0:00:00.878,0:00:04.403
The world is filled[br]with incredible objects

0:00:04.427,0:00:06.355
and rich cultural heritage.

0:00:07.283,0:00:09.419
And when we get access to them,

0:00:09.443,0:00:11.827
we are blown away, we fall in love.

0:00:12.652,0:00:14.249
But most of the time,

0:00:14.273,0:00:19.065
the world's population is living[br]without real access to arts and culture.

0:00:19.946,0:00:25.299
What might the connections be[br]when we start exploring our heritage,

0:00:25.323,0:00:28.530
the beautiful locations[br]and the art in this world?

0:00:29.427,0:00:32.201
Before we get started[br]in this presentation,

0:00:32.225,0:00:35.200
I just want to take care[br]of a few housekeeping points.

0:00:35.651,0:00:38.999
First, I am no expert in art or culture.

0:00:39.023,0:00:41.749
I fell into this by mistake,[br]but I'm loving it.

0:00:42.356,0:00:45.352
Secondly, all of what[br]I'm going to show you

0:00:45.376,0:00:49.341
belongs to the amazing museums,[br]archives and foundations

0:00:49.365,0:00:50.791
that we partner with.

0:00:50.815,0:00:52.671
None of this belongs to Google.

0:00:53.235,0:00:56.424
And finally, what you see behind me

0:00:56.448,0:00:59.799
is available right now[br]on your mobile phones,

0:00:59.823,0:01:00.997
on your laptops.

0:01:01.021,0:01:03.425
This is our current platform,[br]where you can explore

0:01:03.449,0:01:06.860
thousands of museums[br]and objects at your fingertips,

0:01:06.884,0:01:09.395
in extremely high-definition detail.

0:01:09.419,0:01:12.312
The diversity of the content[br]is what's amazing.

0:01:12.336,0:01:14.505
If we just had European paintings,

0:01:14.529,0:01:16.260
if we just had modern art,

0:01:16.284,0:01:18.086
I think it gets a bit boring.

0:01:18.110,0:01:21.969
For example, this month,[br]we launched the "Black History" channel

0:01:21.993,0:01:24.582
with 82 curated exhibitions,

0:01:24.606,0:01:28.172
which talk about arts and culture[br]in that community.

0:01:28.196,0:01:31.819
We also have some[br]amazing objects from Japan,

0:01:31.843,0:01:35.947
centered around craftsmanship,[br]called "Made in Japan."

0:01:35.971,0:01:39.098
And one of my favorite exhibitions,

0:01:39.122,0:01:41.527
which actually is the idea of my talk,

0:01:41.551,0:01:45.414
is -- I didn't expect to become[br]a fan of Japanese dolls.

0:01:45.438,0:01:48.602
But I am, thanks to this exhibition,

0:01:48.626,0:01:52.111
that has really taught me[br]about the craftsmanship

0:01:52.135,0:01:54.257
behind the soul of a Japanese doll.

0:01:54.281,0:01:56.509
Trust me, it's very exciting.

0:01:56.533,0:01:57.881
Take my word for it.

0:01:57.905,0:01:59.914
So, moving on swiftly.

0:01:59.938,0:02:02.770
One quick thing I wanted[br]to showcase in this platform,

0:02:02.794,0:02:05.729
which you can share with your kids[br]and your friends right now,

0:02:05.753,0:02:10.009
is you can travel to all these[br]amazing institutions virtually, as well.

0:02:10.033,0:02:13.619
One of our recent ideas was with[br]The Guggenheim Museum in New York,

0:02:13.643,0:02:17.001
where you can get a taste[br]of what it might feel like

0:02:17.025,0:02:18.176
to actually be there.

0:02:18.200,0:02:19.967
You can go to the ground floor

0:02:19.991,0:02:22.680
and obviously, most of you,[br]I assume, have been there.

0:02:22.704,0:02:25.441
And you can see the architectural[br]masterpiece that it is.

0:02:25.465,0:02:28.559
But imagine this accessibility[br]for a kid in Bombay

0:02:28.583,0:02:30.303
who's studying architecture,

0:02:30.327,0:02:33.382
who hasn't had a chance[br]to go to The Guggenheim as yet.

0:02:33.406,0:02:36.491
You can obviously look at objects[br]in the Guggenheim Museum,

0:02:36.515,0:02:39.233
you can obviously get into them[br]and so on and so forth.

0:02:39.257,0:02:41.132
There's a lot of information here.

0:02:41.156,0:02:45.005
But this is not the purpose[br]of my talk today.

0:02:45.543,0:02:47.309
This exists right now.

0:02:47.333,0:02:51.259
What we now have are the building blocks[br]to a very exciting future,

0:02:51.283,0:02:52.998
when it comes to arts and culture

0:02:53.022,0:02:55.633
and accessibility to arts and culture.

0:02:55.657,0:03:00.125
So I am joined today onstage[br]by my good friend and artist in residence

0:03:00.149,0:03:02.492
at our office in Paris, Cyril Diagne,

0:03:02.516,0:03:04.645
who is the professor of interactive design

0:03:04.669,0:03:07.306
at ECAL University[br]in Lausanne, Switzerland.

0:03:07.330,0:03:09.944
What Cyril and our team[br]of engineers have been doing

0:03:09.968,0:03:13.103
is trying to find these connections[br]and visualize a few of these.

0:03:13.127,0:03:14.940
So I'm going to go quite quick now.

0:03:14.964,0:03:18.566
This object you see[br]behind me -- oh, just clarification:

0:03:18.590,0:03:20.541
Always, seeing the real thing is better.

0:03:20.565,0:03:23.428
In case people think[br]I'm trying to replicate the real thing.

0:03:23.452,0:03:24.659
So, moving on.

0:03:25.155,0:03:28.671
This object you see behind me[br]is the Venus of Berekhat Ram.

0:03:28.695,0:03:30.771
It's one of the oldest[br]objects in the world,

0:03:30.795,0:03:33.969
found in the Golan Heights[br]around 233,000 years ago,

0:03:33.993,0:03:37.122
and currently residing[br]at the Israel Museum in Jerusalem.

0:03:37.146,0:03:39.767
It is also one of the oldest[br]objects on our platform.

0:03:40.322,0:03:41.668
So let's zoom.

0:03:41.692,0:03:43.457
We start from this one object.

0:03:43.986,0:03:46.522
What if we zoomed out

0:03:46.546,0:03:51.052
and actually tried to experience[br]our own cultural big bang?

0:03:51.076,0:03:52.654
What might that look like?

0:03:53.474,0:03:57.131
This is what we deal with on a daily basis[br]at the Cultural Institute --

0:03:57.155,0:04:02.117
over six million cultural artifacts[br]curated and given to us by institutions,

0:04:02.141,0:04:04.087
to actually make these connections.

0:04:04.111,0:04:05.510
You can travel through time,

0:04:05.534,0:04:09.045
you can understand more[br]about our society through these.

0:04:09.069,0:04:13.469
You can look at it[br]from the perspective of our planet,

0:04:13.493,0:04:16.128
and try to see how[br]it might look without borders,

0:04:16.152,0:04:18.823
if we just organized art and culture.

0:04:18.847,0:04:21.075
We can also then plot it by time,

0:04:21.099,0:04:25.179
which obviously, for the data geek[br]in me, is very fascinating.

0:04:25.203,0:04:28.589
You can spend hours[br]looking at every decade

0:04:28.613,0:04:31.929
and the contributions[br]in that decade and in those years

0:04:31.953,0:04:34.286
for art, history and cultures.

0:04:34.310,0:04:37.689
We would love to spend hours[br]showing you each and every decade,

0:04:37.713,0:04:39.574
but we don't have the time right now.

0:04:39.598,0:04:42.857
So you can go on your phone[br]and actually do it yourself.

0:04:42.881,0:04:46.343
(Applause)

0:04:46.367,0:04:49.469
But if you don't mind[br]and can hold your applause till later,

0:04:49.493,0:04:51.041
I don't want to run out of time,

0:04:51.065,0:04:53.264
because I want to show you[br]a lot of cool stuff.

0:04:53.288,0:04:54.788
So, just very quickly:

0:04:54.812,0:04:58.798
you can move on from here[br]to another very interesting idea.

0:04:58.822,0:05:00.533
Beyond the pretty picture,

0:05:00.557,0:05:02.346
beyond the nice visualization,

0:05:02.370,0:05:04.697
what is the purpose, how is this useful?

0:05:04.721,0:05:09.450
This next idea comes[br]from discussions with curators

0:05:09.474,0:05:11.109
that we've been having at museums,

0:05:11.133,0:05:13.156
who, by the way, I've fallen in love with,

0:05:13.180,0:05:16.374
because they dedicate their whole life[br]to try to tell these stories.

0:05:16.398,0:05:19.329
One of the curators told me,[br]"Amit, what would it be like

0:05:19.353,0:05:21.832
if you could create[br]a virtual curator's table

0:05:21.856,0:05:24.159
where all these six million objects

0:05:24.183,0:05:28.586
are displayed in a way for us[br]to look at the connections between them?"

0:05:28.610,0:05:32.639
You can spend a lot of time, trust me,[br]looking at different objects

0:05:32.663,0:05:34.616
and understanding where they come from.

0:05:34.640,0:05:36.848
It's a crazy Matrix experience.

0:05:36.872,0:05:37.886
(Laughter)

0:05:37.910,0:05:39.196
Just moving on,

0:05:39.220,0:05:42.391
let's take the world-famous[br]Vincent Van Gogh,

0:05:42.415,0:05:46.130
who is very well-represented[br]on this platform.

0:05:46.154,0:05:48.930
Thanks to the diversity[br]of the institutions we have,

0:05:48.954,0:05:54.658
we have over 211 high-definition,[br]amazing artworks by this artist,

0:05:54.682,0:05:58.429
now organized in one beautiful view.

0:05:58.453,0:06:01.426
And as it resolves,[br]and as Cyril goes deeper,

0:06:01.450,0:06:03.659
you can see all the self-portraits,

0:06:03.683,0:06:05.108
you can see still life.

0:06:05.132,0:06:07.735
But I just wanted to highlight[br]one very quick example,

0:06:07.759,0:06:09.447
which is very timely:

0:06:09.471,0:06:10.782
"The Bedroom."

0:06:10.806,0:06:13.107
This is an artwork[br]where three copies exist --

0:06:13.131,0:06:15.461
one at the Van Gogh Museum in Amsterdam,

0:06:15.485,0:06:17.472
one at the Orsay in Paris

0:06:17.496,0:06:19.430
and one at the Art Institute of Chicago,

0:06:19.454,0:06:21.957
which, actually, currently[br]is hosting a reunion

0:06:21.981,0:06:23.756
of all three artworks physically,

0:06:23.780,0:06:25.917
I think only for the second time ever.

0:06:25.941,0:06:30.860
But, it is united digitally and virtually[br]for anybody to look at

0:06:30.884,0:06:32.805
in a very different way,

0:06:32.829,0:06:35.862
and you won't get pushed[br]in the line in the crowd.

0:06:35.886,0:06:40.628
So let's take you and let's travel[br]through "The Bedroom" very quickly,

0:06:40.652,0:06:44.239
so you can experience what we are doing[br]for every single object.

0:06:44.263,0:06:47.714
We want the image to speak[br]as much as it can

0:06:47.738,0:06:49.382
on a digital platform.

0:06:49.406,0:06:52.553
And all you need is an internet[br]connection and a computer

0:06:52.577,0:06:56.915
(Applause)

0:06:56.939,0:06:59.497
And, Cyril, if you can go deeper, quickly.

0:06:59.521,0:07:01.005
I'm sorry, this is all live,

0:07:01.029,0:07:03.464
so you have to give Cyril[br]a little bit of --

0:07:04.401,0:07:06.461
and this is available for every object:

0:07:06.485,0:07:09.200
modern art, contemporary art,[br]Renaissance -- you name it,

0:07:09.224,0:07:10.411
even sculpture.

0:07:11.802,0:07:15.397
Sometimes, you don't know[br]what can attract you

0:07:15.421,0:07:20.649
to an artwork or to a museum[br]or to a cultural discovery.

0:07:20.673,0:07:23.415
So for me, personally,[br]it was quite a challenge

0:07:23.439,0:07:26.661
because when I decided to make this[br]my full-time job at Google,

0:07:26.685,0:07:29.288
my mother was not very supportive.

0:07:29.312,0:07:30.559
I love my mother,

0:07:30.583,0:07:33.582
but she thought I was wasting my life[br]with this museum stuff.

0:07:33.606,0:07:36.918
And for her, a museum is what[br]you do when you go on vacation

0:07:36.942,0:07:38.909
and you tick-mark and it's over, right?

0:07:38.933,0:07:41.464
And it took around four and a half years

0:07:41.488,0:07:43.702
for me to convince my lovely Indian mother

0:07:43.726,0:07:45.379
that actually, this is worthwhile.

0:07:45.403,0:07:48.844
And the way I did it was,[br]I realized one day that she loves gold.

0:07:48.868,0:07:54.443
So I started showing her all objects[br]that have the material gold in them.

0:07:54.467,0:07:57.118
And the first thing my mom asks me is,

0:07:57.142,0:07:59.570
"How can we buy these?"

0:07:59.594,0:08:01.412
(Laughter)

0:08:01.436,0:08:04.578
And obviously, my salary is not that high,

0:08:04.602,0:08:06.980
so I was like, "We can't[br]actually do that, mom.

0:08:07.004,0:08:09.115
But you can explore them virtually."

0:08:09.139,0:08:12.558
And so now my mom -- every time[br]I meet her, she asks me,

0:08:12.582,0:08:16.048
"Any more gold, any more silver[br]in your project? Can you show me?"

0:08:16.072,0:08:18.246
And that's the idea[br]I'm trying to illustrate.

0:08:18.270,0:08:20.408
It does not matter how you get in,

0:08:20.432,0:08:21.794
as long as you get in.

0:08:21.818,0:08:23.428
Once you get in, you're hooked.

0:08:23.452,0:08:25.801
Moving on from here very quickly,

0:08:25.825,0:08:28.674
there is kind of a playful idea, actually,

0:08:28.698,0:08:30.333
to illustrate the point of access,

0:08:30.357,0:08:32.779
and I'm going to go[br]quite quickly on this one.

0:08:32.803,0:08:38.210
We all know that seeing the artwork[br]in person is amazing.

0:08:38.234,0:08:40.817
But we also know[br]that most of us can't do it,

0:08:40.841,0:08:43.599
and the ones that can afford[br]to do it, it's complicated.

0:08:43.623,0:08:47.430
So -- Cyril, can we load[br]up our art trip, what do we call it?

0:08:47.454,0:08:49.216
We don't have a good name for this.

0:08:49.240,0:08:54.211
But essentially, we have[br]around 1,000 amazing institutions,

0:08:54.235,0:08:55.402
68 countries.

0:08:55.426,0:08:57.126
But let's start with Rembrandt.

0:08:57.150,0:08:59.392
We might have time for only one example.

0:08:59.416,0:09:01.645
But thanks to the diversity,

0:09:01.669,0:09:05.573
we've got around 500 amazing[br]Rembrandt object artworks

0:09:05.597,0:09:07.866
from 46 institutions and 17 countries.

0:09:07.890,0:09:10.442
Let's say that on your next vacation,

0:09:10.466,0:09:12.516
you want to go see[br]every single one of them.

0:09:12.540,0:09:13.874
That is your itinerary,

0:09:13.898,0:09:17.552
you will probably travel[br]53,000 kilometers,

0:09:17.576,0:09:20.163
visit around, I think, 46 institutions,

0:09:20.187,0:09:24.378
and just FYI, you might release[br]10 tons of CO2 emissions.

0:09:24.402,0:09:25.865
(Laughter)

0:09:25.889,0:09:27.362
But remember, it's art,

0:09:27.386,0:09:30.051
so you can justify it,[br]perhaps, in some way.

0:09:30.837,0:09:32.759
Moving on swiftly from here,

0:09:32.783,0:09:35.718
is something a little bit[br]more technical and more interesting.

0:09:35.742,0:09:40.679
All that we've shown you so far[br]uses metadata to make the connections.

0:09:40.703,0:09:42.813
But obviously we have[br]something cool nowadays

0:09:42.837,0:09:45.749
that everyone likes to talk about,[br]which is machine learning.

0:09:45.773,0:09:48.847
So what we thought is,[br]let's strip out all the metadata,

0:09:48.871,0:09:52.483
let's look at what machine learning can do

0:09:52.507,0:09:56.830
based purely on visual recognition[br]of this entire collection.

0:09:56.854,0:10:01.047
What we ended up with[br]is this very interesting map,

0:10:01.071,0:10:05.381
these clusters that have[br]no reference point information,

0:10:05.405,0:10:09.370
but has just used visuals[br]to cluster things together.

0:10:09.394,0:10:12.976
Each cluster is an art to us[br]by itself of discovery.

0:10:13.000,0:10:15.663
But one of the clusters we want[br]to show you very quickly

0:10:15.687,0:10:18.339
is this amazing cluster of portraits

0:10:18.363,0:10:20.765
that we found from museums[br]around the world.

0:10:20.789,0:10:22.972
If you could zoom in[br]a little bit more, Cyril.

0:10:22.996,0:10:26.828
Just to show you, you can[br]just travel through portraits.

0:10:26.852,0:10:29.626
And essentially, you can do nature,[br]you can do horses

0:10:29.650,0:10:31.406
and clusters galore.

0:10:32.309,0:10:33.857
When we saw all these portraits,

0:10:33.881,0:10:37.533
we were like, "Hey, can we do[br]something fun for kids,

0:10:37.557,0:10:39.239
or can we do something playful

0:10:39.263,0:10:41.288
to get people interested in portraits?"

0:10:41.312,0:10:43.492
Because I haven't really seen

0:10:43.516,0:10:46.883
young kids really excited[br]to go to a portrait gallery.

0:10:47.357,0:10:49.289
I wanted to try to figure something out.

0:10:49.313,0:10:51.780
So we created something[br]called the portrait matcher.

0:10:51.804,0:10:53.121
It's quite self-explanatory,

0:10:53.145,0:10:55.883
so I'm just going to let Cyril[br]show his beautiful face.

0:10:55.907,0:11:00.039
And essentially what's happening is,[br]with the movement of his head,

0:11:00.063,0:11:04.230
we are matching different portraits[br]around the world from museums.

0:11:04.254,0:11:06.450
(Applause)

0:11:17.445,0:11:19.151
And I don't know about you,

0:11:19.175,0:11:21.191
but I've shown it to my nephew and sister,

0:11:21.215,0:11:23.064
and the reaction is just phenomenal.

0:11:23.088,0:11:25.794
All they ask me is,[br]"When can we go see this?"

0:11:27.492,0:11:29.065
And by the way, if we're nice,

0:11:29.089,0:11:31.910
maybe, Cyril, you can smile[br]and find a happy one?

0:11:32.696,0:11:34.094
Oh, perfect.

0:11:34.118,0:11:36.106
By the way, this is not rehearsed.

0:11:36.474,0:11:39.054
Congrats, Cyril. Great stuff. Oh wow.

0:11:39.078,0:11:42.554
OK, let's move on; otherwise,[br]this will just take the whole time.

0:11:42.578,0:11:44.296
(Applause)

0:11:44.320,0:11:46.900
So, art and culture[br]can be fun also, right?

0:11:47.900,0:11:51.297
For our last quick experiment --

0:11:51.321,0:11:53.084
we call all of these "experiments" --

0:11:53.108,0:11:55.997
our last quick experiment[br]comes back to machine learning.

0:11:56.021,0:11:58.862
We show you clusters, visual clusters,

0:11:58.886,0:12:03.787
but what if we could ask the machine[br]to also name these clusters?

0:12:03.811,0:12:08.564
What if it could automatically tag[br]them, using no actual metadata?

0:12:08.588,0:12:12.688
So what we have is this kind of explorer,

0:12:12.712,0:12:16.323
where we have managed to match,[br]I think, around 4,000 labels.

0:12:16.347,0:12:19.511
And we haven't really[br]done anything special here,

0:12:19.535,0:12:21.193
just fed the collection.

0:12:21.217,0:12:22.987
And we found interesting categories.

0:12:23.011,0:12:25.916
We can start with horses,[br]a very straightforward category.

0:12:25.940,0:12:28.239
You would expect to see[br]that the machine has put

0:12:28.263,0:12:30.004
images of horses, right?

0:12:30.028,0:12:33.470
And it has, but you also notice,[br]right over there,

0:12:33.494,0:12:35.573
that it has a very abstract image

0:12:35.597,0:12:38.792
that it has still managed to recognize[br]and cluster as horses.

0:12:38.816,0:12:42.506
We also have an amazing head[br]in terms of a horse.

0:12:42.530,0:12:46.532
And each one has the tags[br]as to why it got categorized in this.

0:12:46.556,0:12:50.530
So let's move to another one[br]which I found very funny and interesting,

0:12:50.554,0:12:53.045
because I don't understand[br]how this category came up.

0:12:53.069,0:12:54.819
It's called "Lady in Waiting."

0:12:54.843,0:12:57.455
If, Cyril, you do it very quickly,

0:12:57.479,0:13:01.021
you will see that we have[br]these amazing images

0:13:01.045,0:13:03.806
of ladies, I guess, in waiting or posing.

0:13:03.830,0:13:05.276
I don't really understand it.

0:13:05.300,0:13:07.540
But I've been trying to ask[br]my museum contacts,

0:13:07.564,0:13:09.809
you know, "What is this?[br]What's going on here?"

0:13:09.833,0:13:11.445
And it's fascinating.

0:13:11.469,0:13:14.158
Coming back to gold very quickly,

0:13:14.182,0:13:15.745
I wanted to search for gold

0:13:15.769,0:13:17.871
and see how the machine[br]tagged all the gold.

0:13:17.895,0:13:19.975
But, actually, it doesn't tag it as gold.

0:13:20.320,0:13:21.835
We are living in popular times.

0:13:21.859,0:13:23.962
It tags it as "bling-bling."

0:13:23.986,0:13:25.800
(Laughter)

0:13:27.200,0:13:30.121
I'm being hard on Cyril,[br]because I'm moving too fast.

0:13:30.145,0:13:32.772
Essentially, here you have[br]all the bling-bling

0:13:32.796,0:13:35.957
of the world's museums organized for you.

0:13:35.981,0:13:39.497
And finally, to end this talk[br]and these experiments,

0:13:39.521,0:13:42.838
what I hope you feel after this talk[br]is happiness and emotion.

0:13:42.862,0:13:45.301
And what would we see[br]when we see happiness?

0:13:45.325,0:13:48.484
If we actually look at all the objects

0:13:48.508,0:13:50.452
that have been tagged under "happiness,"

0:13:50.476,0:13:53.540
you would expect happiness, I guess.

0:13:53.564,0:13:59.667
But there was one that came up[br]that was very fascinating and interesting,

0:13:59.691,0:14:04.066
which was this artwork[br]by Douglas Coupland,

0:14:04.090,0:14:06.185
our friend and artist[br]in residence as well,

0:14:06.209,0:14:08.470
called, "I Miss My Pre-Internet Brain."

0:14:08.494,0:14:12.109
I don't know why the machine feels like[br]it misses its pre-Internet brain

0:14:12.133,0:14:13.403
and it's been tagged here,

0:14:13.427,0:14:15.341
but it's a very interesting thought.

0:14:15.365,0:14:17.548
I sometimes do miss my pre-Internet brain,

0:14:17.572,0:14:20.572
but not when it comes to exploring[br]arts and culture online.

0:14:20.596,0:14:22.909
So take out your phones,[br]take out your computers,

0:14:22.933,0:14:24.090
go visit museums.

0:14:24.114,0:14:26.793
And just a quick call-out[br]to all the amazing archivists,

0:14:26.817,0:14:28.201
historians, curators,

0:14:28.225,0:14:30.997
who are sitting in museums,[br]preserving all this culture.

0:14:31.021,0:14:34.339
And the least we can do is get[br]our daily dose of art and culture

0:14:34.363,0:14:35.694
for ourselves and our kids.

0:14:35.718,0:14:36.871
Thank you.

0:14:36.895,0:14:47.536
(Applause)