align="left"> Today, the development of the robotics industry like a raging fire, the popularity of industrial robots. In the days before the Chinese held conference on artificial intelligence, Chinese University Hong Kong, academician of Academy of Engineering China (Shenzhen) President Xu Yangsheng told reporters that a long time in the past, research scientists have focused on the robot action, now is the time to consider the problem of intelligent robot, the robot ", can be more intelligent action is more important than".
Academician Xu Yangsheng engaged in robotics research has been nearly 30 years of history. He believes that the robot contains three aspects, one is the action, one is the perception (sensing) part, and the other is the cognitive (Artificial Intelligence) part. "Like the wind. Do you know the wind? This is the problem of perception. After the wind should add clothes, this is the cognitive problems. Then we might decide to add clothes, which is an action problem." Xu Yangsheng believes that the robot has been around the science for 50 years to wear clothes this thing, did not want to wind it, you need to add clothes, and so on.
Xu Yangsheng pointed out that, even before scientists in the study of the robot's motion problems, there are many problems to be solved in the process of intelligent. The climbing robot as an example, there are many challenges, such as the robot to adapt to the different thickness of the branches, the weight can not be too heavy, climb and climb up with problems in different ways etc.. In addition, the calligraphy robot design process, but also contains the contents of intelligent learning.
Xu Yangsheng introduced, in fact, as early as 20 years ago, scientists have started to discuss the problem of distributed intelligence, but before it is a difficult thing to do, with the progress of the development of network and technology, can now carry out research in this area.
Xu Yangsheng believes that in robotics research, development and action related intelligence, first of all should consider how to learn from nature, learn from. For example, over the past two thousand years, people have been doing the dream of flying, want to do big wings, tied to the arm, hoping to fly like a bird. But has not been successful, of course, the plane is another matter, with the original people's dream is not the same. Why not do it now? I think it is not enough to learn from nature, nature has a lot of smart things have not learned."
Where does the intelligence of the robot come from? Now researchers are accustomed to realize the algorithm, that is, learning from the model". Xu Yangsheng felt that you can learn from people, you can also learn from animals, now there are too few people to study animals, especially animal intelligence".
In addition, Xu Yangsheng on the role of big data is also very seriously. He believes that big data with intelligence as a train of two wheels, which is not a lack of. After the big data, if there is no intelligence, there is no effect; in turn, rely on big data intelligence, so these two things are linked together."
"In the future, we also have to consider the scalability problem of intelligence." Xu Yangsheng said, if you can expand, what is the extension of the conditions in the end is the hardware, software, or data, this problem is worth pondering."