Shulou(Shulou.com)06/02 Report--

This paper analyzes the two concepts of perceptual intelligence and cognitive intelligence based on the prediction of ten scientific and technological trends of Ali Dama Institute.

A few days ago, Alibaba Dharma House released the top 10 scientific and technological trends in 2020, with the trend of "artificial intelligence evolving from perceptual intelligence to cognitive intelligence" topping the list.

What is perceptual intelligence and cognitive intelligence?

It is generally believed that the main development directions of artificial intelligence include computational intelligence, perceptual intelligence, and cognitive intelligence:

Computational intelligence, mainly refers to fast computing and memory storage capacity.

Perceptual intelligence, that is, visual, auditory, tactile and other perception. Both humans and animals have the ability to interact with nature through a variety of intelligent perception.

For example, autopilot calculates driving information through perceptual devices such as lidar and artificial intelligence algorithms, and face payment is a device that confirms identity by perceiving face data information. The advantage of machines in perceiving the world lies in active perception, because machines can make full use of the achievements of DNN and big data.

Cognitive intelligence, that is, "understanding and thinking", has the ability to deal with complex facts and situations. The development of cognitive intelligence is divided into three levels:

The first layer is language understanding, the second layer is analysis and reasoning, and the third layer is the necessity of personality and emotion developing from perceptual intelligence to cognitive intelligence.

The trend report released by the Dharma Institute points out that artificial intelligence has reached or surpassed the human level in the field of perceptual intelligence such as "listening, speaking and seeing", but it is still in its infancy in the field of cognitive intelligence that requires external knowledge, logical reasoning or domain transfer. Cognitive intelligence is the key factor of whether the future artificial intelligence boom can further open the ceiling and form a larger industrial scale.

Take the public security industry as an example, face recognition only enables machines to identify themselves, but when it comes to investigating and solving cases, it is necessary to sort out a large number of clues and infer possible results, which can not be accomplished only by perceptual intelligence.

With the emergence of cognitive intelligence, artificial intelligence system is no longer just a simple statistical fitting of data, but actively understand the law and causality behind the development of things. Cognitive intelligence can help machines understand data across modes, learn the "general expression" closest to human brain cognition, and obtain multimode perception ability similar to human brain, thus bringing subversive industrial value.

The advanced path from perceptual intelligence to cognitive intelligence

The Dharma Academy believes that cognitive intelligence will combine technologies such as natural language processing, cross-domain knowledge graph, causal reasoning, and continuous learning to establish an effective mechanism for stable acquisition and expression of knowledge, so that knowledge can be understood and applied by machines, so as to achieve a key breakthrough from perceptual intelligence to cognitive intelligence.

Natural language processing technology and knowledge graph are the cornerstones of technological development in the stage of cognitive intelligence. Under this general trend, the enterprises that take the lead in laying out natural language processing and knowledge graph will have strong competitiveness in the era of cognitive intelligence.

Machine learning needs to be based on text, but there is great technical difficulty in understanding the text, and the text is processed differently in different scenes, so the value contained in the text has been difficult to be used effectively.

The function of natural language processing technology, popularly speaking, is to transform natural language into symbols that machines can understand and build the basis of machine understanding knowledge.

When the natural language recognition is conquered, the data of different industries can be connected, and even the online and offline data can be connected to prepare for the construction of knowledge base by artificial intelligence.

The development of cognitive intelligence is inseparable from natural language processing and knowledge graph, so enterprises that master these two core technologies can naturally gain greater value.

The layout and Development of Cognitive Intelligence in China

At present, in the market, cognitive intelligence technology companies are more based on their own business transformation, which can be divided into two categories, one is the transformation from big data analysis companies, such as Mingluo Technology, percentage points, etc.; the other is voice or language acoustics companies, such as iFLYTEK, Spice, Yunzhisheng, Express, etc., most of these companies are based on voice recognition, voiceprint recognition.

For example, iFLYTEK takes perceptual intelligence + cognitive intelligence as the main research direction of enterprises.

The goal of its perceptual intelligence is to achieve highly available perceptual intelligence, so that the machine has more perceptual ability; tasks include speech recognition, handwriting recognition, image recognition and so on.

The goal of cognitive intelligence is to achieve a breakthrough in cognitive intelligence and partially replace human mental work; its tasks include intelligent customer service, automatic marking, human-computer interaction and so on.

Another Chinese AI company, Quick Shang Tong, uses intelligent biometric technologies such as voiceprint recognition to build perceptual intelligent "intelligent biometric brain" and natural language processing technology to build cognitive intelligent "Aifu brain". Its goal is to hope that machines can take voice and language as the entrance, train expert systems in different fields through cognitive intelligence, and then empower the whole society.

When perceptual intelligence is weak, the emergence of cognitive intelligence can pull industrial upgrading to the fast lane. Nowadays, the technical barriers are still being overcome, and the general cognitive intelligence technology can not be used in various industries, but in the field of segmentation, cognitive intelligence has been able to land commercially.

In the field of public security, cognitive intelligence can extract and analyze the behavioral characteristics of criminals, and can develop a set of AI model for crime prediction, which can improve the early warning ability of social security events and the perception of risk situation to a certain extent. Landing cases can refer to the Express Voice print integrated combat platform, clear data big data research and judgment system, and so on.

Http://www.woshipm.com/ai/3287782.html

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.