ad-free
Channel / Source:
TEDx Talks
Published: 2017-05-17
Source: https://www.youtube.com/watch?v=uawLjkSI7Mo
the our world is changing in many ways and one of the things which is gonna have a huge impact on our future is artificial intelligence a I bringing another industrial revolution previous industrial revolutions spend that humans mechanical power this new revolution this second machine age is going to expand our cognitive abilities are mental power cuter he's are not just going to replace manual labor but also
mental labor so where do we stand today you may have heard about what happened last March when a machine learning system cold Alfa go used deep learning to beat the world champion at the game of go go is an ancient Chinese game which had been much more difficult for computers to master then the game of chess how did we succeed now after decades of AI research
Alfa go was trained play go first by watching over and over tens of millions of moves made by very strong human players then by playing against itself millions of games she learning allows computers to learn from examples to learn from data she learning has turned out to be a key crime knowledge into computers and this is important because knowledge is what enables intelligence putting knowledge into
computers had been a challenge for previous approaches to a I why there are many things which we know intuitively so we cannot communicate them verbally we do not have conscious access to that intuitive knowledge how can we programmed computers without knowledge what's the solution solution is for machines to learn died dot knowledge by themselves just as we do and this is important because knowledge is what
enables intelligence my mission has been to contribute to discover and understand principles of intelligence true learning whether animal or human or machine learning I and others believe that there are a few key principles just like the laws of physics simple principles which could explain our own intelligence and helpless build intelligent machines for example think about the laws of aerodynamics which are general enough to explain the
flight of both birds and planes wouldn't it be amazing to discover such simple but powerful principles that would explain intelligence itself well we've made some progress my collaborators and I have contributed in recent years in a revolution in a I with our research on neural networks and declining a new approach to machine learning which is inspired by the brain started with speech recognition on your phones
with neural networks since two thousand twelve shortly after came a breakthrough in computer vision purists can now do a pretty good job of recognizing the content of images in fact Prue Ching human performance on some benchmarks over the last five years a computer can now get it intuitive understanding of the visual appearance of a go board that is compatible to dot of the best human players
more recently following some discoveries made in my lap declining has been used to translate from one language to another and you know start seeing this in Google translate this is expanding the computer's ability to understand and generate natural language but don't be fooled we are still very very far from a machine doc would be as able as humans to learn to master many aspects of our
will so let's take an example even the eighty two year old child is able to learn things in a way that computers are not able to do right now two year old child actually masters included physics she knows when she dropped the ball that it is going to fall down which she spilled some liquids she expects the resulting mess her parents do not need to teach
her about Newton's laws or differential equations she discovers all these things by herself even unsupervised way unsupervised learning actually remains one of the key challenges four a I and it may take several more decades of fundamental research crack that not unsupervised learning is actually trying to discover representations of the data let me show you an example consider a page on the screen that you're seeing with
your eyes or that the computer is seen as an image a bunch of pixels in order to answer a question about the content of the image you need to understand its high level meeting and this high level meeting corresponds to the highest level of representation in your brain nor down you have the individual meaning of words and even lower down you have characters make up the
words and those characters could be rendered in different ways different strokes that make up the characters and those trucks are made up of edges and those edges are made up of pixels so these are different levels of representation but the pixels are not sufficient by themselves to make sense of the image to answer a highlight of question about the content of the page your brain actually
house these different levels of representation starting with neurons in the first visual area of cortex view one which recognize edges and then your own scan the second visual area of cortex view too which recognize strokes and small shapes higher up you have your own switch the tech parts of objects and then objects in full seems mule networks when they are trained with images can actually discovered
these types of levels of representation that match pretty well what we observe in the brain both biological you'll networks which are what you have in your brain and the deep you'll networks that we train on our machines can learn to transform from one level of representation to the next with the higher levels corresponding to more abstract notions for example the abstract notion of the capture it
eight can be rendered in many different ways at the lowest levels as many different configurations of pixels depending on your the position %HESITATION rotation font and so on so how do we learn these high levels of representations one thing that has been very successful up to now in the applications of the planning is what we called supervise lining with supervised learning the computer needs to be
taken by the hand and humans have to tell the computer the answer to many questions for example on millions and millions of images humans have to tell the machine well for this image it is a cat for this image it is a dog for this image it is a laptop for this image is a keyboard and so on and so on millions of times this is
very painful and we use crowdsourcing to %HESITATION managed to do that although this is very powerful and we already able to soul very interesting problems humans are much stronger and they can learn over many more different aspects of the world in a much more Thomas way just as we've seen with the two year old child learning about into the physics unsupervised learning could also help close
deal with self driving cars meet me explain what I mean unsupervised learning allows computers to project themselves into the future generate plausible futures condition on the current situation and that allows computers to reason and to plan ahead even for circumstances dot they have not been trained on this is important because if we you supervise learning we would have to tell the computers about all the circumstances
where the card could be and that you know how he would react in that situation how did I learn to avoid dangerous driving behavior did I have to die a thousand times in an accident with us that we were trying machines right now so it's not gonna fly or at least not to drive so what we need is to train our models to be able to
generate plausible images plausible futures be creative and we're making progress without so we're training these deep neural networks to go from high level meeting two pixels rather than from pixels to heighten the meetings are going in the other direction through the levels of representation in this way the computer can generate images there are new images different from what the computer has seen while he was trained
but are plausible that look like natural images you can also use these models to dream up strange sometimes scary images just like our dreams and nightmares here are some images that were synthesized by the computer using these deep chart of models they look like natural images but if you look closely you'll see there are different and they're still missing some of the important details that that
we would recognize as natural about ten years ago unsupervised learning has been key to the breakthrough that we obtained discovering D. planning this was happening in just a few laps including mine at a time when you'll networks were not popular they were almost abandoned by the scientific community now things have changed a lot it has become a very hot field there are now hundreds of students
every year applying for grad studies I've my lab with my collaborators Montreal has become the largest academic concentration of declining researchers in the world we just received a huge research grant of ninety four million dollars push the boundaries of a I and data science and also to transfer technology of declining and data science to industry business people stimulated by all this are creating startups industrial labs
many of which nearly universities for example just a few weeks ago we announced the lounge of a started factory called elementary I which is going to focus on the planning application there's just not enough declining experts so they're getting paid crazy salaries and many of my former academic colleagues have accepted generous deals from companies to work in industrial laps I for myself have chosen to stay
in university to work for the public good to work with students to remain independent I got it the next generation of declining expert one thing got we are doing beyond commercial value is thinking about the social implications of a I many of us are now starting to turn our eyes towards social value added applications like health I think that we can use declining to improve treatment
with personalized medicine I believe that in the future as we collect more data from millions and billions of people around the earth we'll be able to provide medical advice to billions of people who don't have access to it right now and we can imagine many other applications for social value of yeah I for example something that will come out of our research on natural language understanding
is providing all kinds of services like legal services to those who can't afford them we are now turning our eyes also towards the social implications of a I in my community but it's not just for experts to think about this I believe got beyond the math and the jargon ordinary people can get a sense of what goes on under the hood I know of to participate
in the importance decisions that will take place in the next few years and decades about a I so please set aside your fees and give yourself some space to learn about it my collaborators and I have written several introductory papers and a book untitled deep learning to help students and engineers jump into this exciting field there are also many online resources software tutorials videos and many
undergraduate students are learning a lot of this about researching declining by themselves to later joined the ranks of labs like mine a I is going to have a profound impact on our society so it's important to ask how are we going to use it immense positives may come along with negatives such as military use or wrap it disruptive changes in the job market to make sure
