AI industry shows three distinct trends: the pace of innovation has become the core competitiveness of industries AI, AI accelerated industrialization, and industrialization of AI. AI is committed to providing industry solutions unicorn business giants, only continue to promote artificial intelligence in the direction of interpretable understandable development, reduce prejudice algorithm generated can be achieved between man and computer, and silicon-carbon-based biological mutual trust between the biological co-governance, better coordination of the world.
According to the China Science and Technology Policy Research Center, Tsinghua University, published "China's Artificial Intelligence Development Report 2018" shows that China's rapid development of artificial intelligence, has become the world a number of areas, including artificial intelligence patent portfolio, the number of relevant research papers, venture capital accounted for ratio, and artificial intelligence market size will continue to grow in 2020 to reach nearly 100 billion yuan of scale.
Short-term development of artificial intelligence can not be done overnight, on the whole, the development of general artificial intelligence and artificial intelligence is still in its infancy. This is not only at the national level have a clear understanding of the field of artificial intelligence research institutions, experts and enterprises have also has a rational judgment. Especially in applications, AI industry, real industry artificial intelligence B-side industry, there is only the beginning.
Simple to understand, computational intelligence that is fast computation, memory and storage capacity; intelligent perception, that is visual, auditory, tactile and other sensory ability, is very important and popular face recognition, voice recognition, that is, aware intelligence; cognitive intelligence is more complex, including analysis, thinking, comprehension, ability to judge.
From the point of view of the present stage of development of artificial intelligence, with the continuous development of computing power, escalating storage means, computational intelligence can be said to have been achieved; and as the popularity of mobile Internet, big data, cloud computing and other technological developments, non more the value of structured data mining and valued, voice, image, video, contacts and other related perception and perception of intelligence is also developing rapidly; in computational intelligence and perception development on the basis of intelligence, artificial intelligence is to be able to analyze, think, understand judgment and other cognitive intelligent extension, real intelligent solutions has emerged clues. We can say that the development of artificial intelligence has reached a critical point by the forward-aware intelligence to cognitive intelligence.
The current artificial intelligence (AI) to do is identify the comparison, after the lack of information into the brain, processing, understanding, thinking 'steps, just stay in the' perception ', rather than' knowledge '. "I want AI have similar brain activity, cognition come to the stage, it needs to acquire knowledge, reasoning .AI from" perceived intelligence "to" Cognition smart "practice currently more difficult under a common state, but in some areas such as money laundering, reconnaissance is being applied.
AI Artificial Intelligence is now in a weak state, it has no brain, to let it form the brain, the core is to have "knowledge." If the existing algorithm, the model is nerve or brain structure, then knowledge is the driving force behind the brain can be up and running. Let AI get knowledge base of knowledge in the industry as "knowledge map", it should not only focus on knowledge but also concerned about the association between knowledge points. These associations will give AI imaginative. Mentioned water, it is to react to the density, transparency and other properties, and also water, drink a function linked more advanced is calculated using a multi-channel vigorously to fetch water.
A knowledge map is very difficult, massive human knowledge of how to translate into the language of the machine, and connect with, very problematic. Especially before the work has been done manually, such as the Valley lyrics library, Baidu Encyclopedia, Wikipedia and so can be converted into a knowledge map, but the heavy workload, the content of abnormal complex.
"Associated density" is insufficient application difficult to achieve another "bottleneck." A knowledge point may occur in up to several relationships, can solve the neighbor's father Bill Gates's mother problem with this single dimension, but from the reflection of the real world "butterfly effect" where the interaction is very much far. In the absence of sufficient density is associated with, AI "brain" of tens of millions of entries even have a body mass spectra of common knowledge is difficult to achieve the level of applications.
Density associated with knowledge of AI is more like a cerebral cortex, the more complex the AI will be more intensive "smart." Some industry itself has its own foundation association graph, which provides help for its development patterns. For example, the public security system a set of systems, including people, places, and things, organizations, institutions and associations will join the immigration data, third-party logistics data came after, it can be fixed into a knowledge map. In studying the behavior of some of the laws of burglary, pickpocketing and other public transportation to do, in addition to the introduction of knowledge map development trend of historical data, risk perception, but also gradually joined the environment, weather and other data, including the value of PM2.5 found PM2.5 value of pocketing public transport is influential.
The more intense the more accurate diagrams AI knowledge map to do like "Sherlock Holmes" as a comprehensive analysis of large amounts of information to form an aggregated knowledge map, which challenges the traditional efficiency of artificial building knowledge map.
In fact, to explore the application of artificial intelligence floor industrial upgrading now the curtain has just opened. In the B-side in particular, is a traditional industry in particular faces a number of difficulties, including corporations, governments for AI's understanding and concerns, data acquisition, storage, deletion, fragmented and complex, application costs AI + industry high, lack of technical personnel team and so on. More importantly, the level of development in artificial intelligence technology itself, but also has not yet reached maturity, many issues still need to overcome.
But no matter how to say, artificial intelligence and depth of industry integration is an irreversible trend, the impact on human artificial intelligence, human-computer relationship will be deepened, the whole world will be more closely linked catalyzed artificial intelligence.