WEBVTT

00:00:17.816 --> 00:00:21.325
Our world is changing in many ways

00:00:21.325 --> 00:00:25.975
and one of the things which is going
to have a huge impact on our future

00:00:25.975 --> 00:00:29.363
is artificial intelligence - AI,

00:00:29.363 --> 00:00:32.953
bringing another industrial revolution.

00:00:33.627 --> 00:00:39.504
Previous industrial revolutions
expanded human's mechanical power.

00:00:40.014 --> 00:00:45.572
This new revolution,
this second machine age

00:00:45.572 --> 00:00:50.122
is going to expand
our cognitive abilities,

00:00:50.122 --> 00:00:52.102
our mental power.

00:00:52.782 --> 00:00:57.177
Computers are not just going
to replace manual labor,

00:00:57.597 --> 00:00:59.897
but also mental labor.

00:01:00.500 --> 00:01:03.450
So, where do we stand today?

00:01:04.034 --> 00:01:07.724
You may have heard
about what happened last March

00:01:07.724 --> 00:01:11.776
when a machine learning system
called AlphaGo

00:01:11.776 --> 00:01:17.708
used deep learning to beat
the world champion at the game of Go.

00:01:18.279 --> 00:01:20.679
Go is an ancient Chinese game

00:01:20.679 --> 00:01:24.159
which had been much more difficult
for computers to master

00:01:24.159 --> 00:01:25.982
than the game of chess.

00:01:26.893 --> 00:01:32.086
How did we succeed,
now, after decades of AI research?

00:01:33.068 --> 00:01:36.698
AlphaGo was trained to play Go.

00:01:37.678 --> 00:01:41.300
First, by watching over and over

00:01:41.814 --> 00:01:46.894
tens of millions of moves made
by very strong human players.

00:01:47.746 --> 00:01:52.496
Then, by playing against itself,
millions of games.

00:01:54.222 --> 00:01:59.941
Machine Learning allows computers
to learn from examples.

00:02:00.465 --> 00:02:02.575
To learn from data.

00:02:03.885 --> 00:02:07.235
Machine learning
has turned out to be a key

00:02:07.235 --> 00:02:11.635
to cram knowledge into computers.

00:02:12.174 --> 00:02:14.066
And this is important

00:02:14.066 --> 00:02:19.296
because knowledge
is what enables intelligence.

00:02:20.438 --> 00:02:26.768
Putting knowledge into computers had been
a challenge for previous approaches to AI.

00:02:27.515 --> 00:02:28.745
Why?

00:02:29.059 --> 00:02:33.859
There are many things
which we know intuitively.

00:02:34.601 --> 00:02:38.081
So we cannot communicate them verbally.

00:02:38.619 --> 00:02:42.780
We do not have conscious access
to that intuitive knowledge.

00:02:43.270 --> 00:02:46.690
How can we program computers
without knowledge?

00:02:47.664 --> 00:02:49.114
What's the solution?

00:02:49.314 --> 00:02:55.343
The solution is for machines to learn
that knowledge by themselves,

00:02:55.343 --> 00:02:56.443
just as we do.

00:02:56.443 --> 00:03:03.194
And this is important because knowledge
is what enables intelligence.

00:03:03.194 --> 00:03:06.974
My mission has been
to contribute to discover

00:03:06.974 --> 00:03:12.676
and understand principles
of intelligence through learning.

00:03:13.166 --> 00:03:18.116
Whether animal, human or machine learning.

00:03:19.450 --> 00:03:25.066
I and others believe that there are
a few key principles,

00:03:25.066 --> 00:03:27.296
just like the law of physics.

00:03:27.885 --> 00:03:32.745
Simple principles which could explain
our own intelligence

00:03:32.745 --> 00:03:36.741
and help us build intelligent machines.

00:03:37.885 --> 00:03:41.595
For example, think about the laws
of aerodynamics

00:03:41.595 --> 00:03:48.036
which are general enough to explain
the flight of both, birds and planes.

00:03:49.146 --> 00:03:55.381
Wouldn't it be amazing to discover
such simple but powerful principles

00:03:55.381 --> 00:03:59.186
that would explain intelligence itself?

00:04:00.026 --> 00:04:03.394
Well, we've made some progress.

00:04:04.384 --> 00:04:10.857
My collaborators and I have contributed
in recent years in a revolution in AI

00:04:11.777 --> 00:04:16.397
with our research on neural networks
and deep learning,

00:04:16.397 --> 00:04:20.946
an approach to machine learning
which is inspired by the brain.

00:04:22.041 --> 00:04:25.243
It started with speech recognition

00:04:25.243 --> 00:04:29.963
on your phones,
with neural networks since 2012.

00:04:30.977 --> 00:04:35.647
Shortly after, came a breakthrough
in computer vision.

00:04:36.680 --> 00:04:43.087
Computers can now do a pretty good job
of recognizing the content of images.

00:04:43.674 --> 00:04:50.049
In fact, they approach human performance
on some benchmarks over the last 5 years.

00:04:50.711 --> 00:04:54.721
A computer can now get
an intuitive understanding

00:04:54.721 --> 00:04:58.191
of the visual appearance of a Go-board

00:04:58.191 --> 00:05:01.763
that is comparable to that
of the best human players.

00:05:01.763 --> 00:05:03.454
More recently,

00:05:03.454 --> 00:05:06.584
following some discoveries made in my lab,

00:05:06.584 --> 00:05:11.408
deep learning has been used to translate
from one language to another

00:05:11.414 --> 00:05:14.441
and you are going to start seeing
this in Google translate.

00:05:15.191 --> 00:05:18.192
This is expanding the computer's ability

00:05:18.192 --> 00:05:22.532
to understand and generate
natural language.

00:05:23.550 --> 00:05:25.517
But don't be fooled.

00:05:25.517 --> 00:05:30.048
We are still very, very far from a machine

00:05:30.048 --> 00:05:34.033
that would be as able as humans

00:05:34.033 --> 00:05:37.593
to learn to master
many aspects of our world.

00:05:38.541 --> 00:05:41.237
So, let's take an example.

00:05:41.637 --> 00:05:46.787
Even a two year old child
is able to learn things

00:05:46.787 --> 00:05:50.657
in a way that computers
are not able to do right now.

00:05:51.767 --> 00:05:56.169
A two year old child actually
masters intuitive physics.

NOTE Paragraph

00:05:56.968 --> 00:06:01.908
She knows when she drops a ball
that it is going to fall down.

00:06:02.493 --> 00:06:06.093
When she spills some liquids
she expects the resulting mess.

00:06:06.586 --> 00:06:09.516
Her parents do not need to teach her

00:06:09.516 --> 00:06:12.980
about Newton's laws
or differential equations.

00:06:13.840 --> 00:06:20.200
She discovers all these things by herself
in a unsupervised way.

00:06:21.352 --> 00:06:27.712
Unsupervised learning actually remains
one of the key challenges for AI.

00:06:28.184 --> 00:06:33.014
And it may take several more decades
of fundamental research

00:06:33.014 --> 00:06:34.674
to crack that knot.

00:06:34.674 --> 00:06:40.895
Unsupervised learning is actually trying
to discover representations of the data.

00:06:41.729 --> 00:06:43.779
Let me show you and example.

00:06:44.364 --> 00:06:49.346
Consider a page on the screen
that you're seeing with your eyes

00:06:49.346 --> 00:06:54.196
or that the computer is seeing
as an image, a bunch of pixels.

00:06:54.993 --> 00:07:00.113
In order to answer a question
about the content of the image

00:07:00.863 --> 00:07:05.211
you need to understand
its high-level meaning.

00:07:05.674 --> 00:07:10.821
This high level meaning corresponds
to the highest level of representation

00:07:10.821 --> 00:07:12.321
in your brain.

00:07:12.906 --> 00:07:18.308
Low down, you have
the individual meaning of words

00:07:19.188 --> 00:07:23.798
and even lower down, you have characters
which make up the words.

00:07:24.810 --> 00:07:27.677
Those characters could be
rendered in different ways

00:07:27.677 --> 00:07:30.879
with different strokes
that make up the characters.

00:07:31.559 --> 00:07:34.839
And those strokes are made up of edges


00:07:34.839 --> 00:07:37.284
and those edges are made up of pixels.

00:07:37.284 --> 00:07:40.454
So these are different
levels of representation.

00:07:41.079 --> 00:07:44.236
But the pixels are not
sufficient by themselves

00:07:44.236 --> 00:07:46.584
to make sense of the image,

00:07:46.584 --> 00:07:51.904
to answer a high level question
about the content of the page.

00:07:52.932 --> 00:07:57.594
Your brain actually has
these different levels of representation

00:07:57.594 --> 00:08:02.291
starting with neurons
in the first visual area of cortex - V1,

00:08:02.291 --> 00:08:04.596
which recognizes edges.

00:08:04.596 --> 00:08:09.334
And then, neurons in the second
visual area of cortex - V2,

00:08:09.334 --> 00:08:12.800
which recognizes strokes and small shapes.

00:08:12.800 --> 00:08:17.060
Higher up, you have neurons
which detect parts of objects

00:08:17.060 --> 00:08:19.992
and then objects and full scenes.

00:08:21.182 --> 00:08:24.757
Neural networks,
when they're trained with images,

00:08:24.757 --> 00:08:28.860
can actually discover these types
of levels of representation

00:08:28.860 --> 00:08:32.778
that match pretty well
what we observe in the brain.

00:08:33.638 --> 00:08:38.798
Both, biological neural networks,
which are what you have in your brain

00:08:38.804 --> 00:08:42.828
and the deep neural networks
that we train on our machines

00:08:42.845 --> 00:08:48.075
can learn to transform from one level
of representation to the next,

00:08:48.369 --> 00:08:53.299
with the high levels corresponding
to more abstract notions.

00:08:53.299 --> 00:08:57.562
For example the abstract notion
of the character A

00:08:57.562 --> 00:09:00.891
can be rendered in many different ways
at the lowest levels

00:09:00.891 --> 00:09:03.887
as many different configurations of pixels

00:09:03.887 --> 00:09:09.097
depending on the position,
rotation, font and so on.

00:09:10.445 --> 00:09:15.815
So, how do we learn
these high levels of representations?

00:09:16.962 --> 00:09:20.681
One thing that has been
very successful up to now

00:09:20.681 --> 00:09:22.853
in the applications of deep learning,

00:09:22.855 --> 00:09:25.985
is what we call supervised learning.

00:09:26.297 --> 00:09:31.588
With supervised learning, the computer
needs to be taken by the hand

00:09:31.594 --> 00:09:35.467
and humans have to tell the computer
the answer to many questions.

00:09:35.467 --> 00:09:41.420
For example, on millions and millions
of images, humans have to tell the machine

00:09:41.420 --> 00:09:44.271
well... for this image, it is a cat.

00:09:44.273 --> 00:09:47.095
For this image, it is a dog.

00:09:47.095 --> 00:09:49.585
For this image, it is a laptop.

00:09:49.605 --> 00:09:55.595
For this image, it is a keyboard,
And so on, and so on millions of times.

00:09:56.066 --> 00:10:01.026
This is very painful and we use
crowdsourcing to manage to do that.

00:10:01.461 --> 00:10:03.396
Although, this is very powerful

00:10:03.416 --> 00:10:06.269
and we are able to solve
many interesting problems,

00:10:06.269 --> 00:10:08.313
humans are much stronger

00:10:08.313 --> 00:10:12.076
and they can learn over many more
different aspects of the world

00:10:12.076 --> 00:10:13.809
in a much more autonomous way,

00:10:13.809 --> 00:10:17.609
just as we've seen with the child
learning about intuitive physics.

00:10:17.625 --> 00:10:23.739
Unsupervised learning could also help us
deal with self-driving cars.

00:10:24.567 --> 00:10:26.097
Let me explain what I mean:

00:10:26.097 --> 00:10:31.835
Unsupervised learning allows computers
to project themselves into the future

00:10:31.835 --> 00:10:37.205
to generate plausible futures
conditioned on the current situation.

00:10:38.369 --> 00:10:42.899
And that allows computers to reason
and to plan ahead.

00:10:43.450 --> 00:10:47.985
Even for circumstances
they have not been trained on.

00:10:48.751 --> 00:10:50.441
This is important

00:10:50.441 --> 00:10:53.951
because if we use supervised learning
we would have to tell the computers

00:10:53.951 --> 00:10:57.395
about all the circumstances
where the car could be

00:10:57.395 --> 00:11:01.375
and how humans
would react in that situation.

00:11:02.451 --> 00:11:06.191
How did I learn to avoid
dangerous driving behavior?

00:11:07.276 --> 00:11:10.791
Did I have to die
a thousand times in an accident?

00:11:10.793 --> 00:11:12.106
(Laughter)

00:11:12.106 --> 00:11:14.606
Well, that's the way we train
machines right now.

00:11:15.175 --> 00:11:18.340
So, it's not going to fly
or at least not to drive.

00:11:18.340 --> 00:11:19.928
(Laughter)

00:11:21.288 --> 00:11:25.657
So, what we need is to train our models

00:11:25.657 --> 00:11:31.294
to be able to generate plausible images
or plausible futures, be creative.

00:11:31.294 --> 00:11:33.934
And we are making progress with that.

00:11:33.934 --> 00:11:37.457
So, we're training
these deep neural networks

00:11:37.463 --> 00:11:40.818
to go from high-level meaning to pixels

00:11:40.818 --> 00:11:43.298
rather than from pixels
to high level meaning,

00:11:43.307 --> 00:11:46.787
going into the other direction
through the levels of representation.

00:11:46.787 --> 00:11:50.461
And this way, the computer
can generate images

00:11:51.191 --> 00:11:55.072
that are new images different
from what the computer has seen

00:11:55.072 --> 00:11:56.488
while it was trained,

00:11:57.018 --> 00:12:00.369
but are plausible and look like natural images.

00:12:01.888 --> 00:12:06.332
We can also use these models
to dream up strange,

00:12:06.342 --> 00:12:09.492
sometimes scary images,

00:12:09.492 --> 00:12:11.795
just like our dreams and nightmares.

00:12:12.682 --> 00:12:16.847
Here's some images
that were synthesized by the computer

00:12:16.847 --> 00:12:19.826
using these deep charted models.

00:12:19.826 --> 00:12:21.651
They look like natural images

00:12:21.651 --> 00:12:24.551
but if you look closely,
you will see they are different

00:12:25.458 --> 00:12:28.697
and they're still missing
some of the important details

00:12:28.697 --> 00:12:31.063
that we would recognize as natural.

00:12:31.995 --> 00:12:33.951
About 10 years ago,

00:12:33.951 --> 00:12:38.921
unsupervised learning has been
a key to the breakthrough

00:12:38.921 --> 00:12:42.443
that we obtained
discovering deep learning.

00:12:44.140 --> 00:12:48.055
This was happening in just few labs,
including mine at the time

00:12:48.055 --> 00:12:51.455
at a time when neural networks
were not popular.

00:12:51.455 --> 00:12:55.217
They were almost abandoned
by the scientific community.

00:12:56.394 --> 00:12:58.935
Now, things have changed a lot.

00:12:58.935 --> 00:13:01.375
It has become a very hard field.

00:13:01.384 --> 00:13:06.933
There are now hundreds of students
every year applying for graduate studies

00:13:06.954 --> 00:13:09.784
at my lab with my collaborators.

00:13:11.010 --> 00:13:16.630
Montreal has become
the largest academic concentration

00:13:16.637 --> 00:13:19.387
of deep learning researchers in the world.

00:13:20.182 --> 00:13:26.115
We just received a huge
research grant of 94 million dollars

00:13:26.127 --> 00:13:29.797
to push the boundaries
of AI and data science

00:13:29.797 --> 00:13:36.067
and also to transfer technology of deep
learning and data science to the industry.

00:13:37.249 --> 00:13:43.791
Business people stimulated by all this
are creating start-ups, industrial labs,

00:13:43.791 --> 00:13:46.914
many of which near the universities.

00:13:48.543 --> 00:13:49.625
For example,

00:13:49.625 --> 00:13:54.733
just a few weeks ago, we announced
the launch of a start-up factory

00:13:54.733 --> 00:13:56.507
called 'Element AI'

00:13:56.507 --> 00:13:59.605
which is going to focus
on the deep learning applications.

00:14:01.562 --> 00:14:05.722
There are just not enough
deep learning experts.

00:14:06.355 --> 00:14:10.677
So, they are getting paid crazy salaries,

00:14:11.027 --> 00:14:17.212
and many of my former academic colleagues
have accepted generous deals

00:14:17.228 --> 00:14:20.518
from companies to work in industrial labs.

00:14:21.081 --> 00:14:25.010
I, for myself, have chosen
to stay in university,

00:14:25.010 --> 00:14:27.166
to work for the public good,

00:14:27.166 --> 00:14:28.886
to work with students,

00:14:28.902 --> 00:14:30.592
to remain independent.

00:14:30.596 --> 00:14:34.836
To guide the next generation
of deep learning experts.

00:14:35.294 --> 00:14:41.024
One thing that we are doing
beyond commercial value

00:14:41.024 --> 00:14:44.654
is thinking about the social
implications of AI.

00:14:45.881 --> 00:14:50.026
Many of us are now starting
to turn our eyes

00:14:50.026 --> 00:14:55.986
towards social value added
applications, like health.

00:14:56.457 --> 00:14:58.956
We think that we can use deep learning

00:14:58.956 --> 00:15:02.696
to improve treatment
with personalized medicine.

00:15:03.956 --> 00:15:05.671
I believe that in the future,

00:15:05.671 --> 00:15:10.361
as we collect more data from millions
and billions people around the earth,

00:15:10.361 --> 00:15:13.856
we will be able to provide medical advice

00:15:13.856 --> 00:15:17.246
to billions of people
who don't have access to it right now.

00:15:17.601 --> 00:15:22.724
And we can imagine many other
applications for social value of AI.

00:15:23.140 --> 00:15:26.238
For example, something
that will come out of our research

00:15:26.238 --> 00:15:28.582
on natural language understanding

00:15:29.328 --> 00:15:31.199
is providing all kinds of services

00:15:31.199 --> 00:15:34.059
like legal services,
to those who can't afford them.

00:15:34.512 --> 00:15:37.342
We are now turning our eyes

00:15:37.342 --> 00:15:41.132
also towards the social implications
of AI in my community.

00:15:41.690 --> 00:15:44.801
But it's not just for experts
to think about this.

00:15:46.026 --> 00:15:49.936
I believe that beyond the math
and the jargon,

00:15:49.936 --> 00:15:53.102
ordinary people can get the sense

00:15:53.138 --> 00:15:55.901
of what goes on under the hood

00:15:55.901 --> 00:16:01.191
enough to participate
in the important decisions

00:16:01.191 --> 00:16:06.547
that will take place, in the next
few years and decades about AI.

00:16:07.580 --> 00:16:09.280
So please,

00:16:09.930 --> 00:16:16.230
set aside your fees and give yourself
some space to learn about it.

00:16:17.842 --> 00:16:22.532
My collaborators and I have written
several introductory papers

00:16:22.542 --> 00:16:25.376
and a book entitled "Deep Learning"

00:16:25.376 --> 00:16:29.619
to help students and engineers
jump into this exciting field.

00:16:30.659 --> 00:16:35.784
There are also many online resources:
softwares, tutorials, videos..

00:16:36.310 --> 00:16:41.210
and many undergraduate students
are learning a lot of this

00:16:41.210 --> 00:16:44.548
about research in deep learning
by themselves,

00:16:44.548 --> 00:16:47.835
to later join the ranks of labs like mine.

00:16:49.370 --> 00:16:55.170
Ai is going to have a profound
impact on our society.

00:16:56.652 --> 00:17:01.672
So, it's important to ask:
How are we going to use it?

00:17:03.368 --> 00:17:07.896
Immense positives may come
along with negatives

00:17:07.896 --> 00:17:10.166
such as military use

00:17:10.797 --> 00:17:15.357
or rapid disruptive changes
in the job market.

00:17:15.948 --> 00:17:21.629
To make sure the collective choices
that will be made about AI

00:17:21.629 --> 00:17:23.074
in the next few years,

00:17:23.074 --> 00:17:25.144
will be for the benefit of all,

00:17:25.144 --> 00:17:28.557
every citizen should take an active role

00:17:28.557 --> 00:17:32.911
in defining how AI will shape our future.

00:17:33.871 --> 00:17:34.891
Thank you.

00:17:35.065 --> 00:17:39.395
(Applause)