In the past decade, artificial intelligence technology gradually transferred from academia to industry, into the digital economy industry. Through continuous fermentation media and venture capital, the new industrial system and artificial intelligence system that is born out. State policies have begun to support the digital economy and intelligent economy, we hope to achieve corner overtaking in the pattern of international competition.
In the context of the new normal economy and upgrading of industrial innovation, data and intelligence industry has become a trend to artificial intelligence industry as the core. Artificial Intelligence covers a wide range of industries, including various types of enterprises. In addition to technology giant's in-depth participation, other companies rely on their own advantages, to participate in competition, to build complex competitive landscape.
From a macro trend, artificial intelligence chain can be divided into the base layer, technology layer, application layer.
Base layer: the rapid development of hardware, giant layout open-source platform; in the field of artificial intelligence, conventional chip computing architecture has been unable to support the demand for deep learning and other massively parallel computing, which requires new hardware to better the underlying data store, accelerate calculation process. Mainly hardware-based core layer, which comprises a GPU / FPGA performance acceleration hardware like a neural network chips, and the sensor middleware, which is a prerequisite for the support of artificial intelligence applications. These operators provide hardware computing power to the entire artificial intelligence, present, many international IT giants based.
Technical level: artificial intelligence technology layer is the core of development, the extent of the application layer intelligence products play a decisive role in this development process, the development of algorithms and computing power of AI plays a major role. For enterprise applications, segments of the competition is more intense, especially in the perception of intelligence and cognitive intelligence-led, intelligence and perception of human-computer interaction to achieve that connect people and information, obtain the necessary data modeling. The cognitive stage of intelligent data acquisition modeling operation, the use of functional depth learning to think like the human brain the result.
Application Layer: After two levels of appeal, can be put into practical. For the application layer is concerned, according to the different objects can be divided into two major parts consumer end-use applications and industry application scenarios, and mode of bringing innovation to improve the efficiency with intelligent means. To telemarketing industry, for example, enterprises can be divided into vertical applications for consumer users and enterprise users into two categories.
For consumer users
of mobile Internet companies: These companies often already have the advantage of users and traffic, you can use the telephone robot optimize the user experience, improve operational efficiency. Personal assistant class applications provide enterprise big data platform can bring great convenience to users in daily life, developed rapidly in recent years, the major giants have also been layout.
For enterprise customers
perceive intelligent interactive classes: for enterprise customers, this type of artificial intelligence integrated into customer applications and products is an important trend. Intelligent hardware: robots, unmanned drones and other external entities relying on artificial intelligence and environments complete system-level products, intelligent hardware which play a key role.
And business intelligence solutions: relying on the large number of data center technology platform, coupled with the semantic analysis techniques, can solve more complete robot phone really easy to use and other issues, providing comprehensive intelligent solutions.
Weak AI has accelerated ***, strong artificial intelligence is not far away. Although the field of artificial intelligence industry usher in a new outbreak, but overall, artificial intelligence is still weak artificial intelligence application-specific field-based. At present, most of the system, including deep learning neural network, machine intelligence, is still not to the extent possible to think about. To continue to climb into the strong artificial intelligence, cognitive computer is necessary, which is currently trying to break point in artificial intelligence. AI from technology, and finally the scene, artificial intelligence will eventually need to achieve real landing by the depth of integration to the industry scene.