Source: DoNews
Original headline: Highlights of Robin Li’s 2020 Baidu World Congress speech.
DoNews, September 15 (reporter Liu Wenxuan): the annual Baidu World 2020 conference was held online on September 15, where Robin Li delivered a speech. Li said Baidu World has run 14 consecutive sessions, and at each one Baidu hopes to present the past year’s newest scientific and technological achievements through all kinds of demonstrations.
Li added that this year’s conference differed in two ways: first, it was a purely online broadcast, run with CCTV News; second, 2019, through the marriage of 5G and AI, became the true first year of AI.
Meanwhile, given the complexity and changes of the domestic and international environment this year, AI’s commercialisation has not moved as fast as many imagined. Through this year’s presentations, Baidu hopes to build confidence that AI can still bring huge change to our lives, huge growth momentum to our economy, and more positive influence to our society.
Li on the 2017 Fifth Ring Road ticket
From 2017 to now, the Apollo project and autonomous driving have transformed completely. When I rode a driverless car on the Fifth Ring Road back then, I never expected such a response. Thinking it over, it is understandable: if one day none of us needs to drive, and sitting in the car a computer carries you wherever you want, that is a huge change for life and society. Since then, I feel the policy environment has changed: Beijing, the Ministry of Industry and Information Technology and many local governments have issued policies and regulations allowing Baidu and many other tech companies to test driverless cars legally on public roads. On the other hand, we feel this is a sacred mission: if we can build it early and let everyone enjoy its convenience sooner, the meaning for society, and for each of our lives, is very great.
Li on China’s first mass-produced purpose-built driverless car
The greatest difference between the Hongqi EV and an ordinary car is the cylindrical object on its roof called a lidar. A true driverless car today must have lidar, like the driverless car’s eyes, seeing any obstacle 100 or 200 metres away. Around the vehicle are various sensors, such as millimetre-wave radar and surround-view cameras, helping the car control, decide and dodge obstacles. This car came from Baidu’s cooperation with FAW and carries the Hongqi badge. The biggest difference from the driverless cars most other companies and institutions build is that ours was designed as a driverless car from the initial design stage, so it deserves the name purpose-built driverless car. The cars often seen on roads are bought from others and then fitted with all kinds of add-ons. Baidu and Hongqi together built China’s first mass-produced purpose-built driverless car. It already runs on roads in many Chinese cities and is in open testing in many of them.
Li on 100,000 people having ridden driverless cars
In several places, ordinary people can hail such a driverless car through the Baidu app or Baidu Maps, for example in Changsha, Cangzhou, Chongqing and Beijing, though in Beijing only in districts such as Yizhuang and Shunyi. By our count about 100,000 people have tried driverless cars, and I believe the number will quickly rise to a million or more. In the past it was not open: you needed an appointment and prior vetting. Now, in Changsha and elsewhere, anyone can hail one, and usage jumped at once.
Li on 5G cloud co-driving
5G cloud co-driving is a new term, the newest progress of Baidu’s driverless technology over the past year. Our driverless cars obey traffic rules 100 percent strictly, and when a car meets a situation where following the rules means it cannot get out, we reach the cloud through 5G cloud co-driving: a safety operator in the cloud takes manual control and frees the car. This is one of the few technologies we have seen that truly uses 5G, something 4G could not do. On your phone, 4G and 5G speeds feel much the same, with no essential difference. But without 5G, cloud co-driving genuinely cannot work, because enormous streams of real-time imagery must return to the cloud before we can operate safely.
What is the meaning of 5G cloud co-driving? In the past, testing driverless cars meant a safety officer in the car, ready to intervene manually when something unsolvable appeared. Today the car can carry no safety officer, because with safety operators configured in the cloud, the car can be freed on the spot.
Li on AI changing transport and the economy
I believe that within five years, that is by 2025, driverless technology will surely enter large-scale commercial use. That does not mean the streets will then be full of driverless cars. What truly affects everyone’s travel and life is AI’s change to the whole transport system.
For example, real-time full-volume data recognition intelligently adjusts traffic light timing, achieving smart signal control; and vehicle-road coordination lets roadside equipment talk with cars. These technologies can raise traffic efficiency sharply. We have calculated that AI can lift the passage efficiency of existing vehicles by 15 to 30 percent.
Beyond efficiency, it clearly promotes economic growth. The automobile industry is a giant industry, about 16 percent of GDP. If passage efficiency rises 15 percent, then 16 percent times 15 percent gives 2.4 percent: GDP’s absolute value grows by 2.4 percent. If traffic efficiency rises 30 percent, GDP grows 4.8 percent, and that is a gain every year. On one hand, with no driving bans or purchase limits, everyone’s travel gets smoother; on the other, our economy can move faster. The stakes are enormous.
Li demos Xiaodu’s newest feature: biometric recognition
On one hand, Xiaodu can tell different people apart by image recognition; on the other, it recognises voices too. Xiaodu is a home device, and a household usually has more than one member, each with different tastes and habits, so different people get different experiences with Xiaodu. By recognising each family member, we deliver personalised service.
Li on Xiaodu’s help for the elderly
The phone’s main human-computer interaction is still touch: users must read tiny characters and tap tiny areas, which is inconvenient for the elderly. But natural-language interaction poses no barrier for the old or the young. In the home scene, Xiaodu’s value for children and the elderly alike is very obvious.
A note on sourcing: this article is reproduced from other media to convey more information and does not represent this site’s endorsement of its views. Its originality and the statements it contains have not been verified by this site, which makes no guarantee or promise as to the truthfulness, completeness or timeliness of this article or any part of its content. Readers should verify relevant content themselves. This site bears no direct or joint liability for any infringement in such works. Should any content here infringe your rights, please contact us promptly and we will handle it within 24 hours.
Want to talk through what you just read?
We work end to end, from product design to launch. Tell us about your scenario and we can start there.
Shenzhen XZS Technology Co., Ltd. · Tel 0755-82494862 · 13760424861