Baidu’s chief scientist Wu En: self-driving cars really mature must for 20 years

if you pay attention to the wind dynamic trend of investment and research fund, you will find that artificial intelligence will greatly change our life in the future. Few innovators in this new field of science and technology have significantly, while Chinese search engine giant baidu’s chief scientist Wu En (Andrew Ng) is one of them. Wu En at Stanford professor of artificial intelligence course, dominated by Google Brain “project, launched Coursera a pioneer in online education. Last year, he joined the baidu, will hope to take this platform to make computers learn to see, listen, will do the thing that people said to be in the service of the most densely populated country in the world.

why in China, the machine’s intelligent applications can have so great opportunity of development, has become a small mystery. The global flow of baidu ranked no. 5, taobao shopping website, QQ instant messaging applications, media company, sina, and sina weibo microblogging platform, these are the top 15 sites, are also belong to the Chinese website. According to Wu En da, when baidu set out to design a new application, first to consider is the mobile platform. In China, the mobile phone is one of the important channels connected consumers.

when it comes to his work, always gentle polite Wu En da is still a surprise ors inner passion. Today, he in the United States the sunnyvale, calif., baidu division manages a growing team. He does not believe that the current uproar Fried robot rebellion revolution, he thinks that at present researchers recognize the robot only potential of the tip of the iceberg. Wu En said, compared with the concern about the killer robot, he is more concerned with the running time of the microprocessor in the machine, or how to correct understanding of what is known speech recognition site. All this is he to do in the future. , despite a long way to go before Wu En always believed, we have so many good ideas and good company, in the near future can use voice control smartphones.

recently, Wu En on Skype for a short interview, expounds his view of machine intelligence. The following is a simple dialogue: after finishing

q: in the aspect of artificial intelligence and the potential of deep learning, which is the most make you excited?

Wu En: many companies, including our (baidu) and other companies, has very good computer vision technology, to complete the past — even a year ago — seemingly impossible task. In my opinion, the most challenging thing is how to design the eye-catching products. I think no one has yet devised such a enough to the application of stunning the world.

in silicon valley, where many start-ups are in agriculture or shopping areas (especially clothes shopping) using computer vision technology. Baidu, for example, if you see a picture of a movie star, we can use facial recognition technology to tell you who is the star, his age and be fond of and so on. If their clothing, which, within the scope of our recognition, we can give you recommend related clothes quickly. Everybody likes such technology.

q: these photos will be advertisers compete for the land?

Wu En: we have no current towards commercialization; We are just for the user to find similar clothing style. But it’s true that there are a lot of vertical can be applied to the technology, such as looking for interested users, by identifying a holiday destination and recommend other similar destination. Obviously, computer vision technology can do much more than this, but what can do, we are still in exploration.

q: artificial intelligence is increasingly devastating discussion, what makes people have such worry?

Wu En da: I think, perhaps hundreds of years people would create an unprecedented new technology, at that time, maybe computer will become a devastating threat. But the future is unpredictable. I don’t even know what will happen the next five years. I’m not worried about artificial intelligence will become a terrible thing, just like I don’t worry about overpopulation on Mars. Hundreds of years later I hope people can conquer Mars, but until now the human is not on the planet, so why should we worry about overpopulation in this matter.

q: artificial intelligence is a what kind of job?

Wu En: probably similar to build a rocket of nature. You need a great engine and a lot of fuel. If you have enough engine without enough fuel, you cannot turn the rocket into orbit. On the other hand, if the fuel is sufficient but engine is too weak, you even the emission problem. So, both be short of one cannot.

depth of learning, creating one of the key process of artificial intelligence). Deep learning model is the rocket engine, the huge amounts of data is what we are used to fuel processing algorithms.

q: do you have any work experience in Google, so what do you think of self-driving cars?

Wu En: next door to my office location is the team self-driving cars, many of them are my good friend, so I know their job. But I have never directly involved.

I think self-driving cars and most people understand there are still a big gap. There are two kinds of development point of view, people have been wrangling over. The first kind of development view, the development of self-driving cars have an increasing path, that is to say you have cruise control first, then the adaptive cruise control, and automatic driving a car can realize automatic driving on the highway, and then you continue to increase over the next 20 years new technology until you make full sense of the automatic driving a car. The second development point of view tend to organizations, perhaps the Carnegie Mellon or Google, unexpectedly overnight self-driving cars appeared in front of people! Yesterday has not self-driving cars, start selling today!

me, I support the first view. I think most people are not really understanding of self-driving cars. Able to drive one thousand miles and want to go to which go to which are two very different concepts. In fact, the most fundamental is still a machine learning techniques, it is a good way to promote autonomous accurate rate, such as increased from 90% to 99%. But, 99.99% accurate rate is a huge breakthrough, also is difficult. I can tell you is: be in self-driving cars than drunk driving safety a lot.

q: you started Coursera, and advocates online education programs. So what do you think about the future of education?

Wu En: our education system so far in the education on the next generation solutions to everyday problems has been a huge success. When the tractor instead of traditional agricultural labor, we education the next generation to work in factories. But in widely guide extraordinary creative work, our education still have limitations.

q: do you agree with the labor in the future will be replaced by automation such a statement? Because automation can reduce costs, so will the people working only need 10-20 hours a week is enough!

Wu En: want me to say, should be zero working time! I think the lowest life guarantee can serve as a long-term solution, but I’m not sure whether I will accept this solution. I think if all mankind can be liberated from the labor to do some important things, our society will be growing by leaps and bounds. Give people a range of skills to do all sorts of important things, it must be very exciting!

to learn more knowledge, entrepreneurial innovation quickly add hunting cloud network micro letter public account: