Home > News content

Facebook CTO dwell decade vision: network connection, AI and VR

via:博客园     time:2016/8/17 10:30:50     readed:1062

FacebookFacebook CTO 详谈十年愿景:网络连接、AI 和 VR

August 16 news, "Forbes" website published the article to restore business and Metis Strategy, president of IT consulting firm Peter & middot; Hay (Peter High) on Facebook CTO Mike & middot; Nova Luopu Fu (Mike Schroepfer) of interview . In the interview, Paul Luo Pufu explain the three pillars of innovation before the company announced a ten-year development plan: network connections, artificial intelligence and virtual reality.

The following are the main contents of the article:

Earlier this year Facebook F8 conference, the company announced the three pillars of its innovative vision for the next decade: Fi, artificial intelligence and virtual reality. Its chief technology officer Paul Luo Pufu responsible for leading three projects. Although this vision is the decade, but the company actually has three areas have made no small progress.

Facebook advances in artificial intelligence everywhere, dynamic information about the company (news feed) from people to mark what way. Innovative virtual reality on through the Oculus Rift has been the best show. Recently, the company on a network connection has taken an important step: Acquila aircraft set sail in Arizona, which will exceed commercial aircraft flight altitude and the weather layer. Facebook's goal is that by the wingspan longer than a Boeing 737 but weighs less than 1,000 pounds V aircraft to bring basic networking services to developing countries.

"Forbes" correspondent recently in Facebook's headquarters in Menlo Park is located in Nova Luopu Fu met to discuss the three pillars of the innovation of social networks and many other topics, including the company's personnel recruitment, how to maintain it innovation advantage, as well as its headquarters (the world's largest open-space office one) behind the design logic.

The following is a summary of the interview:

Q: Facebook F8 Developers Conference held earlier this year on, you have raised the three pillars of innovation. You can tell us about them?

Nova Luopu Fu:We can say that is a very special presence in the industry, we are very happy to share our decades of vision and roadmap, it will be broken down into three core areas:

& Middot; network, connecting nearly 4 billion is still no network available on the person (account for a large part of the global population);

& Middot; artificial intelligence, solve some of the core issues, to build a real intelligent computer systems;

& Middot; virtual reality and augmented reality, to create a promising next-generation computing systems, so that people thousands of miles apart can feel like staying the same room.

Q: Facebook There is absolutely no shortage of resources and talents, how did you pick out the three pillars, and not pick out something else?

Nova Luopu Fu:That many of them directly from Facebook CEO Mark & ​​middot; Zuckerberg (Mark Zuckerberg), derived from the more open the world becomes more interconnected mission. I just see this as the use of technology to connect people. We sat down to discuss the road together: & ldquo; if that is our goal, we are best suited to do that in this world there are specific what the big problem? & Rdquo; so as you start analysis, the answer will come out naturally.

The first problem is that if a large part of the global population even basic networking services are not, then it belongs to the fundamental problem. Then, through discussion and analysis, you will realize that there is a technical problem solutions; there is something that can significantly reduce the cost of deploying network infrastructure, which is a major constraint. In fact, it's just an economic issue.

Once people are connected to the network, it will met you I have encountered the problem, and that is information overload. Nowadays too much information, but my time is very limited, so I could not get the most valuable that information. The only way, we realized that an effective solution to this problem is to begin to build intelligent systems based on artificial intelligence, artificial intelligence so that I can become all-weather real-time assistant, make sure I do not miss any for me very important information to ensure I will not spend time on unimportant things. We know that to achieve this wide range of artificial intelligence is the only way.

Well, the value of networking and access to information problems can be solved. However, many of us have not always our friends and family around, we can not always talk to them to witness important moments in their lives. Now we can spend the advanced technology is video call. If I want to capture my children's life time, and always remember the dots, the video call is currently the best tool. The problem is that if I want to appear on the scene to record those times so that they will still have the feeling of proximity to the scene 20 years later to relive it? This is virtual reality can come in handy. It allows you to wear a helmet can experience this after all, makes you feel really in a real place thousands of miles away, so you can go anywhere you want to go.

Q: How do you see the long-term goals that require some time to achieve, as well as to commercialize the project and do a small number of short-term commercial preparation of?

Nova Luopu Fu:For the ten-year development plan, I would say the most important thing is, when things were not going well to remain patient, because the process to make any major event will not be easy. I had also founded the company in the past, of which the process is never smooth sailing. From the outside, I work in Facebook might have been progressing well, but in fact all the way down will be faced with a variety of ups and downs. Importantly, in the event flow, and do not feel discouraged.

Also, if you can be reached on the way forward can help you know yourself milestone of progress, even though you do it. Artificial intelligence is very simple, because the team has deployed a large number of things to dramatically improve people's daily Facebook experience. They may like to help you better organize your photos, so you will not lose important life pictures such simple technology, it can be a more fundamental technology, as we launched earlier this year of assistive technology, it may have visual impairments, can not see the day to upload photos Facebook billions of people to generate and read caption.

At the same time, our study is much more difficult problem. For example, how to teach computers to digest, such as unstructured data content of Wikipedia articles, and then answer questions related data? This may sound very simple, but it is at the forefront of the field of artificial intelligence research: in the absence of significant give you good value format, the inference is not digested and understood something and not optimized. From this perspective, the problem is very far-reaching.

In the field of virtual reality, we want to bring to the market an excellent product content to entice you to purchase Best Buy Oculus Rift helmet. We know that in 2016 this is still a relatively small niche market. Facebook compared to the current user base of more than 1 billion, and its audience is still limited, but we hope for virtual reality content will be good year after year. We will launch system upgrade every few years, as these systems become more good, cheaper, virtual reality after a decade is expected to reach hundreds of millions of users, not just millions. You have to process accumulated audience to be patient, not to the small amount of short-term users disappointed.

Network connectivity we have made some milestones, such as a few weeks ago sailing Acquila aircraft. But that is only the first flight of our aircraft. It is well away from the full implementation of the ultimate goal of preparation is still far away, the ultimate goal is to lay the network infrastructure costs too expensive areas provide network connection services.

You have to try to have a long-term vision, patience when things may be a problem, while the implementation of some of the previous intermediate step on the road trip, from which progress and add value to see the progress. At least that is what I look at long-term and short-term goals.

Q: You aim to solve the problem of not only mental needs, but also has extensive industry background talent. You need to have a STEM (science, technology, engineering and mathematics) education personnel, the need for advanced neuroscience expertise of personnel, and so on. Moreover, in Silicon Valley, talent competition is so fierce, how do you think people get the problem?

Nova Luopu Fu:Honestly, this is where the real fun work. Today, we in the headhunting do face challenges, but when you find a group of interdisciplinary and professional talent, amazing things will happen. Let's talk about it Acquila aircraft. You have to have high-end composite materials, electrical systems and motors. An important part is that the use of solar cells based on a new generation of battery technology. Next, we need a plane up and down in the transport network of the communications system, which means we have to do free-space optical communications, laser have communications equipment, laser transmitter and receiver. You need software to help control an unprecedented flight of aircraft. We need to look like what it would be to do some simulation. You have a machine learning software, hardware, electronics, aeronautics and materials science, you need to place all things merge into one.

When you put people in these areas to form a small team & mdash; & mdash; the dozens of people gathered in one & mdash; & mdash; so that when they explicitly targeting this goal, a lot of wonderful things happen. Our company's core strategy, our technology is that relating to the people gathered in one piece.

Q: employees choose to explore the field on how much freedom?

Nova Luopu Fu:They have a lot of freedom. I came to Facebook from the beginning, the basic principle is to get in your field belongs to the top level of smart people, and give them the freedom to choose where to work. If you let them do what they were most excited about the things they can play the biggest role. If they wake up in the morning because of a wait to solve the problem at hand and immediately put to work, they will bring much better work performance.

Of course, sometimes we have to try to convince some employees try to do something else, or to explain what is most important to them, but I can basically get the right people in the right position appears, after that I just they need to be clearly articulated our ultimate goal: & ldquo; this is what I want you to do things. You can bring a lot of contributions. Think about what is most of them makes you feel excited, and then concentrated to do on the line. & Rdquo; then we can start to do some things out together, along with the development of technology. So, we work on the content selection is a lot of freedom.

Q: In addition to the formation of the team within the company, also outside the company to build the ecosystem, how do you see?

Paul Luo Pufu: This is the one we tried a lot of innovation in the field. Looking back fifty-six historical open computing projects (Open Compute Project) creation, I feel we all have to accept open source is a good way to develop software. I worked at Mozilla, by the time we developed an open-source browser. Almost every company has to run Linux & mdash; & mdash; that there are open; we are on the same Linux kernel to make a contribution, but we do not for hardware development.

We discuss said, & ldquo; we do together to develop the right hardware. & Rdquo; thus, our data center to design buildings, racks and server & mdash; & mdash; where all things are open, people can collaborate. Today, the entire industry during the collaboration, including Google. We also adopt the same strategy in the field of network connection, such as the laying of telecom infrastructure. If we can make the whole industry involved, that eventually everyone will benefit, because you can share a variety of core IP, you can develop on the same element, you can achieve economies of scale and access to production capacity, products are becoming more cheaper, it is also possible to obtain a wider range of popularity.

As for the study of artificial intelligence, we are active in open source and open our research. All we have to participate in the mainstream of the General Assembly. I major in assessing the progress they have made with the team recently. A few years ago they developed a program called & ldquo; the product, the product is a way to short-term memory is attached to the gyrus neural networks, that was something we had never had a function; Memories & rdquo. That was in 2014 research, then there have been research papers cited the work. Each month, the emergence of new paper demonstrates that improvements were part of the basis of the questions and answers reference the new technology upgrade.

Compare the entire industry together to create and accumulate results, and the results will be achieved if we are alone, you will find a huge difference, because we can work on emerging rapidly advancing research and development than we do to be fast on their own light years time. The basic technology can make a lot of people outside of Facebook can also benefit. Just within the capability, we will be happy to share that kind of technology.

Q: How did you let yourself take place inside and outside the company to keep abreast of the latest innovations?

Nova Luopu Fu:I think that's what makes me think that he might have the best place to work in the industry. First, I'll read anything that I can see. And even better than it is to sit down with the team to discuss their ongoing work together. This is by far the best part of my daily work.

Just a few weeks ago, I told the Facebook team for the exchange, we are doing all the work they carried out a one-day briefing. Has written many articles gyrus neural networks of the thesis in the 1990s, Jahn & middot; music library (Yann LeCun) is leading the team to his company's vision forward. The team not only in the assessment of the work they are doing, because they are public and open their research, they also assess other people based on their research findings to solve similar problems arising from the works.

In addition, I will talk to some of the social development of the latest VR applications communicate with people, to see how we can try in those areas. I also have the opportunity at the Conference on Science and Technology with the industry or are doing work for me very interested in talking to people, I have the opportunity to understand the industry dynamics nowadays. It's interesting. Honestly, to keep abreast of the latest innovations is not easy, because all the time there are a lot of interesting and attractive things happen.

Q: After you get a chance at the headquarters of the cruise, I want to ask is that you have not thought about any place not only in these walls, still other Facebook appears to promote collaborative creation of office space?

Nova Luopu Fu:You're in the building is one of the largest single building in the world. This layer has 2,800 employees, did not set any personal office & mdash; & mdash; you can see that we are all on their desks. The purpose of this design is, as a test, to see thousands of people make the same at a room, we were able to collaborate to the kind of point.

Obviously, we have one secret weapon is Facebook itself, we also have to provide it to others. We spent all day on Facebook, because we work here, it becomes an excellent collaboration tool: you have Facebook groups have Facebook Messenger, there are a variety of good and help you understand the various types of polymerization way of information, such as let you know what a friend recently busy, or learn 16 different teams at the same time do things. Those tools can facilitate your understanding of these.

I think a lot of people tend to overlook is that collaboration is the key, people are able to put forward their views and ideas and expert insights, to have the opportunity to understand another person, to understand each other is doing something concrete. This may be a cross-sectoral collaboration. For example, I was doing machine learning, I want to understand doctors try to do something to find patterns in drug discovery. I have more understanding of their problems, the more I give them help.

Our culture is a lot of time to bring basic empathy, because you're on Facebook, you not only can see what work colleagues are doing, you can also learn about the daily life of them; the children go to school next week , or they resort back. It will bring something I did not experience any other equivalent body mass with our organization had a sense of cohesion.

Q: As the organization has grown into a giant science and technology, for you to maintain a small company atmosphere and entrepreneurship have concerns about how much? Obviously, there are many innovative companies set up earlier than you had glorious moment, but after subvert it into a corner position.

Nova Luopu Fu:That's what we have been thinking about the problem. The company's growth is also good and bad places. The benefit is that we can engage in a variety of exciting things we have in all of these areas top experts. If you are a new company engineers want to know more about AI, aircraft or virtual reality, we are able to meet your needs. That's fine, it is a great advantage. However, let larger group of people united to work together to achieve a common mission, is a very challenging thing.

I think it comes down to we have been doing a few things. First, you want the staff to have a real sense of the ultimate goal line. In large organizations, it may become federalization, that we do different things, so that their goals are not unanimous. We are our mission is very clear & mdash; & mdash; the use of technology to connect people & mdash; & mdash; we made it clear that if you are excited about, that's fine, join us; and If you're not really excited, then you should choose another employer.

I like to think so: a lot of our work is all about culture. If you will be seen as the development of this construction system, then the way we work together is a system, we can take the time to build such a system & mdash; & mdash; the easiest is to provide training to new employees. For many companies, the new training day and a half just paperwork. For us, it is a six-week intensive training camp, this design is intended to allow you access to the technology as much as possible, and as much as possible contact with a company.

To the end of the six-week training camp, no matter what kind of work you are, you've seen the company engineers are able to reach hundreds of people, but you just do not touch the contact's shallow. Period, you wrote some code and let you review; you wander in the company, talking with engineers, but also ask them to vulnerabilities.

So, when you have finished training camp, everyone will get to know each other well, because whether you are employed for over 15 years, or just graduated from college, we are doing is to understand Facebook, which allows us to build relationships . This relationship can be spread to the entire company, you establish contact with those in the collective, and all those who have seen the company a lot of people, so when someone on the team to do something in doubt, he would have thought & ldquo; Oh, I have seen Mary in the study of this vulnerability. I'll ask her, maybe she knew whom to turn to. & Rdquo; this can build up those loose connections, but they are very important to create a kind of collective cohesion.

For this I can say stop, because we have to create a collective cohesion and designed a project. In addition to training camp, we also implement job rotation: Enter the post after 19 months, you can draw a month full rotation to another team. During which you can learn new skills, can work in a completely different work area, you can see a new group of colleagues, that the continuation of the company's internal cohesion.

Q: Have you personally experienced that right, you and Cheryl & middot; Sandberg (Sheryl Sandberg) swap jobs for a week.

Nova Luopu Fu:of course. Language I learned what her job is, and vice versa. We attempt this job rotation can be implemented anywhere, I think it can help. That hard to deal with, so many things to do yet, but we are focused on studying the problem.

Q: In the hiring process, how do you assess the talent? Especially now that the company aimed at long-term goals.

Nova Luopu Fu:The first thing you want to do is, whether they have the skills needed to complete the job & mdash; & mdash; as compared to the past, we now demand for the skills to be much more widely. Mainly in the past it is to find software engineering skills with a talent, and now we will have to find electrical engineering, mechanical engineering or AI professional talent. For these positions, we have different skills assessment methods, but basically is mainly based on a technology validation, to see if they are located in the field of first-class talent.

In addition, collaboration is also very important for us. One part of the interview is a collaborative deal with the problem. It consists of two parts: one is to see if they answer questions correctly, the other part is to see whether they can cooperate to find answers to your questions together with another person, because this is the reality of the way the world works. We want to ensure that people have good communication skills, be able to collaborate well with the team.

China IT News APP

Download China IT News APP

Please rate this news

The average score will be displayed after you score.

Post comment

Do not see clearly? Click for a new code.

User comments