horizon.
90
242974
2183
它表明了黑洞的表面界限。
04:17
Although this picture wouldn't contain
all the detail we see
all the detail we see
91
245181
2918
雖然這張相片沒有包含所有細節，
04:20
in computer graphic renderings,
92
248123
1506
像我們在電腦圖形渲染上看到的那樣，
04:21
it would allow us to safely get
our first glimpse
our first glimpse
93
249653
2299
但是，至少我們可以安全地
04:23
of the immediate environment
around a black hole.
around a black hole.
94
251976
2487
對黑洞附近的環境瞥上一眼。
04:26
However, as you can imagine,
95
254487
1613
然而，如同大家想像的，
04:28
building a single-dish telescope
the size of the Earth is impossible.
the size of the Earth is impossible.
96
256124
3624
建造一個地球大小的
單碟望遠鏡是不可能的。
單碟望遠鏡是不可能的。
04:31
But in the famous words of Mick Jagger,
97
259772
1887
但是在 Mick Jagger 的名言中：
04:33
"You can't always get what you want,
98
261683
1791
「你無法一直得到你所想要的，
04:35
but if you try sometimes,
you just might find
you just might find
99
263498
2187
但是如果你去嘗試，你可能會發現
04:37
you get what you need."
100
265709
1215
你得到了你所需要的。」
04:38
And by connecting telescopes
from around the world,
from around the world,
101
266948
2464
藉由連結世界各地的望遠鏡，
04:41
an international collaboration
called the Event Horizon Telescope
called the Event Horizon Telescope
102
269436
3538
名為「事件視界望遠鏡」的國際組織
04:44
is creating a computational telescope
the size of the Earth,
the size of the Earth,
103
272998
3109
正著手創建一個地球大小的
計算型望遠鏡，
計算型望遠鏡，
04:48
capable of resolving structure
104
276131
1537
它能夠解析黑洞的
04:49
on the scale of a black
hole's event horizon.
hole's event horizon.
105
277692
2199
表面界限的結構。
04:51
This network of telescopes is scheduled
to take its very first picture
to take its very first picture
106
279915
3387
這個望遠鏡網路預計明年
04:55
of a black hole next year.
107
283326
1815
拍攝黑洞的第一張相片。
04:57
Each telescope in the worldwide
network works together.
network works together.
108
285165
3338
世界各地的望遠鏡網路同時運作。
05:00
Linked through the precise timing
of atomic clocks,
of atomic clocks,
109
288527
2712
透過原子鐘的精準時間鏈結，
05:03
teams of researchers at each
of the sights freeze light
of the sights freeze light
110
291263
2657
每個地點的研究團隊
藉由蒐集數千兆兆字節的數據
藉由蒐集數千兆兆字節的數據
05:05
by collecting thousands
of terabytes of data.
of terabytes of data.
111
293944
2962
將光線「定格」。
05:08
This data is then processed in a lab
right here in Massachusetts.
right here in Massachusetts.
112
296930
5017
麻薩諸塞州這裡的實驗室
接下來處理這些資料。
接下來處理這些資料。
05:13
So how does this even work?
113
301971
1794
那麼，這些資料是如何運作的呢？
05:15
Remember if we want to see the black hole
in the center of our galaxy,
in the center of our galaxy,
114
303789
3403
還記得嗎？如果我們想要
看清在銀河中間的黑洞，
看清在銀河中間的黑洞，
05:19
we need to build this impossibly large
Earth-sized telescope?
Earth-sized telescope?
115
307216
2982
我們就需要建造地球大小的
望遠鏡，這是不現實的。
望遠鏡，這是不現實的。
05:22
For just a second,
let's pretend we could build
let's pretend we could build
116
310222
2232
等一下，假設我們能夠建造
05:24
a telescope the size of the Earth.
117
312478
1842
地球般大小的望遠鏡。
05:26
This would be a little bit
like turning the Earth
like turning the Earth
118
314344
2455
就有點像將地球
05:28
into a giant spinning disco ball.
119
316823
1747
想像成舞廳裡的迪斯可旋轉球。
05:30
Each individual mirror would collect light
120
318594
2200
每一面鏡子會蒐集光線，
05:32
that we could then combine
together to make a picture.
together to make a picture.
121
320818
2597
然後我們能將這些
影像整合成一張圖片。
影像整合成一張圖片。
05:35
However, now let's say
we remove most of those mirrors
we remove most of those mirrors
122
323439
2661
但是，現在讓我們
移除大多數的鏡子，
移除大多數的鏡子，
05:38
so only a few remained.
123
326124
1972
只剩下少數幾個。
05:40
We could still try to combine
this information together,
this information together,
124
328120
2877
我們仍可試著整合這些資訊，
05:43
but now there are a lot of holes.
125
331021
1993
但是，現在只能看到很多「孔洞」。
05:45
These remaining mirrors represent
the locations where we have telescopes.
the locations where we have telescopes.
126
333038
4373
這些剩下的鏡子代表
那些有望遠鏡的地方。
那些有望遠鏡的地方。
05:49
This is an incredibly small number
of measurements to make a picture from.
of measurements to make a picture from.
127
337435
4079
測量數據少之又少，
甚至無法形成一張圖片。
甚至無法形成一張圖片。
05:53
But although we only collect light
at a few telescope locations,
at a few telescope locations,
128
341538
3838
雖然我們只在少數
有望遠鏡的地方蒐集光線，
有望遠鏡的地方蒐集光線，
05:57
as the Earth rotates, we get to see
other new measurements.
other new measurements.
129
345400
3423
地球旋轉時，我們可以
獲得一些新的測量數據。
獲得一些新的測量數據。
06:00
In other words, as the disco ball spins,
those mirrors change locations
those mirrors change locations
130
348847
3819
換句話說，就像迪斯可球旋轉時，
那些鏡子也會改變位置，
那些鏡子也會改變位置，
06:04
and we get to observe
different parts of the image.
different parts of the image.
131
352690
2899
我們得以觀察不同面向的影像。
06:07
The imaging algorithms we develop
fill in the missing gaps of the disco ball
fill in the missing gaps of the disco ball
132
355613
4018
我們所開發的成像算法填補了
「迪斯可球」的不可見縫隙，
「迪斯可球」的不可見縫隙，
06:11
in order to reconstruct
the underlying black hole image.
the underlying black hole image.
133
359655
3033
目的在重建黑洞的相片。
06:14
If we had telescopes located
everywhere on the globe --
everywhere on the globe --
134
362712
2636
如果地球的每個地方都有望遠鏡，
06:17
in other words, the entire disco ball --
135
365372
1941
也就是整個迪斯可球佈滿了鏡子，
06:19
this would be trivial.
136
367337
1284
這是最簡潔、理想的情況。
06:20
However, we only see a few samples,
and for that reason,
and for that reason,
137
368645
3322
但是，我們只看得到
某些局部的成像，因此，
某些局部的成像，因此，
06:23
there are an infinite number
of possible images
of possible images
138
371991
2388
有無數可能的相片
06:26
that are perfectly consistent
with our telescope measurements.
with our telescope measurements.
139
374403
2964
可以與現有望遠鏡的
局部成像相吻合。
局部成像相吻合。
06:29
However, not all images are created equal.
140
377391
3016
當然，並不是每一張「相片」的
優先級別都相同。
優先級別都相同。
06:32
Some of those images look more like
what we think of as images than others.
what we think of as images than others.
141
380849
4458
有些相片比別的
更近似我們所想像的。
更近似我們所想像的。
06:37
And so, my role in helping to take
the first image of a black hole
the first image of a black hole
142
385331
3222
因此，為了協助拍攝黑洞
第一張相片，我的任務就是
第一張相片，我的任務就是
06:40
is to design algorithms that find
the most reasonable image
the most reasonable image
143
388577
2932
設計發現最合理影像的演算法，
06:43
that also fits the telescope measurements.
144
391533
2222
當然也必須符合望遠鏡的量測數據。
06:46
Just as a forensic sketch artist
uses limited descriptions
uses limited descriptions
145
394727
3942
就像法庭的素描家一樣，
利用有限的相貌描述以及
利用有限的相貌描述以及
06:50
to piece together a picture using
their knowledge of face structure,
their knowledge of face structure,
146
398693
3514
他們對於臉部結構的知識，
將表現相貌特點的圖片拼湊出來，
將表現相貌特點的圖片拼湊出來，
06:54
the imaging algorithms I develop
use our limited telescope data
use our limited telescope data
147
402231
3315
我開發的影像演算法
使用有限的望遠鏡資料
使用有限的望遠鏡資料
06:57
to guide us to a picture that also
looks like stuff in our universe.
looks like stuff in our universe.
148
405570
4322
為我們生成這種影像：
類似於宇宙中的事物的影像。
類似於宇宙中的事物的影像。
07:01
Using these algorithms,
we're able to piece together pictures
we're able to piece together pictures
149
409916
3651
利用這些演算法，
讓我們能夠利用零零散散的資料
讓我們能夠利用零零散散的資料
07:05
from this sparse, noisy data.
150
413591
2180
拼湊出黑洞可能的樣子。
07:07
So here I show a sample reconstruction
done using simulated data,
done using simulated data,
151
415795
4529
在這裡，讓大家看一個利用模擬資料
重建的影像樣本，
重建的影像樣本，
07:12
when we pretend to point our telescopes
152
420348
1933
這是我們假設將望遠鏡指向
07:14
to the black hole
in the center of our galaxy.
in the center of our galaxy.
153
422305
2585
銀河系中心的黑洞時所得到的。
07:16
Although this is just a simulation,
reconstruction such as this give us hope
reconstruction such as this give us hope
154
424914
4455
雖然這只是一個模擬，
但是這讓我們充滿了希望：
但是這讓我們充滿了希望：
07:21
that we'll soon be able to reliably take
the first image of a black hole
the first image of a black hole
155
429393
3453
我們能夠仰賴這樣的模擬演算法，
很快地得到黑洞的第一張相片，
很快地得到黑洞的第一張相片，
07:24
and from it, determine
the size of its ring.
the size of its ring.
156
432870
2595
同時也能計算「光環」的大小。
07:28
Although I'd love to go on
about all the details of this algorithm,
about all the details of this algorithm,
157
436118
3199
雖然我很樂意繼續說明
這個演算法的所有細節，
這個演算法的所有細節，
07:31
luckily for you, I don't have the time.
158
439341
2174
但由於時間不夠，所以
你們也不用費腦子聽了。
你們也不用費腦子聽了。
07:33
But I'd still like
to give you a brief idea
to give you a brief idea
159
441539
2001
但是，我還是很樂意
跟大家做個簡短的說明：
跟大家做個簡短的說明：
07:35
of how we define
what our universe looks like,
what our universe looks like,
160
443564
2302
我們如何定義宇宙看起來像什麼？
07:37
and how we use this to reconstruct
and verify our results.
and verify our results.
161
445890
4466
以及我們如何
利用這個演算法重建並驗證結果。
利用這個演算法重建並驗證結果。
07:42
Since there are an infinite number
of possible images
of possible images
162
450380
2496
因為有無數可能的影像
07:44
that perfectly explain
our telescope measurements,
our telescope measurements,
163
452900
2365
與地球上望遠鏡的量測完全符合，
07:47
we have to choose
between them in some way.
between them in some way.
164
455289
2605
我們必須在它們之間
找個方法進行挑選。
找個方法進行挑選。
07:49
We do this by ranking the images
165
457918
1838
我們對影像進行打分，
07:51
based upon how likely they are
to be the black hole image,
to be the black hole image,
166
459780
2834
打分的根據是：看起來有多像黑洞，
07:54
and then choosing the one
that's most likely.
that's most likely.
167
462638
2482
然後選擇最像的影像。
07:57
So what do I mean by this exactly?
168
465144
2195
那麼，這到底是什麼意思呢？
07:59
Let's say we were trying to make a model
169
467862
1978
假設我們試著建立一個模型，
08:01
that told us how likely an image
were to appear on Facebook.
were to appear on Facebook.
170
469864
3183
它告訴我們這個影像在
Facebook 上出現的可能性。
Facebook 上出現的可能性。
08:05
We'd probably want the model to say
171
473071
1701
我們希望這個模型會這樣判斷：
08:06
it's pretty unlikely that someone
would post this noise image on the left,
would post this noise image on the left,
172
474796
3557
大家應該不太可能會上傳
像左邊這張亂亂的圖，
像左邊這張亂亂的圖，
08:10
and pretty likely that someone
would post a selfie
would post a selfie
173
478377
2419
而比較可能會上傳自拍照，
08:12
like this one on the right.
174
480820
1334
像右邊這張。
08:14
The image in the middle is blurry,
175
482178
1639
中間這張圖像片是模糊的，
08:15
so even though it's more likely
we'd see it on Facebook
we'd see it on Facebook
176
483841
2639
即使模糊，和亂亂的圖像比較的話，
我們還是很有可能
我們還是很有可能
08:18
compared to the noise image,
177
486504
1360
在 Facebook 上看到，
08:19
it's probably less likely we'd see it
compared to the selfie.
compared to the selfie.
178
487888
2960
只不過不如自拍照那樣常見。
08:22
But when it comes to images
from the black hole,
from the black hole,
179
490872
2290
但是，如果是黑洞的影像，
08:25
we're posed with a real conundrum:
we've never seen a black hole before.
we've never seen a black hole before.
180
493186
3502
我們遇到一個真正的難題：
我們從來沒見過黑洞的樣子。
我們從來沒見過黑洞的樣子。
08:28
In that case, what is a likely
black hole image,
black hole image,
181
496712
2291
在這種情況下，
黑洞可能的影像是什麼？
黑洞可能的影像是什麼？
08:31
and what should we assume
about the structure of black holes?
about the structure of black holes?
182
499027
2938
我們應該假設黑洞的結構是什麼？
08:33
We could try to use images
from simulations we've done,
from simulations we've done,
183
501989
2632
我們可能會試著使用
之前生成的模擬結果，
之前生成的模擬結果，
08:36
like the image of the black hole
from "Interstellar,"
from "Interstellar,"
184
504645
2530
像「星際效應」裡的黑洞影像，
08:39
but if we did this,
it could cause some serious problems.
it could cause some serious problems.
185
507199
2938
但是，如果這樣做的話，
可能會造成一些嚴重的問題。
可能會造成一些嚴重的問題。
08:42
What would happen
if Einstein's theories didn't hold?
if Einstein's theories didn't hold?
186
510161
3380
如果愛因斯坦的理論不適用的話，
會發生什麼事？
會發生什麼事？
08:45
We'd still want to reconstruct
an accurate picture of what was going on.
an accurate picture of what was going on.
187
513565
3961
我們還是想要重建
一個準確的圖像。
一個準確的圖像。
08:49
If we bake Einstein's equations
too much into our algorithms,
too much into our algorithms,
188
517550
3371
如果將太多愛因斯坦的方程式
融入我們的演算法中，
融入我們的演算法中，
08:52
we'll just end up seeing
what we expect to see.
what we expect to see.
189
520945
2755
最後只會得到我們期望的結果，
而不一定是事實。
而不一定是事實。
08:55
In other words,
we want to leave the option open
we want to leave the option open
190
523724
2276
換句話說，我們不能
貿然確定實際情況如何，
貿然確定實際情況如何，
08:58
for there being a giant elephant
at the center of our galaxy.
at the center of our galaxy.
191
526024
2923
因為銀河系中央有一隻巨象。
09:00
(Laughter)
192
528971
1057
(笑聲)
09:02
Different types of images have
very distinct features.
very distinct features.
193
530052
2989
不同類型的影像有著
各自非常顯著的特徵。
各自非常顯著的特徵。
09:05
We can easily tell the difference
between black hole simulation images
between black hole simulation images
194
533065
3548
我們可以很容易地區分
黑洞模擬影像
黑洞模擬影像
09:08
and images we take
every day here on Earth.
every day here on Earth.
195
536637
2276
以及我們在地球上日常生活中的照片。
09:10
We need a way to tell our algorithms
what images look like
what images look like
196
538937
3104
我們需要一種方法來告訴演算法
影像看起來像什麼，
影像看起來像什麼，
09:14
without imposing one type
of image's features too much.
of image's features too much.
197
542065
3249
而不是去強加特定一種影像的特徵給它。
09:17
One way we can try to get around this
198
545865
1893
我們可以用一個方法
試著解決這個問題：
試著解決這個問題：
09:19
is by imposing the features
of different kinds of images
of different kinds of images
199
547782
3062
通過導入不同類型影像的特徵
讓演算法重建影像，
讓演算法重建影像，
09:22
and seeing how the type of image we assume
affects our reconstructions.
affects our reconstructions.
200
550868
4130
然後觀察預先假設的影像類型
如何影響我們重建的影像。
如何影響我們重建的影像。
09:27
If all images' types produce
a very similar-looking image,
a very similar-looking image,
201
555712
3491
如果所有不同類型的影像特徵
產生的結果都很類似，
產生的結果都很類似，
09:31
then we can start to become more confident
202
559227
2057
那麼我們可以充滿信心地說：
09:33
that the image assumptions we're making
are not biasing this picture that much.
are not biasing this picture that much.
203
561308
4173
對於這個影像所做的假設
沒有與事實偏差太多。
沒有與事實偏差太多。
09:37
This is a little bit like
giving the same description
giving the same description
204
565505
2990
這有點像是將相同的相貌描述
09:40
to three different sketch artists
from all around the world.
from all around the world.
205
568519
2996
提供給三個來自世界各地不同的素描家，
09:43
If they all produce
a very similar-looking face,
a very similar-looking face,
206
571539
2860
如果他們都畫出很相像的臉，
09:46
then we can start to become confident
207
574423
1793
那麼我們可以充滿信心地說：
09:48
that they're not imposing their own
cultural biases on the drawings.
cultural biases on the drawings.
208
576240
3616
他們的作品沒有受到
本人的文化偏見的影響。
本人的文化偏見的影響。
09:51
One way we can try to impose
different image features
different image features
209
579880
3315
我們導入不同類型影像
的特徵的一個方法
的特徵的一個方法
09:55
is by using pieces of existing images.
210
583219
2441
就是藉由現存的影像去拼湊。
09:58
So we take a large collection of images,
211
586214
2160
所以我們要蒐集大量的影像，
10:00
and we break them down
into their little image patches.
into their little image patches.
212
588398
2718
然後將它們分解成許多碎片。
10:03
We then can treat each image patch
a little bit like pieces of a puzzle.
a little bit like pieces of a puzzle.
213
591140
4285
之後我們可以把這些碎片
當作拼圖的碎片。
當作拼圖的碎片。
10:07
And we use commonly seen puzzle pieces
to piece together an image
to piece together an image
214
595449
4278
我們使用常見的「碎片」拼湊成圖片，
10:11
that also fits our telescope measurements.
215
599751
2452
這張圖片當然也要
符合望遠鏡的量測數據。
符合望遠鏡的量測數據。
10:15
Different types of images have
very distinctive sets of puzzle pieces.
very distinctive sets of puzzle pieces.
216
603040
3743
不同類型的影像有其獨特的拼圖碎片。
10:18
So what happens when we take the same data
217
606807
2806
所以，當我們利用相同的數據資料
10:21
but we use different sets of puzzle pieces
to reconstruct the image?
to reconstruct the image?
218
609637
4130
卻使用不同類型的拼圖碎片
來重建這個影像，會發生什麼事？
來重建這個影像，會發生什麼事？
10:25
Let's first start with black hole
image simulation puzzle pieces.
image simulation puzzle pieces.
219
613791
4766
讓我們先從黑洞模擬
圖像的拼圖碎片開始。
圖像的拼圖碎片開始。
10:30
OK, this looks reasonable.
220
618581
1591
好的，這看起來很合理。
10:32
This looks like what we expect
a black hole to look like.
a black hole to look like.
221
620196
2694
這看起來像我們
所期待的黑洞的樣子。
所期待的黑洞的樣子。
10:34
But did we just get it
222
622914
1193
但是，僅僅是導入了
一些些黑洞模擬影像的碎片，
一些些黑洞模擬影像的碎片，
10:36
because we just fed it little pieces
of black hole simulation images?
of black hole simulation images?
223
624131
3314
我們就得出了結果嗎？
10:39
Let's try another set of puzzle pieces
224
627469
1880
讓我們來試試另一組拼圖，
10:41
from astronomical, non-black hole objects.
225
629373
2509
這些是天文學影像的拼圖，不是黑洞的。
10:44
OK, we get a similar-looking image.
226
632914
2126
沒錯，我們得到一個類似的影像。
10:47
And then how about pieces
from everyday images,
from everyday images,
227
635064
2236
那麼如果是日常生活的影像呢？
10:49
like the images you take
with your own personal camera?
with your own personal camera?
228
637324
2785
就像用相機所照的照片一樣？
10:53
Great, we see the same image.
229
641312
2115
很好，我們得到相同的影像。
10:55
When we get the same image
from all different sets of puzzle pieces,
from all different sets of puzzle pieces,
230
643451
3366
當我們從不同類型的拼圖
得到相同的影像，
得到相同的影像，
10:58
then we can start to become more confident
231
646841
2046
我們更有信心了，
11:00
that the image assumptions we're making
232
648911
1966
我們所假定的影像
11:02
aren't biasing the final
image we get too much.
image we get too much.
233
650901
2921
和我們最後得到的影像
並沒有差距太多。
並沒有差距太多。
11:05
Another thing we can do is take
the same set of puzzle pieces,
the same set of puzzle pieces,
234
653846
3253
我們可以做的另一件事
就是使用同一組拼圖，
就是使用同一組拼圖，
11:09
such as the ones derived
from everyday images,
from everyday images,
235
657123
2489
比如日常生活中的影像碎片，
11:11
and use them to reconstruct
many different kinds of source images.
many different kinds of source images.
236
659636
3600
並利用它們來重組
各種不同素材來源的影像。
各種不同素材來源的影像。
11:15
So in our simulations,
237
663260
1271
那麼，在模擬實驗當中，
11:16
we pretend a black hole looks like
astronomical non-black hole objects,
astronomical non-black hole objects,
238
664555
3775
我們假設黑洞看起來就像是
天文學裡那些非黑洞的物體，
天文學裡那些非黑洞的物體，
11:20
as well as everyday images like
the elephant in the center of our galaxy.
the elephant in the center of our galaxy.
239
668354
3849
或者又把它看成「銀河系中央的大象」
這樣的日常生活影像。
這樣的日常生活影像。
11:24
When the results of our algorithms
on the bottom look very similar
on the bottom look very similar
240
672227
3168
我們下方的演算結果
11:27
to the simulation's truth image on top,
241
675419
2096
和上方的模擬實驗中的真實影像很相像，
11:29
then we can start to become
more confident in our algorithms.
more confident in our algorithms.
242
677539
3346
我們就可以對我們的演算法更有信心。
11:32
And I really want to emphasize here
243
680909
1867
我真的想要強調這一點：
11:34
that all of these pictures were created
244
682800
1934
這些所有的圖片都是
11:36
by piecing together little pieces
of everyday photographs,
of everyday photographs,
245
684758
2936
由日常生活照片的碎片
拼湊出來的，
拼湊出來的，
11:39
like you'd take with your own
personal camera.
personal camera.
246
687718
2215
就是那種用私人相機照出來的照片。
11:41
So an image of a black hole
we've never seen before
we've never seen before
247
689957
3276
我們之前從沒看過黑洞的相片，
11:45
may eventually be created by piecing
together pictures we see all the time
together pictures we see all the time
248
693257
3943
但最後黑洞的相片也許是由我們
常常看到的日常生活照片拼湊出來的：
常常看到的日常生活照片拼湊出來的：
11:49
of people, buildings,
trees, cats and dogs.
trees, cats and dogs.
249
697224
2745
人像、建築物、樹木、貓、狗等等。
11:51
Imaging ideas like this
will make it possible for us
will make it possible for us
250
699993
2645
這些成像方法讓我們能夠
11:54
to take our very first pictures
of a black hole,
of a black hole,
251
702662
2619
拍攝出黑洞的第一張相片，
11:57
and hopefully, verify
those famous theories
those famous theories
252
705305
2447
我們同時也希望
能夠驗證那些著名的理論，
能夠驗證那些著名的理論，
11:59
on which scientists rely on a daily basis.
253
707776
2421
那些科學家平常所依賴的理論。
12:02
But of course, getting
imaging ideas like this working
imaging ideas like this working
254
710221
2608
當然，提出這些成像的方法與理論，
12:04
would never have been possible
without the amazing team of researchers
without the amazing team of researchers
255
712853
3322
沒有一個驚人的研究團隊
是不可能達到這種成果的，
是不可能達到這種成果的，
12:08
that I have the privilege to work with.
256
716199
1887
我很榮幸身為這個團隊的一員。
12:10
It still amazes me
257
718110
1163
我對這件事感到驚異：
12:11
that although I began this project
with no background in astrophysics,
with no background in astrophysics,
258
719297
3351
雖然我沒有任何天文物理的背景
而加入這個專案，
而加入這個專案，
12:14
what we have achieved
through this unique collaboration
through this unique collaboration
259
722672
2619
我們透過這獨特的合作所得到的，
12:17
could result in the very first
images of a black hole.
images of a black hole.
260
725315
2759
能夠獲得第一張黑洞的相片。
12:20
But big projects like
the Event Horizon Telescope
the Event Horizon Telescope
261
728098
2698
但是像「事件視界望遠鏡」
這樣的大專案，
這樣的大專案，
12:22
are successful due to all
the interdisciplinary expertise
the interdisciplinary expertise
262
730820
2814
多虧有跨學科領域的專業知識而成功，
12:25
different people bring to the table.
263
733658
1790
不同的專家共同合作着。
12:27
We're a melting pot of astronomers,
264
735472
1706
我們像是個熔爐，集結了天文學家、
12:29
physicists, mathematicians and engineers.
265
737202
2232
物理學家、數學家和工程師。
12:31
This is what will make it soon possible
266
739458
2554
這就是我們讓不可思議的事情
12:34
to achieve something
once thought impossible.
once thought impossible.
267
742036
2853
快速實現的原因。
12:36
I'd like to encourage all of you to go out
268
744913
2256
我很想鼓勵大家
12:39
and help push the boundaries of science,
269
747193
2096
去協助推動科學的前沿，
12:41
even if it may at first seem
as mysterious to you as a black hole.
as mysterious to you as a black hole.
270
749313
3901
即使第一步可能像黑洞那樣神秘。
12:45
Thank you.
271
753238
1174
謝謝大家。
12:46
(Applause)
272
754436
2397
(掌聲)
ABOUT THE SPEAKER
Katie Bouman - Imaging scientistKatie Bouman is part of an international team of astronomers that's creating the world's largest telescope to take the very first picture of a black hole.
Why you should listen
It is believed that the heart of the Milky Way hosts a four-million-solar-mass black hole feeding off a spinning disk of hot gas. An image of the shadow cast by the event horizon of the black hole could help to address a number of important scientific questions. For instance, does Einstein's theory of general relativity hold in extreme conditions? Unfortunately, the event horizon of this black hole appears so small in the sky that imaging it would require a single-dish radio telescope the size of the entire Earth. Although a single-dish telescope this large is unrealizable, by connecting disjoint radio telescopes located all around the globe, Katie Bouman and a team of astronomers are creating an Earth-sized computational telescope -- the Event Horizon Telescope -- that is capable of taking the very first up-close picture of a black hole.
Bouman is a PhD candidate in the Computer Science and Artificial Intelligence Laboratory (CSAIL) at the Massachusetts Institute of Technology (MIT). The focus of her research is on using emerging computational methods to push the boundaries of interdisciplinary imaging. By combining techniques from both astronomy and computer science, Bouman has been working on developing innovative ways to combine the information from the Event Horizon Telescope network to produce the first picture of a black hole. Her work on imaging for the Event Horizon Telescope has been featured on BBC, The Boston Globe, The Washington Post, Popular Science and NPR.
Katie Bouman | Speaker | TED.com