This round of the CCF-GAIR Global Artificial Intelligence and Robotics Summit held a roundtable forum focused on the commercialization of robots. The host was Shenzhen Bi Robot Association Secretary General Bi Yalei, Senior Director of Ninebot Technology Zhang Jiachen, ROOBO Intelligent Butler CEO Liu Yingbo, Dahua Le Orange CEO Liang Lei and Wu Ling, co-founder of Wuling Technology, participated in the discussion.
Moderator: First hand over the time to Zhang Jiachen.
Zhang Jiachen: I am a product manager and I may introduce more about the things we are doing today and why we do things. There are three points for everyone to discuss.
First, the essence of robots is the upgrading of computers. The desktop-to-PC development breaks through the usual 2D interactive mode, implements the 3D interactive mode, and the input and output methods begin with physical methods. This is a robot from our point of view. It is not the only answer, but it is our idea.
Second, "mobile" will greatly expand the application of the robot. From the previous point of view, the robot is a new type of computer, so that this very powerful computer can follow you and accompany you to various places and interact with you through 3D physical output input. We hope that the robot is a small warm bear instead of a scarecrow. This is the idea that guides us to design products.
Third, there are two phases of innovation and entrepreneurship. The first phase is from 0 to 1, and the second phase is from 1 to 100. In the second phase, the allocation of resources is uneven and knowledgeable. At this time, for software engineers, there is absolutely no need to make a prototype. They can develop a lot of apps, lower the entry threshold for application developers, and let robots not only be used by people in small circles, but it is Let more people can play it.
Due to these three viewpoints, we made several decisions:
First of all, we decided to focus on R&D work focusing on mobility within 5 kilometers. This means that we have done a lot of computer vision and we have used the company's original two-wheeled balance vehicle chassis to explore the fusion approach and to move within 5 kilometers. Why is the balance car a good chassis? Because of the combination of the modes and vision that can be carried on people, there are actually three types of interactions, one is the mode of human control; the second is the mode of autonomous navigation; the third is the mode of human and vehicle, because you can ride The pattern moves with this robot. From the product point of view, this is a very clever way to solve the challenges brought by the mixed environment. What is a mixed environment? In fact, mobile robots and autonomous driving technology face very similar challenges.
As for reducing the entry barrier for developers, we have made a robot platform. According to the robot platform, we want to be a semi-open platform for software and hardware, and a win-win situation for developers. What we do is a wheeled robot machine that will provide a tabletop system, as well as some built-in apps. The whole machine has hardware development. We hope to leave a lot of space for developers. With more people involved, we can do this thing. As a detonation product, Ninebot is a genetic company with a very deep consumer-level goal. Therefore, we will also focus on consumer-level goals.
For example, at the 2016 CES theme opening speech, we released a new product. Its chassis is a balance car with more than 400 patented technologies. It also integrates the Segway Robot and can keep its riding mode forever. .
Now, there are already thousands of third-party developers from around the world, including Facebook, who have submitted third-party development applications to us, most of them in the name of the organization, as well as individual developers.
How to open the diversification path of robot commercialization?Moderator: Each company has a different brand, so how to open the path? I believe the path is very diverse. Ask intelligent expert Liu Yingbo to introduce your company first and then express your opinion on this issue.
Liu Yingbo: Our company can be simply positioned as an artificial intelligence + robot system company, through a series of smart robot products to carry out our business. For example, pudding robots accompanied by children, pet robots, assistant robots, etc. At present, we have 300 colleagues operating around the entire line.
Moderator: So many scenes, how to choose between and how to face the operator? How to develop the commercialization of products with everyone? Because now there are many robots, but in fact it can be like Zhang Jiachen just mentioned, thousands of collaborators to develop applications, this scenario is actually very difficult to achieve. Or do you think about parenting or pets you just mentioned as a breakthrough?
Yingbo Liu: The reason why we call Smart Robot Systems Inc., certainly represents us on the general direction of intelligent robots. May not be able to clearly answer whether it is good for children's robots or pet robots, or assistant robots. We call smart butlers and are optimistic about the application of intelligent robots. A product can not be big and complete, we will be divided according to the use of the scene and each user to use the group, the core is to solve the technology, product experience, as well as the so-called degree of just needed.
Moderator: Explain to us one of the cases, what kind of scene users can use your robot to do perception?
Yingbo Liu: For example, we are mass production of pudding robots. This little robot uses some of AI's visual capabilities because it is applied to children. If a 0-6 year old child appears in front of this robot, the robot will automatically take pictures or videos to the children and send the videos to their parents. If you are a parent, you must have enjoyed the videos or photos of your family friends. And after we introduced this feature, it does represent the user’s appeal in this area. There are about seven to eight hundred daily activities.
Moderator: Many of the Korean robot's publicity perspectives are very similar to what you said. Combining IT with RT, there will always be a lot of information services behind. The commercialization scenario that you are currently talking about is a bit like combining the Internet and the Internet of Things to find the specific needs of users in one direction and let the robot be one of the tools . This can be understood, right?
Yingbo Liu: Yes.
How do big companies enter the robot market?Moderator: Dahua is a very famous enterprise for video. Dahua is involved in the robot market from the perspective of parenting robots. What do we think big companies will do in the robot market?
Liang Lei: We are not a pure robot company, but we say an open platform company with video as its core. Ninebot also mentioned that for robots to be open platforms, we also open platforms, including video technology, cloud. Originally, we were a 2B industrial company. In order to enter the consumer market in the civil market, we have the Le Orange brand. At the time, it was also a coincidence that the robot still focused on the vertical field to make it transparent.
At that time, we wanted to make videos at home and we could see old people and children. Therefore, we had a small music education robot. Everyone can see that it includes voice intelligence technology, video intelligence technology, etc. But in fact, the space for video is very large. From the detection and collection to the identification of the final features, the application of the scene can be enriched. For example, when a child appears, the robot can record his bits and pieces. It may be seven or eight hours a day, and it is very difficult for parents to have so much time. Our video technology is condensed to one minute.
Moderator: The robotics industry is more penetrating than the medical machinery and biomedical engineering fields. Because the transition to all industries is relevant, there will be opportunities for the sticky words. Robotic application scenario The ecological imagination space itself is very large. What I would like to ask is, in addition to your own products, who needs to work in the field of education?
Liang Lei: Our traditional ID companies are not good at Internet operations. Xiaole Robots is a typical Internet add-on networking product. It cooperates with Baidu, Himalayan, iQiyi, and Sina Parenting Educational Institutions through the new interactive method of robotics. .
Moderator: Changed a gameplay.
Liang Lei: For an interactive way, there will be new entrances. If there is no new interaction or mobile app, it will still be part of the mobile Internet. The parenting robot's interaction can be separated from the mobile phone and no need to give the child a mobile phone.
Is there a most valuable robot application scenario?Moderator: Ecology itself is cross-border, but still have a fundamental foothold. Because Zhang Jiachen has just introduced the company, ask a question related to your business. Just now you listed a lot of application scenarios for us. To the purpose, it is inconvenient to introduce which scenario you now think is the development goal that you think is valuable.
Zhang Jiachen: It's hard to say which one. The mature technology can find something that can be commercialized. Let me give you an example. At the invitation of the Singapore Department of Economic Development, I arrived in Singapore two months ago. Because the aging of Singapore is even more serious than ours, while many elderly people are living at home, even if the robot can only do a very small thing, it can be used in all public rental housing in the Singapore government. In Singapore, 70% of elderly people are hospitalized for dehydration. If they can be detected from a visual point of view to some extent, or through other methods to monitor whether the elderly drink enough water on time, then remind them, even if it is so small One thing, from the point of view of health economics, can greatly reduce medical costs.
Moderator: You just said that on the one hand, we can supplement the capabilities of robot intelligence through wearables. On the other hand, we can also combine many social services, such as the childcare service just mentioned. Until now, what kind of support can you give these developers at the beginning? What promise should you give at the beginning? Or what kind of costs should be paid for him to enter this field?
Zhang Jiachen: Our developer program is still at an early stage to do things in a more experimental way, but it is more convincing that developers plan to charge. We need to find developers who are willing to pay for this to enter the platform and complement each other. We will provide very close technical support. At the same time, we will evaluate developers in three stages. One is to look at the market potential he has developed, the second is the strength of the past, and the third is whether the scene he developed is a wheeled platform of ours.
Moderator: They created a platform that was open to developers and attracted developers. In the process, they chose their own partners and gave each other very close technical support. Just now you also introduced your company's current platform. How do you think about the key commercialization path?
Gu Jiawei: It was not long before the industry got up and it was talking about the business model. This will make the business model more difficult to do in the early days. First of all, free is hard . I think this is a time to be tricky: Can it be a business model that changes the mindset of consumers and help users change the model when connecting to a new service? This is what we are considering. From this we have found a very interesting phenomenon - if today's robots are hard to make others think that they can solve the problem, then we must find two cores: First, can we break the business model, then find the user Reconnect innovation.
Liu Yingbo: As mentioned by Zhang Jiachen, the robot is an extension of the computer, and many of the components used are common things in the industry. However, Gu Jiawei also mentioned that the chances of doing a single product are gone now, and this industry obviously has not yet reached the stage of market explosion.
Moderator: Heard a lot of such views, and it is very hard to do it. It seems to be impossible to do it.
Liu Yingbo: A very high threshold has already been formed. Our wave of Internet or mobile internet companies are looking at the platform benefits. Other technologies are mature enough, the technology is better applied in the product, the cost is reduced, and when the market breaks out, the opportunity of the platform comes. Everyone must definitely look at the long-term market. But when it comes to when today, I feel that only fortune telling can be said.
Liang Lei: I think the future is certainly bright, but the status quo is definitely cruel. Now that there is no flash point, we do not clearly see which one can break out. The platforms, products, services, hardware, and software are integrated. We haven't yet served users well. How do you say one of the breakout points? The prospects for the future development of this market are certainly promising, but at present it is necessary for everyone to do their job well. The future of each market area, metaphor of mobile phones, mobile phone market was not so big a decade ago, there are still people doing the platform, doing services, finished products, and key hardware, each one doing a great job. The robot is still such a rule. When it really breaks out, every one of them will have strong players. Everyone will serve the users well and stare at their vertical fields.
Moderator: Regardless of whether it is a component or a platform, we must finally look for customer needs from the perspective of the user.
Liang Lei: Right.
to sum upModerator: Several guests discussed the path of several robot commercialization scenarios from the perspective of their own business and their own practice. There is actually a commonality among them: We don’t actually emphasize that it looks like people, but we must Do valuable work for people. How to reflect its intelligence? AI, intelligence itself is the ability to perceive, cognize, and interact with. I also tell you that our Institute of Technology and Technology is looking at the ability of human-computer collaboration. It is how we perceive that some people's actions can have some reflections that are valuable to people. Simple opening and closing is just a need. What we are doing now may be a structure that helps people stand up, or an action that uses mind to direct a prosthetic limb. These products themselves will stimulate people's just need, where these are just needed, and where the ultimate commercial solution is.