ABOUT THE SPEAKER
Deborah Gordon - Ecologist
By studying how ant colonies work without any one leader, Deborah Gordon has identified striking similarities in how ant colonies, brains, cells and computer networks regulate themselves.

Why you should listen

Ecologist Deborah M. Gordon has learned that ant colonies can work without central control by using simple interactions like how often the insects touch antennae. Contrary to the notion that colonies are organized by efficient ants, she has instead discovered that evolution has produced “noisy” systems that tolerate accident and respond flexibly to the environment. When conditions are tough, natural selection favors colonies that conserve resources.

Her studies of ant colonies have led her and her Stanford colleagues to the discovery of the “Anternet,” which regulates foraging in ants in the same way the internet regulates data traffic. But as she said to Wired in 2013, "Insect behavior mimicking human networks ... is actually not what’s most interesting about ant networks. What’s far more interesting are the parallels in the other direction: What have the ants worked out that we humans haven’t thought of yet?" Her latest exploration: How do ants behave in space?

More profile about the speaker
Deborah Gordon | Speaker | TED.com
TED2014

Deborah Gordon: What ants teach us about the brain, cancer and the Internet

黛博拉·高登: 關於大腦、癌症及網路,螞蟻教了我們什麼

Filmed:
1,481,089 views

生物學家黛博拉·高登研究螞蟻,無論她在何處找到它們——在沙漠、在熱帶地區、在她的廚房裡… 她在這個精彩的演說中,解釋她對這種我們連想都不想,就很開心地拍走的昆蟲,是如何為之著迷。她主張螞蟻的生活能提供有用的模式,讓我們學習很多其他的主題,包括疾病、科技及人類的大腦。
- Ecologist
By studying how ant colonies work without any one leader, Deborah Gordon has identified striking similarities in how ant colonies, brains, cells and computer networks regulate themselves. Full bio

Double-click the English transcript below to play the video.

00:13
I study研究 ants螞蟻
0
1099
1619
我研究螞蟻,
00:14
in the desert沙漠, in the tropical熱帶 forest森林
1
2718
3297
在沙漠中,在熱帶雨林中,
00:18
and in my kitchen廚房,
2
6015
1528
還有我家的廚房,
00:19
and in the hills丘陵 around Silicon Valley where I live生活.
3
7543
4020
以及我住的矽谷四周的山丘上。
00:23
I've recently最近 realized實現 that ants螞蟻
4
11563
1676
我最近才領悟到原來螞蟻
00:25
are using運用 interactions互動 differently不同
5
13239
2278
使用不同的互動法
00:27
in different不同 environments環境,
6
15517
1828
在不同的環境中,
00:29
and that got me thinking思維
that we could learn學習 from this
7
17345
2000
而這使我不禁去想,
我們或許能從這件事
00:31
about other systems系統,
8
19345
1555
了解其他系統,
00:32
like brains大腦 and data數據 networks網絡 that we engineer工程師,
9
20900
5588
像大腦及我們設計出的數據網路,
00:38
and even cancer癌症.
10
26488
2846
甚至癌症。
00:41
So what all these systems系統 have in common共同
11
29334
1752
所以這些系統的共通點
00:43
is that there's no central中央 control控制.
12
31086
2857
在於沒有中央控制。
00:45
An ant螞蟻 colony殖民地 consists of sterile無菌 female workers工人 --
13
33943
3820
蟻群是由不孕的雌性工蟻——
00:49
those are the ants螞蟻 you see walking步行 around —
14
37763
2386
就是你看到四處走動的那些螞蟻——
00:52
and then one or more reproductive生殖 females女性
15
40149
2156
還有一隻或多隻能生育的雌蟻組成,
00:54
who just lay鋪設 the eggs.
16
42305
2103
而這種雌蟻只下蛋。
00:56
They don't give any instructions說明.
17
44408
2009
牠們不下任何指令。
00:58
Even though雖然 they're called queens皇后,
18
46417
1984
即使牠們被稱為蟻后,
01:00
they don't tell anybody任何人 what to do.
19
48401
2165
牠們也不會告訴任何螞蟻要做什麼。
01:02
So in an ant螞蟻 colony殖民地, there's no one in charge收費,
20
50566
3276
所以在蟻群內,沒有負責人。
01:05
and all systems系統 like this without central中央 control控制
21
53842
2778
所有像這樣沒有中央控制的系統,
01:08
are regulated調控 using運用 very simple簡單 interactions互動.
22
56620
3919
要以非常簡單的互動來規範。
01:12
Ants螞蟻 interact相互作用 using運用 smell.
23
60549
2727
螞蟻以嗅覺互動。
01:15
They smell with their antennae天線,
24
63276
2187
牠們以觸角聞味道,
01:17
and they interact相互作用 with their antennae天線,
25
65463
2761
而且牠們也以觸角互動,
01:20
so when one ant螞蟻 touches觸摸 another另一個 with its antennae天線,
26
68224
3009
所以當一隻螞蟻以觸角
碰觸另外一隻螞蟻,
01:23
it can tell, for example, if the other ant螞蟻
27
71233
1607
牠就能分辨,譬如說這一隻
01:24
is a nestmatenestmate
28
72840
1634
是不是同窩蟻,
01:26
and what task任務 that other ant螞蟻 has been doing.
29
74474
4236
還有這隻螞蟻正在做什麼任務。
01:30
So here you see a lot of ants螞蟻 moving移動 around
30
78710
3200
所以這裡你看到很多螞蟻
01:33
and interacting互動 in a lab實驗室 arena競技場
31
81910
1933
在這個實驗場所四處走動及互動,
01:35
that's connected連接的 by tubes to two other arenas競技場.
32
83843
3438
這個場所與另外兩個以管子相連。
01:39
So when one ant螞蟻 meets符合 another另一個,
33
87281
2504
所以當某隻螞蟻碰到另一隻,
01:41
it doesn't matter which哪一個 ant螞蟻 it meets符合,
34
89785
2089
碰到的是哪一隻螞蟻並不重要,
01:43
and they're actually其實 not transmitting發射
35
91874
1885
而且牠們其實並沒有傳送
01:45
any kind of complicated複雜 signal信號 or message信息.
36
93759
3465
任何複雜的信號或訊息。
01:49
All that matters事項 to the ant螞蟻 is the rate
37
97224
2111
對螞蟻而言最重要的是
01:51
at which哪一個 it meets符合 other ants螞蟻.
38
99335
2821
牠們碰到其他螞蟻的頻率。
01:54
And all of these interactions互動, taken採取 together一起,
39
102156
2757
而所有的互動,全部一起看,
01:56
produce生產 a network網絡.
40
104913
2505
會產生一種網路。
01:59
So this is the network網絡 of the ants螞蟻
41
107418
2210
那麼這就是螞蟻的網路,
02:01
that you just saw moving移動 around in the arena競技場,
42
109628
2644
由你剛剛看到的那個場所
裡面的移動所形成。
02:04
and it's this constantly經常 shifting network網絡
43
112272
3287
就是這個不斷移位的網路,
02:07
that produces產生 the behavior行為 of the colony殖民地,
44
115559
2520
產生蟻群的行為,
02:10
like whether是否 all the ants螞蟻 are hiding inside the nest,
45
118079
2881
像是不是所有的螞蟻都躲在窩裡,
02:12
or how many許多 are going out to forage飼料.
46
120960
2694
或有多少在外面覓食。
02:15
A brain actually其實 works作品 in the same相同 way,
47
123654
1759
大腦其實也以同樣的方法運作,
02:17
but what's great about ants螞蟻 is
48
125413
1500
但研究螞蟻最棒的地方,
02:18
that you can see the whole整個 network網絡 as it happens發生.
49
126913
4867
在於你能親眼目睹整個網路形成。
02:23
There are more than 12,000 species種類 of ants螞蟻,
50
131780
3100
有超過一萬兩千種螞蟻,
02:26
in every一切 conceivable可以想像 environment環境,
51
134880
2125
存在於每一種可想到的環境中,
02:29
and they're using運用 interactions互動 differently不同
52
137005
2649
而牠們互動的方法也不同,
02:31
to meet遇到 different不同 environmental環境的 challenges挑戰.
53
139654
2768
以因應不同的環境挑戰。
02:34
So one important重要 environmental環境的 challenge挑戰
54
142422
2305
所以有一種很重要的環境挑戰,
02:36
that every一切 system系統 has to deal合同 with
55
144727
1870
是每一種系統都必須面對的,
02:38
is operating操作 costs成本, just what it takes
56
146597
2083
就是營業成本,就是到底要花多少
02:40
to run the system系統.
57
148680
2265
來經營系統。
02:42
And another另一個 environmental環境的 challenge挑戰 is resources資源,
58
150945
2548
而另一種環境挑戰則是資源,
02:45
finding發現 them and collecting蒐集 them.
59
153493
2336
要找尋及收集資源。
02:47
In the desert沙漠, operating操作 costs成本 are high
60
155829
2783
在沙漠裡,營業成本很高,
02:50
because water is scarce稀缺,
61
158612
1773
因為水很稀少,
02:52
and the seed-eating種子為食 ants螞蟻 that I study研究 in the desert沙漠
62
160385
2499
而且我在沙漠中研究的
一種吃種子的螞蟻,
02:54
have to spend water to get water.
63
162884
2900
必須先用掉水才能得到水。
02:57
So an ant螞蟻 outside foraging覓食,
64
165784
2139
所以一隻螞蟻在外面覓食,
02:59
searching搜索 for seeds種子 in the hot sun太陽,
65
167923
2037
在烈日下尋找種子,
03:01
just loses失去 water into the air空氣.
66
169960
2007
就會失去水分,釋放到空氣中。
03:03
But the colony殖民地 gets得到 its water
67
171967
1664
但這個蟻群會因此得到水份,
03:05
by metabolizing代謝 the fats脂肪 out of the seeds種子
68
173631
1766
也就是從代謝牠們所吃的種子
03:07
that they eat.
69
175397
1934
所含的脂肪以得到水份,
03:09
So in this environment環境, interactions互動 are used
70
177331
3159
所以在這樣的環境下,互動是要
03:12
to activate啟用 foraging覓食.
71
180490
1549
動員覓食行為。
03:14
An outgoing傳出 forager覓食 doesn't go out unless除非
72
182039
2264
指派為覓食者的螞蟻,
03:16
it gets得到 enough足夠 interactions互動 with returning回國 foragers徵糧,
73
184303
2573
在與回來的覓食蟻
得到足夠的互動前不會出去,
03:18
and what you see are the returning回國 foragers徵糧
74
186876
1990
而你現在看到的是回來的覓食蟻,
03:20
going into the tunnel隧道, into the nest,
75
188866
1824
進入隧道中,進入窩裡,
03:22
and meeting會議 outgoing傳出 foragers徵糧 on their way out.
76
190690
2636
與正要出去的覓食蟻互動。
03:25
This makes品牌 sense for the ant螞蟻 colony殖民地,
77
193326
1623
這種方法對蟻群而言很有道理,
03:26
because the more food餐飲 there is out there,
78
194949
2323
因為外面的食物愈多,
03:29
the more quickly很快 the foragers徵糧 find it,
79
197272
1851
覓食蟻就能愈快找到食物,
03:31
the faster更快 they come back,
80
199123
1236
牠們也就愈快回來,
03:32
and the more foragers徵糧 they send發送 out.
81
200359
2801
所以就會送更多的覓食蟻出去。
03:35
The system系統 works作品 to stay stopped停止,
82
203160
2507
這個系統的原理是保持不動,
03:37
unless除非 something positive happens發生.
83
205667
2015
直到有好事發生。
03:39
So interactions互動 function功能 to activate啟用 foragers徵糧.
84
207682
3881
所以互動的功用
在促使覓食蟻開始活動。
03:43
And we've我們已經 been studying研究
the evolution演化 of this system系統.
85
211563
2907
我們一直在研究這個系統的演化。
03:46
First of all, there's variation變異.
86
214470
1091
首先,這系統裡有變數。
03:47
It turns out that colonies群落 are different不同.
87
215561
2479
結果證明每一種蟻群都不一樣。
03:50
On dry days, some colonies群落 forage飼料 less,
88
218040
2696
在乾燥的日子,
某些蟻群的覓食行為會少一點,
03:52
so colonies群落 are different不同 in how
89
220736
1505
所以每個蟻群的不同點在於
03:54
they manage管理 this trade-off交易
90
222241
1354
牠們如何權衡
03:55
between之間 spending開支 water to search搜索 for seeds種子
91
223595
3168
是要花掉水分以尋找種子,
03:58
and getting得到 water back in the form形成 of seeds種子.
92
226763
3320
還是要找種子回來以取得裡面的水。
04:02
And we're trying to understand理解 why
93
230083
1719
我們試圖瞭解為什麼
04:03
some colonies群落 forage飼料 less than others其他
94
231802
2140
某些蟻群的覓食行為較少,
04:05
by thinking思維 about ants螞蟻 as neurons神經元,
95
233942
2325
透過將螞蟻視為神經元的方法,
04:08
using運用 models楷模 from neuroscience神經科學.
96
236267
2338
並使用神經科學的模式。
04:10
So just as a neuron神經元 adds增加 up its stimulation促進
97
238605
2674
所以就像神經元會累積
04:13
from other neurons神經元 to decide決定 whether是否 to fire,
98
241279
2006
從別的神經元送來的刺激,
以決定是否發射,
04:15
an ant螞蟻 adds增加 up its stimulation促進 from other ants螞蟻
99
243285
2885
螞蟻也會累積
從別的螞蟻傳來的刺激,
04:18
to decide決定 whether是否 to forage飼料.
100
246170
2023
以決定是否出去覓食。
04:20
And what we're looking for is whether是否 there might威力 be
101
248193
1637
我們在找的是,是否在
04:21
small differences分歧 among其中 colonies群落
102
249830
1992
各個蟻群間有微小的差異,
04:23
in how many許多 interactions互動 each ant螞蟻 needs需求
103
251822
3314
也就是螞蟻需要多少互動
04:27
before it's willing願意 to go out and forage飼料,
104
255136
1926
牠才願意出去覓食,
04:29
because a colony殖民地 like that would forage飼料 less.
105
257062
3767
因為像那樣的蟻群可能較少覓食。
04:32
And this raises加薪 an analogous類似 question about brains大腦.
106
260829
3139
這也引發大家對大腦
產生類似的問題。
04:35
We talk about the brain,
107
263968
1412
我們談論大腦,
04:37
but of course課程 every一切 brain is slightly different不同,
108
265380
2890
但當然每個大腦都有少許不同,
04:40
and maybe there are some individuals個人
109
268270
1379
而或許有某些人,
04:41
or some conditions條件
110
269649
1319
或在某些情況下,
04:42
in which哪一個 the electrical電動 properties性能 of neurons神經元 are such這樣
111
270968
3222
神經元的電性質也像這樣,
04:46
that they require要求 more stimulus刺激物 to fire,
112
274190
3870
需要更多的刺激才能發射,
04:50
and that would lead to differences分歧 in brain function功能.
113
278060
3726
而那會導致大腦功能產生差異。
04:53
So in order訂購 to ask evolutionary發展的 questions問題,
114
281786
2334
所以為了要問進化的問題,
04:56
we need to know about reproductive生殖 success成功.
115
284120
2679
我們必須瞭解生殖成功率。
04:58
This is a map地圖 of the study研究 site現場
116
286799
2325
這是研究地點的地圖,
05:01
where I have been tracking追踪 this population人口
117
289124
2553
我已經在那裡追蹤這種
05:03
of harvester收割機 ant螞蟻 colonies群落 for 28 years年份,
118
291677
3293
收割蟻群 28 年了,
05:06
which哪一個 is about as long as a colony殖民地 lives生活.
119
294970
2262
這大約就是一個蟻群的壽命。
05:09
Each symbol符號 is a colony殖民地,
120
297232
2099
每一個符號代表一個蟻群,
05:11
and the size尺寸 of the symbol符號 is
how many許多 offspring子孫 it had,
121
299331
3266
而符號的大小代表
這個蟻群有多少後代,
05:14
because we were able能夠 to use genetic遺傳 variation變異
122
302597
1913
因為我們能用遺傳變異
05:16
to match比賽 up parent and offspring子孫 colonies群落,
123
304510
2711
去匹配親子蟻群,
05:19
that is, to figure數字 out which哪一個 colonies群落
124
307221
3297
也就是說,去找出哪個蟻群
05:22
were founded成立 by a daughter女兒 queen女王
125
310518
1925
是哪一隻蟻后的女兒
05:24
produced生成 by which哪一個 parent colony殖民地.
126
312443
2068
所產生的。
05:26
And this was amazing驚人 for me, after all these years年份,
127
314511
2002
這對我而言很奇妙,在這些年之後,
05:28
to find out, for example, that colony殖民地 154,
128
316513
3128
能發現到,舉個例,蟻群 154,
05:31
whom I've known已知 well for many許多 years年份,
129
319641
2176
多年來我非常瞭解的一個蟻群,
05:33
is a great-grandmother曾祖母.
130
321817
1819
居然是曾祖母。
05:35
Here's這裡的 her daughter女兒 colony殖民地,
131
323636
1740
這是牠的女兒的蟻群,
05:37
here's這裡的 her granddaughter孫女 colony殖民地,
132
325376
2684
這是牠的孫女的蟻群,
05:40
and these are her great-granddaughter曾孫女 colonies群落.
133
328060
2392
而這些是牠的曾孫女的蟻群。
05:42
And by doing this, I was able能夠 to learn學習
134
330452
2025
經由這麼做,我學到
05:44
that offspring子孫 colonies群落 resemble類似 parent colonies群落
135
332477
3283
後代蟻群與親代蟻群,
05:47
in their decisions決定 about which哪一個 days are so hot
136
335760
2217
在決定哪一天太熱,
05:49
that they don't forage飼料,
137
337977
1778
不出去覓食這方面很相似,
05:51
and the offspring子孫 of parent colonies群落
138
339755
1464
而親代蟻群的後代,
05:53
live生活 so far from each other that the ants螞蟻 never meet遇到,
139
341219
2906
兩者住得很遠,
這些螞蟻從來沒有見過面,
05:56
so the ants螞蟻 of the offspring子孫 colony殖民地
140
344125
2289
所以後代蟻群的螞蟻,
05:58
can't be learning學習 this from the parent colony殖民地.
141
346414
2244
不可能從牠們的親代蟻群學到這個。
06:00
And so our next下一個 step is to look
142
348658
1381
因此我們下一步要去找
06:02
for the genetic遺傳 variation變異
underlying底層 this resemblance相似.
143
350039
5276
產生這種相似度的遺傳變異。
06:07
So then I was able能夠 to ask, okay, who's誰是 doing better?
144
355315
4125
所以我才能問,好,誰做得比較好?
06:11
Over the time of the study研究,
145
359440
1460
在做研究的這段時間,
06:12
and especially特別 in the past過去 10 years年份,
146
360900
1465
特別是過去十年,
06:14
there's been a very severe嚴重 and deepening深化 drought乾旱
147
362365
3308
曾有非常嚴重且日益加深的乾旱,
06:17
in the Southwestern西南 U.S.,
148
365673
2127
發生在美國西南部,
06:19
and it turns out that the
colonies群落 that conserve養護 water,
149
367800
3053
結果是這些節約用水的蟻群,
06:22
that stay in when it's really hot outside,
150
370853
4429
就是在外面真的
很熱的時候還留在窩裡,
06:27
and thus從而 sacrifice犧牲 getting得到 as much food餐飲 as possible可能,
151
375282
2632
從而犧牲盡可能找最多食物的蟻群,
06:29
are the ones那些 more likely容易 to have offspring子孫 colonies群落.
152
377914
2945
愈有可能產生後代蟻群。
06:32
So all this time, I thought that colony殖民地 154
153
380859
2269
所以這些時間,我以為蟻群 154
06:35
was a loser失敗者, because on really dry days,
154
383128
2670
是輸家,因為在很乾燥的日子,
06:37
there'd這紅色 be just this trickle of foraging覓食,
155
385798
1868
牠們只有稀稀落落的覓食行為,
06:39
while the other colonies群落 were out
156
387666
1591
而其他的蟻群都跑出去
06:41
foraging覓食, getting得到 lots of food餐飲,
157
389257
2117
覓食,獲得大量的食物,
06:43
but in fact事實, colony殖民地 154 is a huge巨大 success成功.
158
391374
3020
但其實,蟻群 154 非常成功。
06:46
She's a matriarch女家長.
159
394394
1332
她是女族長。
06:47
She's one of the rare罕見 great-grandmothers曾祖母 on the site現場.
160
395726
3052
她是當地少有的曾祖母輩之一。
06:50
To my knowledge知識, this is the first time
161
398778
2785
據我所知,這也是第一次
06:53
that we've我們已經 been able能夠 to track跟踪
162
401563
1639
我們能追蹤
06:55
the ongoing不斷的 evolution演化 of collective集體 behavior行為
163
403202
3001
在自然的動物族群中
06:58
in a natural自然 population人口 of animals動物
164
406203
2117
持續進行的集體行為演化,
07:00
and find out what's actually其實 working加工 best最好.
165
408320
4657
並找出什麼是最佳的運作方式。
07:04
Now, the Internet互聯網 uses使用 an algorithm算法
166
412977
2421
那麼,網際網路使用一種演算法
07:07
to regulate調節 the flow of data數據
167
415398
2853
以管理資料流,
07:10
that's very similar類似 to the one
168
418251
2227
與這個非常相似,
07:12
that the harvester收割機 ants螞蟻 are using運用 to regulate調節
169
420478
2376
就是收獲蟻在使用以管理
07:14
the flow of foragers徵糧.
170
422854
1541
覓食蟻的流程。
07:16
And guess猜測 what we call this analogy比喻?
171
424395
3366
你猜我們怎麼叫這種類比?
07:19
The anternetanternet is coming未來.
172
427761
1518
蟻際網路來了!
07:21
(Applause掌聲)
173
429279
1721
(掌聲)
07:23
So data數據 doesn't leave離開 the source資源 computer電腦
174
431000
3454
所以資料不會從源計算機輸出,
07:26
unless除非 it gets得到 a signal信號 that there's enough足夠 bandwidth帶寬
175
434454
2861
直到它得到信號,有足夠的頻寬
07:29
for it to travel旅行 on.
176
437315
2729
讓資料傳出去。
07:32
In the early days of the Internet互聯網,
177
440044
1441
在網際網路發展早期,
07:33
when operating操作 costs成本 were really high
178
441485
2274
當營業成本還很高,
07:35
and it was really important重要 not to lose失去 any data數據,
179
443759
3227
而且很重要
不能失去任何資料的時期,
07:38
then the system系統 was set up for interactions互動
180
446986
2157
系統被設成互動
07:41
to activate啟用 the flow of data數據.
181
449143
3062
以啟動資料流。
07:44
It's interesting有趣 that the ants螞蟻 are using運用 an algorithm算法
182
452205
2385
有趣的是螞蟻會使用一種演算法,
07:46
that's so similar類似 to the one that we recently最近 invented發明,
183
454590
3896
與我們最近發明的非常相似,
07:50
but this is only one of a handful少數 of ant螞蟻 algorithms算法
184
458486
2953
但這只是我們所知的
07:53
that we know about,
185
461439
1419
螞蟻演算法之一罷了,
07:54
and ants螞蟻 have had 130 million百萬 years年份
186
462858
3193
螞蟻有一億三千萬年的時間
07:58
to evolve發展 a lot of good ones那些,
187
466051
2026
發展很多很好的演算法,
08:00
and I think it's very likely容易
188
468077
1506
我想非常有可能
08:01
that some of the other 12,000 species種類
189
469583
2557
在其他一萬二千種螞蟻中,
08:04
are going to have interesting有趣 algorithms算法
190
472140
2697
能找到有趣的演算法
08:06
for data數據 networks網絡
191
474837
1024
給資料網路使用,
08:07
that we haven't沒有 even thought of yet然而.
192
475861
2697
是我們想都沒想過的。
08:10
So what happens發生 when operating操作 costs成本 are low?
193
478558
3085
所以,營業成本低的時候
會發生什麼?
08:13
Operating操作 costs成本 are low in the tropics熱帶,
194
481643
1787
熱帶地方的營業成本低,
08:15
because it's very humid, and it's easy簡單 for the ants螞蟻
195
483430
2096
因為那裡很濕,螞蟻很容易
08:17
to be outside walking步行 around.
196
485526
2824
在外面走來走去。
08:20
But the ants螞蟻 are so abundant豐富
197
488350
1653
但在熱帶地方,
08:22
and diverse多種 in the tropics熱帶
198
490003
1818
螞蟻量很大,種類極多,
08:23
that there's a lot of competition競爭.
199
491821
2598
因此競爭也很激烈。
08:26
Whatever隨你 resource資源 one species種類 is using運用,
200
494419
1952
某種螞蟻在使用的資源,
08:28
another另一個 species種類 is likely容易 to be using運用 that
201
496371
3172
別的螞蟻可能在同時間
08:31
at the same相同 time.
202
499543
2379
也需要使用。
08:33
So in this environment環境, interactions互動 are used
203
501922
2608
所以在這種環境,互動的使用法
08:36
in the opposite對面 way.
204
504530
1945
恰好相反。
08:38
The system系統 keeps保持 going
205
506475
1395
系統要持續運轉,
08:39
unless除非 something negative happens發生,
206
507870
1554
直到負面事件發生。
08:41
and one species種類 that I study研究 makes品牌 circuits電路
207
509424
2167
而我在研究的某種螞蟻
還會形成迴路,
08:43
in the trees樹木 of foraging覓食 ants螞蟻
208
511591
2159
覓食蟻在蟻窩到食物源
08:45
going from the nest to a food餐飲 source資源 and back,
209
513750
2981
往來的樹上
08:48
just round回合 and round回合,
210
516731
1329
不斷繞圈圈,
08:50
unless除非 something negative happens發生,
211
518060
1442
直到某件負面事件發生為止,
08:51
like an interaction相互作用
212
519502
1609
像是與別種螞蟻
08:53
with ants螞蟻 of another另一個 species種類.
213
521111
2739
有了互動。
08:55
So here's這裡的 an example of ant螞蟻 security安全.
214
523850
2937
所以這是螞蟻的保全案例。
08:58
In the middle中間, there's an ant螞蟻
215
526787
1858
在中間有一隻螞蟻
09:00
plugging堵漏 the nest entrance入口 with its head
216
528645
2463
正以牠的頭堵住窩的入口,
09:03
in response響應 to interactions互動 with another另一個 species種類.
217
531108
2993
回應牠與另一種螞蟻互動的結果。
09:06
Those are the little ones那些 running賽跑 around
218
534101
1659
這些小東西跑來跑去
09:07
with their abdomens腹部 up in the air空氣.
219
535760
2751
腹部朝上。
09:10
But as soon不久 as the threat威脅 is passed通過,
220
538511
2046
一但威脅消失,
09:12
the entrance入口 is open打開 again,
221
540557
2755
入口就又打開了,
09:15
and maybe there are situations情況
222
543312
1790
也許在某些情況下,
09:17
in computer電腦 security安全
223
545102
1089
電腦的安全性
09:18
where operating操作 costs成本 are low enough足夠
224
546191
2206
在營運成本夠低時,
09:20
that we could just block access訪問 temporarily暫時
225
548397
3375
我們只要暫時把存取擋住,
09:23
in response響應 to an immediate即時 threat威脅,
226
551772
2193
以回應立即性的威脅,
09:25
and then open打開 it again,
227
553965
2026
然後再把它打開就好,
09:27
instead代替 of trying to build建立
228
555991
1269
而不用試著建立
09:29
a permanent常駐 firewall火牆 or fortress堡壘.
229
557260
3980
一個永久性的防火牆或要塞。
09:33
So another另一個 environmental環境的 challenge挑戰
230
561240
1940
那麼另一種環境挑戰,
09:35
that all systems系統 have to deal合同 with
231
563180
1695
所有系統都要面對的,
09:36
is resources資源, finding發現 and collecting蒐集 them.
232
564875
5452
是資源,要找尋及收集資源。
09:42
And to do this, ants螞蟻 solve解決 the problem問題
233
570327
1673
要做這個,螞蟻解決了
09:44
of collective集體 search搜索,
234
572000
1258
集體搜尋的問題,
09:45
and this is a problem問題 that's of great interest利益
235
573258
1576
而這在機器人科學
09:46
right now in robotics機器人,
236
574834
1484
是目前極感興趣的問題,
09:48
because we've我們已經 understood了解 that,
237
576318
1732
因為我們已經瞭解,
09:50
rather than sending發出 a single,
238
578050
1614
與其送出一個單獨操作、
09:51
sophisticated複雜的, expensive昂貴 robot機器人 out
239
579664
3495
複雜又很貴的機器人出去
09:55
to explore探索 another另一個 planet行星
240
583159
1417
探險另一個星球,
09:56
or to search搜索 a burning燃燒 building建造,
241
584576
2567
或去搜索一棟燃燒的建築物,
09:59
that instead代替, it may可能 be more effective有效
242
587143
2600
還不如這樣可能更有效,
10:01
to get a group of cheaper便宜 robots機器人
243
589743
4377
就是找一組便宜的機器人
10:06
exchanging交換 only minimal最小 information信息,
244
594120
2545
只交換最少的資訊,
10:08
and that's the way that ants螞蟻 do it.
245
596665
2799
而那正是螞蟻做的方式。
10:11
So the invasive侵入的 Argentine阿根廷 ant螞蟻
246
599464
1765
所以入侵種阿根廷蟻
10:13
makes品牌 expandable擴張 search搜索 networks網絡.
247
601229
2330
建造可擴充的搜尋網路。
10:15
They're good at dealing交易 with the main主要 problem問題
248
603559
2273
牠們對解決集體搜尋的
10:17
of collective集體 search搜索,
249
605832
1331
主要問題很在行,
10:19
which哪一個 is the trade-off交易 between之間
250
607163
2566
那是權衡了
10:21
searching搜索 very thoroughly
251
609729
1336
要完全徹底的搜尋,
10:23
and covering覆蓋 a lot of ground地面.
252
611065
1997
還是要涵蓋大片土地的結果。
10:25
And what they do is,
253
613062
895
牠們所做的是,
10:25
when there are many許多 ants螞蟻 in a small space空間,
254
613957
2387
如果在一個小空間裡有很多螞蟻,
10:28
then each one can search搜索 very thoroughly
255
616344
2213
那麼每一隻都可以徹底搜尋,
10:30
because there will be another另一個 ant螞蟻 nearby附近
256
618557
1624
因為附近一定會有另一隻螞蟻
10:32
searching搜索 over there,
257
620181
1357
在另一邊搜尋,
10:33
but when there are a few少數 ants螞蟻
258
621538
1647
但當在一塊很大的空間裡,
10:35
in a large space空間,
259
623185
2055
只有少數的螞蟻的時候,
10:37
then they need to stretch伸展 out their paths路徑
260
625240
2414
那麼牠們需要延伸牠們的路徑
10:39
to cover more ground地面.
261
627654
1803
以涵蓋更多的地面。
10:41
I think they use interactions互動 to assess評估 density密度,
262
629457
2954
我想牠們使用互動以評估密度,
10:44
so when they're really crowded,
263
632411
1229
所以當牠們真的很擠的時候,
10:45
they meet遇到 more often經常,
264
633640
1102
牠們就更常碰到彼此,
10:46
and they search搜索 more thoroughly.
265
634742
2465
牠們就更徹底搜尋。
10:49
Different不同 ant螞蟻 species種類 must必須 use different不同 algorithms算法,
266
637207
3400
不同的螞蟻物種
必須使用不同的演算法,
10:52
because they've他們已經 evolved進化 to deal合同 with
267
640607
2522
因為牠們已進化以面對
10:55
different不同 resources資源,
268
643129
1671
不同的資源,
10:56
and it could be really useful有用 to know about this,
269
644800
2559
而知道這一點可能非常有用,
10:59
and so we recently最近 asked ants螞蟻
270
647359
1642
所以我們最近要求螞蟻
11:01
to solve解決 the collective集體 search搜索 problem問題
271
649001
2450
解決集體搜尋問題,
11:03
in the extreme極端 environment環境
272
651451
1368
在極端環境下,
11:04
of microgravity微重力
273
652819
1558
即微重力狀態,
11:06
in the International國際 Space空間 Station.
274
654377
1976
在國際太空站裡。
11:08
When I first saw this picture圖片, I thought,
275
656353
1545
當我第一次看到這張照片,我在想,
11:09
Oh no, they've他們已經 mounted安裝 the habitat棲息地 vertically垂直,
276
657898
2857
不會吧,他們居然把棲地裝成垂直的,
11:12
but then I realized實現 that, of course課程, it doesn't matter.
277
660755
2618
但隨後我意識到,當然,那無所謂。
11:15
So the idea理念 here is that the ants螞蟻
278
663373
2637
所以這個想法就是,螞蟻
11:18
are working加工 so hard to hang on
279
666010
1970
要非常努力地緊緊抓住
11:19
to the wall or the floor地板 or whatever隨你 you call it
280
667980
3057
牆壁或地板,隨便你怎麼叫,
11:23
that they're less likely容易 to interact相互作用,
281
671037
3009
所以牠們就不太可能互動,
11:26
and so the relationship關係 between之間
282
674046
1244
如此一來這兩者之間的關係,
11:27
how crowded they are and how often經常 they meet遇到
283
675290
2120
就是有多擠及多常碰到
11:29
would be messed搞砸 up.
284
677410
1725
就會被搞亂。
11:31
We're still analyzing分析 the data數據.
285
679135
1395
我們仍在分析那些數據。
11:32
I don't have the results結果 yet然而.
286
680530
1964
我還沒拿到結果。
11:34
But it would be interesting有趣 to know
287
682494
1694
但這應該滿有趣的,如果你知道
11:36
how other species種類 solve解決 this problem問題
288
684188
2459
其他物種在地球上不同的環境裡,
11:38
in different不同 environments環境 on Earth地球,
289
686647
2564
如何解決這個問題,
11:41
and so we're setting設置 up a program程序
290
689211
1266
所以我們辦了一個計劃,
11:42
to encourage鼓勵 kids孩子 around the world世界
291
690477
2160
鼓勵全世界各地的孩子,
11:44
to try this experiment實驗 with different不同 species種類.
292
692637
2536
以不同的物種試做這個實驗。
11:47
It's very simple簡單.
293
695173
1767
這很簡單。
11:48
It can be doneDONE with cheap低廉 materials物料.
294
696940
2090
用很便宜的材料就能做。
11:51
And that way, we could make a global全球 map地圖
295
699030
2833
這樣一來,我們或許可以
做一張全球地圖,
11:53
of ant螞蟻 collective集體 search搜索 algorithms算法.
296
701863
3347
畫著螞蟻的集體搜尋演算法。
11:57
And I think it's pretty漂亮 likely容易 that the invasive侵入的 species種類,
297
705210
2483
而我想入侵物種很有可能,
11:59
the ones那些 that come into our buildings房屋,
298
707693
2149
就是跑進我們房屋裡的那些,
12:01
are going to be really good at this,
299
709842
1742
在這方面的表現會非常好,
12:03
because they're in your kitchen廚房
300
711584
1848
因為牠們在你的廚房裡,
12:05
because they're really good
at finding發現 food餐飲 and water.
301
713432
3907
因為牠們真的很會找食物及水。
12:09
So the most familiar resource資源 for ants螞蟻
302
717339
3265
所以對螞蟻最熟悉的資源
12:12
is a picnic野餐,
303
720604
1315
就是野餐。
12:13
and this is a clustered集群 resource資源.
304
721919
2145
而且這還是叢集資源。
12:16
When there's one piece of fruit水果,
305
724064
999
有一片水果出現在這裡,
12:17
there's likely容易 to be another另一個 piece of fruit水果 nearby附近,
306
725063
2515
有另一片在附近的可能性就很大。
12:19
and the ants螞蟻 that specialize專攻 on clustered集群 resources資源
307
727578
3432
而那些專做叢集資源的螞蟻
12:23
use interactions互動 for recruitment招聘.
308
731010
1942
就會用互動招募螞蟻大軍。
12:24
So when one ant螞蟻 meets符合 another另一個,
309
732952
1277
所以當某隻螞蟻碰到另一隻螞蟻,
12:26
or when it meets符合 a chemical化學 deposited沉積
310
734229
1625
或當這隻螞蟻碰到另一隻螞蟻
12:27
on the ground地面 by another另一個,
311
735854
1736
留在地面的某種化學物品,
12:29
then it changes變化 direction方向 to follow跟隨
312
737590
1833
那牠就會改變方向,
12:31
in the direction方向 of the interaction相互作用,
313
739423
1573
依照互動的指示,
12:32
and that's how you get the trail落後 of ants螞蟻
314
740996
1977
而這就是你如何得到一條螞蟻線
12:34
sharing分享 your picnic野餐.
315
742973
1413
來分享你的野餐。
12:36
Now this is a place地點 where I think we might威力 be able能夠
316
744386
1695
那這是我認為我們或許可以
12:38
to learn學習 something from ants螞蟻 about cancer癌症.
317
746081
3741
從螞蟻身上學到關於癌症的地方。
12:41
I mean, first, it's obvious明顯 that we could do a lot
318
749822
1981
我是說第一,很明顯我們能做很多事
12:43
to prevent避免 cancer癌症
319
751803
1610
以避免癌症,
12:45
by not allowing允許 people to spread傳播 around
320
753413
2577
藉著不讓人們傳播
12:47
or sell the toxins毒素 that promote促進
321
755990
1944
或販賣會在我們
12:49
the evolution演化 of cancer癌症 in our bodies身體,
322
757934
2780
體內致癌的毒素,
12:52
but I don't think the ants螞蟻 can help us much with this
323
760714
2346
但我不認為螞蟻在這方面
能幫我們多少,
12:55
because ants螞蟻 never poison their own擁有 colonies群落.
324
763060
3358
因為螞蟻從不毒化自己的蟻群。
12:58
But we might威力 be able能夠 to learn學習 something from ants螞蟻
325
766418
1332
但我們或許能從螞蟻學到
12:59
about treating治療 cancer癌症.
326
767750
1763
與治療癌症有關的事。
13:01
There are many許多 different不同 kinds of cancer癌症.
327
769513
2225
癌症有許多不同的類型。
13:03
Each one originates起源 in a particular特定 part部分 of the body身體,
328
771738
2978
每一種都源自身體的特定部位,
13:06
and then some kinds of cancer癌症 will spread傳播
329
774716
2966
然後某些癌症會擴散
13:09
or metastasize轉移 to particular特定 other tissues組織
330
777682
2830
或轉移到某些特定的組織,
13:12
where they must必須 be getting得到
resources資源 that they need.
331
780512
2880
這些癌細胞一定是在那裡
得到它們所需的資源。
13:15
So if you think from the perspective透視
332
783392
1808
所以如果你從
13:17
of early metastatic轉移性 cancer癌症 cells細胞
333
785200
1950
癌細胞早期轉移的角度想,
13:19
as they're out searching搜索 around
334
787150
1623
好像它們是去外面四處尋找
13:20
for the resources資源 that they need,
335
788773
2317
它們所需的資源,
13:23
if those resources資源 are clustered集群,
336
791090
1983
假如這些資源是叢集的,
13:25
they're likely容易 to use interactions互動 for recruitment招聘,
337
793073
3013
它們就很有可能使用互動
來招募其它細胞,
13:28
and if we can figure數字 out how
cancer癌症 cells細胞 are recruiting招聘,
338
796086
3093
如果我們能找出癌細胞
如何招兵買馬,
13:31
then maybe we could set traps陷阱
339
799179
2347
那也許我們可以設下陷阱,
13:33
to catch抓住 them before they become成為 established既定.
340
801526
4049
在它們變得穩定前捕捉它們。
13:37
So ants螞蟻 are using運用 interactions互動 in different不同 ways方法
341
805575
3235
所以螞蟻在大不同的各種環境下,
13:40
in a huge巨大 variety品種 of environments環境,
342
808810
2602
以不同的方法互動,
13:43
and we could learn學習 from this
343
811412
1821
而我們可以從中學到
13:45
about other systems系統 that operate操作
344
813233
1777
其他系統如何運作,
13:47
without central中央 control控制.
345
815010
2337
不靠中央控制。
13:49
Using運用 only simple簡單 interactions互動,
346
817347
1979
僅僅使用簡單的互動,
13:51
ant螞蟻 colonies群落 have been performing執行
347
819326
1795
蟻群已經執行
13:53
amazing驚人 feats功勳 for more than 130 million百萬 years年份.
348
821121
3633
驚人壯舉超過一億三千萬年。
13:56
We have a lot to learn學習 from them.
349
824754
2140
我們還有很多要向牠們學習。
13:58
Thank you.
350
826894
2738
謝謝。
14:01
(Applause掌聲)
351
829632
2733
(掌聲)
Translated by Regina Chu
Reviewed by Helen Chang

▲Back to top

ABOUT THE SPEAKER
Deborah Gordon - Ecologist
By studying how ant colonies work without any one leader, Deborah Gordon has identified striking similarities in how ant colonies, brains, cells and computer networks regulate themselves.

Why you should listen

Ecologist Deborah M. Gordon has learned that ant colonies can work without central control by using simple interactions like how often the insects touch antennae. Contrary to the notion that colonies are organized by efficient ants, she has instead discovered that evolution has produced “noisy” systems that tolerate accident and respond flexibly to the environment. When conditions are tough, natural selection favors colonies that conserve resources.

Her studies of ant colonies have led her and her Stanford colleagues to the discovery of the “Anternet,” which regulates foraging in ants in the same way the internet regulates data traffic. But as she said to Wired in 2013, "Insect behavior mimicking human networks ... is actually not what’s most interesting about ant networks. What’s far more interesting are the parallels in the other direction: What have the ants worked out that we humans haven’t thought of yet?" Her latest exploration: How do ants behave in space?

More profile about the speaker
Deborah Gordon | Speaker | TED.com