Wednesday, July 19, 2017

Intel: The Integration of AI and SI Makes Robot Industry Grow and Prosper

Song Zhiqiang, the Dean ofIntel Labs China, ILC, claimed the integration of artificial intelligence  (AI) and smart interaction (SI) would make intelligent robot industry burst out. The whole article is as follows.


Today, we are embracing a new world of Internet of Everything. More and more things and equipment have achieved interconnection through network, which makes data outbreak. Data surges like a blood and has becoming the most important driving force in the technical field. Artificial intelligence, automatic drive, 5G, VR / MR and a series of forward-looking technology, give us the opportunity to fully release the potential of these data, and constantly upgrade the human life experience.

We would like to share and communicate with you about Intel's observation of forward-looking technology, and investigate the present and the future of Internet of Everything. Let us begin from the perspective of robot and artificial intelligence to find out industry opportunities in the era of data deluge.

As the most important application area of artificial intelligence, robot has been always paid high attention. Intelligent humanoid robots appeared in Hollywood blockbuster give the public high expectations. However, the intelligent capability of real intelligent service robot is far from that in the movie. With the trend of artificial intelligence promoted by deep learning, it has become a very important issue how intelligent robot industry breaks out. I think there are two key aspects: one is the deep integration of AI and SI; the other is the security of intelligent robots. We'll take about the security later, and now let's focus on the integration of AI and SI.

In the academic world, AI has a commonly-used definition, is to understand the essence of intelligence, and be able to produce a kind of intelligent machine that can respond to external input, similar with human being. The ultimate goal of this discipline is to give the machine intelligent reaction capacity, so advanced AI robot can be seen as the ultimate goal of artificial intelligence.

There are two ways to understand the essence of intelligence: One is to speculate how people is thinking in an intelligent way from the observers' behaviors by virtue of philosophical or psychological methods; the other is to cut apart human brain to check the neuro anatomical structure, and find out the laws of intelligent activities through brain activity detection technology and well-designed experiments.

We have a knowledge of intelligence (not exactly right, but at least the model comes out), but how to make the machine smart need? It needs mathematicians, computer scientists and automation experts to study thoroughly. Therefore, there are many disciplines related to artificial intelligence. Besides soft theory, models and algorithms, it also needs the practice of hardware. For example, intelligent machine needs chips to provide it with a strong computing and storage ability. In addition, advanced humanoid robot requires the technical support of complex system control, either.

In fact, this wave of artificial intelligence is also benefiting from the improvements of calculations and storage technology. ANN is to train machine with massive data, and then machine learns about the inherent laws from the data and build a model, finally, the model will help to speculate new data. This is called a machine learning process. It requires a lot of storage and computing power, but now we are living in an extraordinary appropriate era of machine learning's big development.

No comments:

Post a Comment