TED2013
Eric Berlow and Sean Gourley: Mapping ideas worth spreading
Filmed:
Readability: 4.3
1,131,373 views
What do 24,000 ideas look like? Ecologist Eric Berlow and physicist Sean Gourley apply algorithms to the entire archive of TEDx Talks, taking us on a stimulating visual tour to show how ideas connect globally.
Eric Berlow - Ecologist
TED Senior Fellow Eric Berlow studies ecology and networks, exposing the interconnectedness of our ecosystems with climate change, government, corporations and more. Full bioSean Gourley - Physicist and military theorist
Sean Gourley, trained as a physicist, has turned his scientific mind to analyzing data about a messier topic: modern war and conflict. He is a TED Fellow. Full bio
TED Senior Fellow Eric Berlow studies ecology and networks, exposing the interconnectedness of our ecosystems with climate change, government, corporations and more. Full bioSean Gourley - Physicist and military theorist
Sean Gourley, trained as a physicist, has turned his scientific mind to analyzing data about a messier topic: modern war and conflict. He is a TED Fellow. Full bio
Double-click the English transcript below to play the video.
00:12
Eric Berlow: I'm an ecologist, and Sean's a physicist,
0
562
3061
00:15
and we both study complex networks.
1
3623
2108
00:17
And we met a couple years ago when we discovered
2
5731
1835
00:19
that we had both given a short TED Talk
3
7566
2000
00:21
about the ecology of war,
4
9566
2303
00:23
and we realized that we were connected
5
11869
1447
00:25
by the ideas we shared before we ever met.
6
13316
2818
00:28
And then we thought, you know, there are thousands
7
16134
1556
00:29
of other talks out there, especially TEDx Talks,
8
17690
2114
00:31
that are popping up all over the world.
9
19804
2211
00:34
How are they connected,
10
22015
923
00:34
and what does that global conversation look like?
11
22938
2010
00:36
So Sean's going to tell you a little bit about how we did that.
12
24948
2810
00:39
Sean Gourley: Exactly. So we took 24,000 TEDx Talks
13
27758
3767
00:43
from around the world, 147 different countries,
14
31525
3046
00:46
and we took these talks and we wanted to find
15
34571
2123
00:48
the mathematical structures that underly
16
36694
2040
00:50
the ideas behind them.
17
38734
1722
00:52
And we wanted to do that so we could see how
18
40456
1370
00:53
they connected with each other.
19
41826
2053
00:55
And so, of course, if you're going to do this kind of stuff,
20
43879
1676
00:57
you need a lot of data.
21
45555
956
00:58
So the data that you've got is a great thing called YouTube,
22
46511
3686
01:02
and we can go down and basically pull
23
50197
1768
01:03
all the open information from YouTube,
24
51965
2267
01:06
all the comments, all the views, who's watching it,
25
54232
2349
01:08
where are they watching it, what are they saying in the comments.
26
56581
2779
01:11
But we can also pull up, using speech-to-text translation,
27
59360
3292
01:14
we can pull the entire transcript,
28
62652
2128
01:16
and that works even for people with kind of funny accents like myself.
29
64780
2680
01:19
So we can take their transcript
30
67460
2106
01:21
and actually do some pretty cool things.
31
69566
2098
01:23
We can take natural language processing algorithms
32
71664
2160
01:25
to kind of read through with a computer, line by line,
33
73824
2629
01:28
extracting key concepts from this.
34
76453
2359
01:30
And we take those key concepts and they sort of form
35
78812
2525
01:33
this mathematical structure of an idea.
36
81337
3565
01:36
And we call that the meme-ome.
37
84902
1757
01:38
And the meme-ome, you know, quite simply,
38
86659
2151
01:40
is the mathematics that underlies an idea,
39
88810
2426
01:43
and we can do some pretty interesting analysis with it,
40
91236
1932
01:45
which I want to share with you now.
41
93168
1981
01:47
So each idea has its own meme-ome,
42
95149
2190
01:49
and each idea is unique with that,
43
97339
1951
01:51
but of course, ideas, they borrow from each other,
44
99290
2488
01:53
they kind of steal sometimes,
45
101778
1184
01:54
and they certainly build on each other,
46
102962
1827
01:56
and we can go through mathematically
47
104789
1616
01:58
and take the meme-ome from one talk
48
106405
1840
02:00
and compare it to the meme-ome from every other talk,
49
108245
2454
02:02
and if there's a similarity between the two of them,
50
110699
1973
02:04
we can create a link and represent that as a graph,
51
112672
3250
02:07
just like Eric and I are connected.
52
115922
2394
02:10
So that's theory, that's great.
53
118316
1394
02:11
Let's see how it works in actual practice.
54
119710
2526
02:14
So what we've got here now is the global footprint
55
122236
2788
02:17
of all the TEDx Talks over the last four years
56
125024
2293
02:19
exploding out around the world
57
127317
1550
02:20
from New York all the way down to little old New Zealand in the corner.
58
128867
3329
02:24
And what we did on this is we analyzed the top 25 percent of these,
59
132196
3835
02:28
and we started to see where the connections occurred,
60
136031
2534
02:30
where they connected with each other.
61
138565
1537
02:32
Cameron Russell talking about image and beauty
62
140102
1874
02:33
connected over into Europe.
63
141976
1575
02:35
We've got a bigger conversation about Israel and Palestine
64
143551
2412
02:37
radiating outwards from the Middle East.
65
145963
2255
02:40
And we've got something a little broader
66
148218
1298
02:41
like big data with a truly global footprint
67
149516
2156
02:43
reminiscent of a conversation
68
151672
2179
02:45
that is happening everywhere.
69
153851
2016
02:47
So from this, we kind of run up against the limits
70
155867
2173
02:50
of what we can actually do with a geographic projection,
71
158040
2530
02:52
but luckily, computer technology allows us to go out
72
160570
2052
02:54
into multidimensional space.
73
162622
1546
02:56
So we can take in our network projection
74
164168
1875
02:58
and apply a physics engine to this,
75
166043
1750
02:59
and the similar talks kind of smash together,
76
167793
1885
03:01
and the different ones fly apart,
77
169678
2004
03:03
and what we're left with is something quite beautiful.
78
171682
2072
03:05
EB: So I want to just point out here that every node is a talk,
79
173754
2957
03:08
they're linked if they share similar ideas,
80
176711
2589
03:11
and that comes from a machine reading
81
179300
2084
03:13
of entire talk transcripts,
82
181384
2067
03:15
and then all these topics that pop out,
83
183451
2231
03:17
they're not from tags and keywords.
84
185682
1790
03:19
They come from the network structure
85
187472
1725
03:21
of interconnected ideas. Keep going.
86
189197
2168
03:23
SG: Absolutely. So I got a little quick on that,
87
191365
2022
03:25
but he's going to slow me down.
88
193387
1475
03:26
We've got education connected to storytelling
89
194862
2034
03:28
triangulated next to social media.
90
196896
1643
03:30
You've got, of course, the human brain right next to healthcare,
91
198539
2475
03:33
which you might expect,
92
201014
1386
03:34
but also you've got video games, which is sort of adjacent,
93
202400
2395
03:36
as those two spaces interface with each other.
94
204795
2740
03:39
But I want to take you into one cluster
95
207535
1535
03:41
that's particularly important to me, and that's the environment.
96
209070
2868
03:43
And I want to kind of zoom in on that
97
211938
1493
03:45
and see if we can get a little more resolution.
98
213431
2363
03:47
So as we go in here, what we start to see,
99
215794
2347
03:50
apply the physics engine again,
100
218141
1504
03:51
we see what's one conversation
101
219645
1676
03:53
is actually composed of many smaller ones.
102
221321
2560
03:55
The structure starts to emerge
103
223881
1929
03:57
where we see a kind of fractal behavior
104
225810
2070
03:59
of the words and the language that we use
105
227880
1619
04:01
to describe the things that are important to us
106
229499
1702
04:03
all around this world.
107
231201
1433
04:04
So you've got food economy and local food at the top,
108
232634
2332
04:06
you've got greenhouse gases, solar and nuclear waste.
109
234966
2719
04:09
What you're getting is a range of smaller conversations,
110
237685
2631
04:12
each connected to each other through the ideas
111
240316
2301
04:14
and the language they share,
112
242617
1301
04:15
creating a broader concept of the environment.
113
243918
2450
04:18
And of course, from here, we can go
114
246368
1532
04:19
and zoom in and see, well, what are young people looking at?
115
247900
3534
04:23
And they're looking at energy technology and nuclear fusion.
116
251434
2345
04:25
This is their kind of resonance
117
253779
1674
04:27
for the conversation around the environment.
118
255453
2406
04:29
If we split along gender lines,
119
257859
1899
04:31
we can see females resonating heavily
120
259758
1987
04:33
with food economy, but also out there in hope and optimism.
121
261745
3645
04:37
And so there's a lot of exciting stuff we can do here,
122
265390
2482
04:39
and I'll throw to Eric for the next part.
123
267872
1762
04:41
EB: Yeah, I mean, just to point out here,
124
269634
1602
04:43
you cannot get this kind of perspective
125
271236
1538
04:44
from a simple tag search on YouTube.
126
272774
3360
04:48
Let's now zoom back out to the entire global conversation
127
276134
4188
04:52
out of environment, and look at all the talks together.
128
280322
2534
04:54
Now often, when we're faced with this amount of content,
129
282856
2927
04:57
we do a couple of things to simplify it.
130
285783
2431
05:00
We might just say, well,
131
288214
1314
05:01
what are the most popular talks out there?
132
289528
2829
05:04
And a few rise to the surface.
133
292357
1397
05:05
There's a talk about gratitude.
134
293754
1828
05:07
There's another one about personal health and nutrition.
135
295582
3344
05:10
And of course, there's got to be one about porn, right?
136
298926
2929
05:13
And so then we might say, well, gratitude, that was last year.
137
301855
3234
05:17
What's trending now? What's the popular talk now?
138
305089
2522
05:19
And we can see that the new, emerging, top trending topic
139
307611
3321
05:22
is about digital privacy.
140
310932
2666
05:25
So this is great. It simplifies things.
141
313598
1693
05:27
But there's so much creative content
142
315291
1827
05:29
that's just buried at the bottom.
143
317118
1921
05:31
And I hate that. How do we bubble stuff up to the surface
144
319039
3318
05:34
that's maybe really creative and interesting?
145
322357
2458
05:36
Well, we can go back to the network structure of ideas
146
324815
2931
05:39
to do that.
147
327746
1430
05:41
Remember, it's that network structure
148
329176
2114
05:43
that is creating these emergent topics,
149
331290
2268
05:45
and let's say we could take two of them,
150
333558
1515
05:47
like cities and genetics, and say, well, are there any talks
151
335073
3047
05:50
that creatively bridge these two really different disciplines.
152
338120
2569
05:52
And that's -- Essentially, this kind of creative remix
153
340689
2275
05:54
is one of the hallmarks of innovation.
154
342964
1840
05:56
Well here's one by Jessica Green
155
344804
1606
05:58
about the microbial ecology of buildings.
156
346410
2379
06:00
It's literally defining a new field.
157
348789
2010
06:02
And we could go back to those topics and say, well,
158
350799
2103
06:04
what talks are central to those conversations?
159
352902
2768
06:07
In the cities cluster, one of the most central
160
355670
1690
06:09
was one by Mitch Joachim about ecological cities,
161
357360
3952
06:13
and in the genetics cluster,
162
361312
1720
06:15
we have a talk about synthetic biology by Craig Venter.
163
363032
3193
06:18
These are talks that are linking many talks within their discipline.
164
366225
3353
06:21
We could go the other direction and say, well,
165
369578
1843
06:23
what are talks that are broadly synthesizing
166
371421
2272
06:25
a lot of different kinds of fields.
167
373693
1448
06:27
We used a measure of ecological diversity to get this.
168
375141
2533
06:29
Like, a talk by Steven Pinker on the history of violence,
169
377674
2736
06:32
very synthetic.
170
380410
1180
06:33
And then, of course, there are talks that are so unique
171
381590
2078
06:35
they're kind of out in the stratosphere, in their own special place,
172
383668
3090
06:38
and we call that the Colleen Flanagan index.
173
386758
2514
06:41
And if you don't know Colleen, she's an artist,
174
389272
3034
06:44
and I asked her, "Well, what's it like out there
175
392306
1543
06:45
in the stratosphere of our idea space?"
176
393849
1672
06:47
And apparently it smells like bacon.
177
395521
3255
06:50
I wouldn't know.
178
398776
1791
06:52
So we're using these network motifs
179
400567
2248
06:54
to find talks that are unique,
180
402815
1186
06:56
ones that are creatively synthesizing a lot of different fields,
181
404001
2710
06:58
ones that are central to their topic,
182
406711
1659
07:00
and ones that are really creatively bridging disparate fields.
183
408370
3374
07:03
Okay? We never would have found those with our obsession
184
411744
2102
07:05
with what's trending now.
185
413846
2313
07:08
And all of this comes from the architecture of complexity,
186
416159
2886
07:11
or the patterns of how things are connected.
187
419045
2960
07:14
SG: So that's exactly right.
188
422005
1625
07:15
We've got ourselves in a world
189
423630
2479
07:18
that's massively complex,
190
426109
2044
07:20
and we've been using algorithms to kind of filter it down
191
428153
2867
07:23
so we can navigate through it.
192
431020
1786
07:24
And those algorithms, whilst being kind of useful,
193
432806
2338
07:27
are also very, very narrow, and we can do better than that,
194
435144
3476
07:30
because we can realize that their complexity is not random.
195
438620
2566
07:33
It has mathematical structure,
196
441186
1954
07:35
and we can use that mathematical structure
197
443140
1803
07:36
to go and explore things like the world of ideas
198
444943
2214
07:39
to see what's being said, to see what's not being said,
199
447157
3000
07:42
and to be a little bit more human
200
450157
1407
07:43
and, hopefully, a little smarter.
201
451564
1867
07:45
Thank you.
202
453431
966
07:46
(Applause)
203
454397
4220
ABOUT THE SPEAKERS
Eric Berlow - EcologistTED Senior Fellow Eric Berlow studies ecology and networks, exposing the interconnectedness of our ecosystems with climate change, government, corporations and more.
Why you should listen
Eric Berlow is an ecologist and network scientist who specializes in not specializing. A TED Senior Fellow, Berlow is recognized for his research on food webs and ecological networks and for creative approaches to complex problems. He was the founding director of the University of California's first environmental science center inside Yosemite National Park, where he continues to develop data-driven approaches to managing natural ecosystems.
In 2012 Berlow founded Vibrant Data Labs, which builds tools to use data for social good. Berlow's current projects range from helping spark an egalitarian personal data economy to protecting endangered amphibians in Yosemite to crowd-sourcing novel insights about human creativity. Berlow holds a Ph.D. from Oregon State University in marine ecology.
Eric Berlow | Speaker | TED.com
Sean Gourley - Physicist and military theorist
Sean Gourley, trained as a physicist, has turned his scientific mind to analyzing data about a messier topic: modern war and conflict. He is a TED Fellow.
Why you should listen
Sean Gourley's twin passions are physics (working on nanoscale blue-light lasers and self-assembled quantum nanowires) and politics (he once ran for a national elected office back home in New Zealand).
A Rhodes scholar, he's spent the past five years working at Oxford on complex adaptive systems and collective intelligent systems -- basically, using data to understand the nature of human conflict. As he puts it, "This research has taken me all over the world from the Pentagon, to the House of Lords, the United Nations and most recently to Iraq". Originally from New Zealand, he now lives in San Francisco, where he is the co-founder and CTO of Quid which is building a global intelligence platform. He's a 2009 TED Fellow.
In December 2009, Gourley and his team's research was published in the scientific journal Nature. He is co-founder and CTO of Quid.
Sean Gourley | Speaker | TED.com