The ten hurdles of the artificial intelligence big data industry in 2017

Source：Updated：2018-01-08 08:15:39

In 2017, artificial intelligence technology has driven the development of big data industry, a scene of prosperity, hiding the following concerns.

The development of artificial intelligence and big data in 2017 has encountered the following 10 growing pains:

1. Artificial intelligence has no IQ standard

One of the best areas in the field of artificial intelligence is unmanned, and the reason is that it has global standards from L0 to L5. But in other areas, too much artificial intelligence niche, the relative lack of standardization, voice recognition to do what successful, image recognition how to calculate the best not qualitative, humans have IQ to assess whether intelligent, but the AI did not. The industry has no standard to let the market, the manufacturer and the user can only move in the dark.

2. Artificial intelligence extending edge

The application of artificial intelligence is calculated by the center for support in the past, but because of the marginal expansion of application requirements, the Internet of things will be included, the next development of artificial intelligence and edge computing began to become a leading role in the field of intelligent.

3. AI embedding overcomputation and cloud computing

Artificial intelligence technology not only extends to the marginalization of user level, but also extends to the deeper computing demand. The development of cloud computing and supercomputing technology is leading artificial intelligence into an unprecedented field. As one of the three main factors of artificial intelligence, the development of computing power has allowed artificial intelligence to dabble in things.

4. Artificial intelligence is not gaining momentum

The limiting factor is the industry application of artificial intelligence is still not development, at present the main application in the search, intelligent voice assistant and intelligent household fringes of small-scale use, there is no play their own advantages. While unmanned driving requires at least three years of preparation time, and the industry rules are not yet defined, in general, ai is just a publicity stunt that has no ability to convert profits.

5. Artificial intelligence employment is wrong

The conflict between artificial intelligence and human employment has been on the agenda. Many worry that the development of artificial intelligence will lead to mass unemployment, and a large number of people in the industry, led by hawking, fear that the development of artificial intelligence will lead to human destruction. These concerns, though not unreasonable, could also lead to other jobs, such as the operation and maintenance of ai devices.

6. Big data vents have been passed

Big data from wind to clearly show now has been cover of artificial intelligence, in other words the big data is of s has passed, and even used to advocate big data, data must be set up enterprise's chief officer of the company, said that after 5 years has dropped. The fact that big data is doing better than it is is also one reason why big data is hard to do.

7. Data preservation is difficult

When big data enters the non-internet industry, the main problem is data collection. Incomplete data collection may be of no value to the analysis results, and the comprehensive collection may make the collection cost too high. At the same time, the big data of any industry is faced with the shortness of data preservation cycle and the difficult problem of data expiration.

8. The black market of data

The development of big data industry has led to increasing data value, so the main target of hackers has been focused on the data. Enterprises not only need to timely classify, process, store and analyze data collected, but also be responsible for the safety of data. In the increasingly complex network environment, the black market data is increasing and the cost of protecting data is increasing.

9. Data explosion management is out of sync

Big data after the outbreak of rounds, enterprises face the huge amounts of data in processing data data filtering and data cleaning is required, but the data cleaning process is complicate data management, the meaning of unstructured data contains is increasing, data cleaning can remove the useful information to leave only want to see the incomplete information, this way of data management will lead to the final processing result deviation.

10. Big data commercialization island formation

Every enterprise has its own data, and these data in commercial applications, due to factors such as personal privacy and competitive relationship, does not communicate with each other, is a direct result of the lack of liquidity and update data in the data vary in different enterprises, islanding data forming, the islanding data lack of comprehensiveness, allowing most enterprises can only get the answer within their own bias analysis results, far from the real data.