Li Feifei: How to teach computer to understand pictures _ computer

2016 early reading paper stage, I first contact artificial Intelligence, at that time just feel this vocabulary, really hard to fight, sounded very high, and cool;2017 years, artificial intelligence has become the media of the year new words, even led a vote AI concept stocks rose up, For example, the current time (2017/12/29-15:00:00 Beijing time) PE value has been flying at Hkust 365.29. May 2017 holiday, and listened to Andrew Ng's half section "Machine Learning", halfway. Missed two times to understand the new technology, or even the opportunity of the technological revolution, the first day of 2018, the Li Feifei of this computer vision Ted Brush Well, just a long time did not practice English dictation.

Let me show you something ...
Some pictures are shown, and a girl is describing what-does the picture have ...
The boy is ...
Those are the ...
That ' s a big airplane ...

This is a three the describing what she sees in a series of photos. She may still have a lot to learn about this world, but she ' d already a expert at one very important task-make sense W Hat she sees. Our society are more technologically advanced than ever. We sent people to the moon and we make phones which talk to us or customise radio stations so can play only music we like. Yet, our most advanced machines and computers still struggle in this task. So I ' m here today, to give for you a progress, "latest advances in-out" in computer vision, one of the most FR Ontier and potentially revolutionary technology in computer. Yes, we have prototyped of cars that can drive by themselves, but without smart vision, they cannot-really tell the differ ence between a crumpled paper bag no the road, which can be run over, and a rock that size, which should to be avoided. We have made fabulous megapixel cameras, but we have not delivered to the sight. Drones can fly OVer massive land, but don ' t have enough vision to help us technology the track of the changes. Security cameras are everywhere, but they don't alert us when their the child are drawning in a swimming pool. Photos and videos are becoming a integral part of the global life. They ' re being generated at a pace that ' s far beyond the any what, the or team's human, human could to view. And you, and I are contributing to, this is TED. Yet our most advanced software are still struggling at understanding or managing content. So in the other words, collectively as a society, we ' re very much blind, because our smartest machine are, still blind.

"Why is this so hard?" Camera can take pictures like this one, by converting lights into a two-dimensional array of numbers known as Pixels, but These are just lifeless numbers. They don't carry meaning in themselves. Just as to hear isn't the same as to listen, to take pictures are not the same as to, and by seeing we really by Understanding.

In fact, the IT took Mother nature 540 millions year of hard work to does this task, and more that effort went into Developin g The visual processing apparatus of brains, not the eyes themselves. So the vision begins and the eye, but it truly take place in the brain.

So for fifteen years now, starting to my Ph.D. at Caltech and then leading Standford ' s Vision Lab, I ' ve been working With my mentors, collaborators and students to teach computers to. Our filed is called Computer Vision and machine learning. It ' s part of the general filed of Aritificail Intelligence.

So ultimately, we want to teach the machines to = just what we do:naming objects, identify people, inferring 3D Geom Etry of things, understanding relations, emotions, actions and intensions. You and I weave together entire stories of people, places and things the moment we lay in gaze. In a simplest terms, imagine this teching process as showing the computer some training to images of a particular object, let ' s say cats, and designing a model that learns from these training images. How to hard can this be? Tomorrow answer'll been shown haha