500,000 people have entered the AI ​​field in more than a year

Have you heard of the two new professions?

  Artificial intelligence trainers are defined as new occupations. In just over a year, about 500,000 practitioners have entered the AI ​​field, and a variety of new artificial intelligence occupations such as 5G cloud driver safety officers and traffic accident appraisers have also been derived. .

AI new career one

A traffic accident appraiser who controls the "clairvoyance"

  Next to the monitoring equipment, the traffic accident appraiser quickly glanced at the screen.

He found that in the rain, a silver-gray car was parked in the right lane of a two-lane highway, and a red triangle road sign was placed 10 meters behind the car... After only 5 seconds of analysis and judgment, he obtained the "accident conclusion: Information such as "fault", "silver gray", "car" etc...

  "800 meters ahead, a white car on Shifoying East Road smashed and occupied the leftmost lane. Please change lanes in advance..." The accident information was released quickly, and the driver Mr. Li prepared in advance after learning the information on the vehicle terminal. A driving route is planned.


  "Drivers equipped with this Didi navigation system will be like me. As long as they walk this section of the road, they can click on the device to view real-time photos and videos of the accident site and avoid traffic jams in advance. "

  The reporter learned that at the beginning of this year, the official Didi Navigation-"Clairvoyance", clearly showed the accidents, traffic jams, and control conditions on the road in front of drivers to help them drive better and safely.

Liu Bing, the person in charge of Didi Navigation, said: “In fact, many navigation software can display road conditions such as accidents and traffic jams, but we also found that the navigation software currently in use lacks the types of accidents that drivers are most concerned about, the number of lanes occupied, and even It's how to change lanes and other details. Therefore, in the "clairvoyance", artificial intelligence technology is used to visualize the details of the traffic accident in front of the driver."

  As a result, a new profession-traffic accident appraiser was born.

At present, 20 appraisers look at more than 2,000 sets of photos every day to help "Clairvoyance" continuously optimize various functions and update the advantages of AI technology in real time.

This technology will also be expanded to more scenarios in the future, such as bridge water accumulation, roadside water stains, road construction and so on.

  "Speaking of behind the clairvoyance, it's not as simple as sitting there by the appraiser." Liu Bing said that it took a lot of thought to develop this technology.

In order to get first-hand pictures of the road conditions, Liu Bing and the R&D staff walked the streets, took the subway, took the bus... Finally, they focused on online car-hailing.

  "We use the on-board video equipment installed on the online car-hailing system to add an automatic orientation recognition system. In this way, each online car-hailing system becomes our camera, collecting information on various road sections in real time. These pictures and data are transmitted to the background. Terminal screening, and after in-depth analysis by the appraiser, a set of practical real-life pictures are returned to the "Clairvoyance", which can help drivers travel safely and conveniently." Liu Bing said that the automatic identification system itself also incorporates AI technology. It can eliminate general pedestrians, trees and buildings, but locks and recognizes triangular warning signs, cones, isolation lines, and traffic police and other specific signs to ensure that the view information returned to the terminal is accurate and effective.

Although clairvoyance is extremely intelligent, it can even automatically identify the lane where the accident and the accident occurred, but it still requires a traffic accident forensician to conduct an in-depth analysis to determine the cause of the accident, the type and color of the vehicle, and whether the traffic accident should be shown to the car owner. and many more……

  Liu Bing introduced that for an operable AI "clairvoyance", only recognition is not enough, it must be accurate and timely.

In this process, traffic accident appraisers played a key role.

  Liu Bing remembered that once, a picture appeared on the "Clairvoyance" screen, and a white van with the trunk lifted up was parked next to the fence of a section of closed construction road.

  "There are basically no social vehicles passing by on this road. Why did the algorithm grab this one?" The appraiser quickly determined that this was an invalid picture.

After the incident, after investigation, it was found that the van was indeed moved here after the accident.

AI did identify the accident car, but it has been towed away from the accident road section, and will not affect the traffic.

"Liu Bing said.

AI new profession two

5G cloud driving to help unmanned vehicles get out of trouble

  Located in Baidu Apollo park in Yizhuang, a 5G cloud driver safety officer is sitting in the cockpit.

This cockpit completely simulates the environment inside the car, with steering wheel, throttle, brakes, etc. all available.

The five screens in front of the security officer showed seven angles of the self-driving car he controlled.

On the front screen, the speed, gear and delay of the vehicle are displayed.

  The self-driving car on the screen is located in the training ground of Haidian Driving School.

Cone barrels are placed next to the vehicle to imitate the difficulties encountered by the vehicle on the road.

The safety officer in the cockpit is 80 kilometers away from the vehicle, and he wants to help the vehicle get out of the predicament with a 20 millisecond delay.

  The person in charge of Baidu Apollo autonomous driving introduced that at the end of 2019, autonomous driving began to be deployed at the unmanned level and is ready to start testing. In order to ensure the safety of road testing, the entire test system needs to have remote assistance functions.

Since then, the selection of 5G cloud driver safety officers has also been officially launched.

  One time when an autonomous vehicle was taking a road test in Shunyi, it encountered that the local area was irrigating farmland with plastic water pipes crossing the asphalt road.

In order to protect the water pipes, the irrigation workers blocked the entire road with wavy red banners.

This is a scene that has never been encountered by an autonomous vehicle before, so the vehicle stopped in front of the banner and automatically sent a remote help signal.

After receiving this signal, the 5G cloud driving safety officer can remotely view the road conditions, take over and drive the vehicle, and help the vehicle get out of trouble.

  "Therefore, the main responsibility of the cloud driving safety officer is to help the unmanned car get out of trouble when it encounters a scene that cannot be handled during the road test of the autonomous vehicle." The person in charge said.

  Wanting to be a cloud driving safety officer is more than just driving.

The reporter learned that every cloud driving safety officer is selected from car-side safety officers with more than two years of experience.

The person in charge explained: "The safety officer of the cloud driver needs to go through a week or so of training. The first is a series of training courses on safety knowledge, cloud driver and autonomous driving knowledge, and then a mature cloud driver safety officer gets on the car one-on-one. Guidance and undergoes rigorous practical road test training. Later, the cloud driver safety officer to be selected will have to pass a strict written test and road test before becoming an official cloud driver safety officer."

  Relevant technicians said: "With the advancement of autonomous driving technology, a true unmanned driving state of the entire vehicle will eventually be realized. However, 5G cloud driving safety personnel will always be needed, and they need to remotely monitor in real time on the road. Every self-driving car is guaranteed to be safe."


  New careers make up for the lack of artificial intelligence

  In the past two years, the Ministry of Human Resources and Social Security has officially released more than 20 new occupations to the society, some of which involve the AI ​​field, and employ around 500,000 people.

  Zhu Wei, an expert member of the Internet Society of China, believes that the world is currently committed to the research and development of artificial intelligence technology, and artificial intelligence has penetrated into various industries, such as data analysis, security, logistics, and manufacturing.

Because the development of artificial intelligence is still in its infancy, and many jobs require manual assistance, there are more and more types of occupations serving AI.

  With the development of artificial intelligence, some occupations are gradually being replaced.

Zhu Wei said: “For example, the guidance robots in the lobby of some office buildings have replaced the work of welcomers; the face recognition installed at the gates of some communities has also replaced the functions of security guards.”

  According to Zhu Wei's analysis, the biggest feature of AI is that it can accumulate experience through repeated learning, and constantly update and enrich its own intelligence.

In this process, most of the so-called new professions are transitional and will be gradually replaced with the development of AI.

Our reporter Zhang Qunchen