According to the Washington Observer reported on the 6th, recently, in the United States, a “high-end atmosphere” of the government meeting, there has been a Chinese kindergarten textbooks. An attendee has reportedly presented a Chinese science book on artificial intelligence at a conference on artificial intelligence and national security hosted by the National Security Council on Artificial Intelligence in the United States to illustrate the importance China attaches to artificial intelligence.
It is reported that the book was published by Henan People’s Publishing House, is designed for kindergarten children “artificial intelligence experimental materials.”
U.S. officials at the meeting include former Secretary of State Henry Kissinger, current Defense Secretary John Espe, current Energy Secretary Rick Perry and Senate Democratic Leader Chuck Schumer. Topics discussed at the meeting included the importance of artificial intelligence for future national security and how to respond to challenges from China.
Behind the Move, the Americans have revealed deep concerns about the rapid development of Chinese’s industrial intelligence, as well as concerns that the United States may be lagging behind on AI.
Although the United States has always viewed China as an imaginary enemy, often alarmist, Americans’ fears this time are not entirely unfounded.
As many Chinese companies expand into overseas markets, China’s achievements in artificial intelligence are becoming more and more known.
China has the world’s largest ai-intelligence start-up, worth more than $1billion, according to a report released in February by CB Insights, a research firm that tracks venture capital activity. Six of the 11 unicorn companies in the annual AI Start-up Ranking are from China, with SenseTime topping the list with a valuation of $4.5bn.
In 2017, China’s AI start-ups raised $4.9 billion, more than the $4.4 billion in the U.S.
In September, U.S. lawmakers warned that “China will close the gap between artificial intelligence and the United States by the end of 2018.” While the U.S. has been a world leader in research and development, spending $497 billion in 2015, China’s research and development spending increased 200 percent between 2000 and 2015 and is expected to surpass U.S. research and development spending by the end of 2018, the report said.
According to the Prospective Industry Research Institute, as of the end of March 2019, there were 5,386 active AI companies in the world. Of these, there are 2,169 in the United States, 1189 in mainland China, 404 in the United Kingdom, 303 in Canada and 169 in India.
In terms of the number of unicorn companies, there are 18 in the US and 17 in China, with Beijing leading the way with 468 in the Urban Division of Artificial Intelligence, followed by San Francisco with 328, London in third place with 290, Shanghai at 233 and New York at 207.
China is beginning to show some signs that it may be able to close some of the gap with the US, at least in the area of artificial intelligence chips, according to the Center for Data Innovation, a Washington-based think tank.
According to a previous study, the think-tank found that china had only one company in the top 15 global semiconductor sales in 2019, compared with two in the European Union and six in the US. But take the number of companies designing artificial intelligence chips in 2019, for example, the gap has narrowed , with 55 in the US, Followed by China with 26 and the European Union in third place with 12.
“The ‘American Dream’ is re-emerging in China, ” says one author who has grown up in the United States and has been to China several times. ”
(Screenshot of US media website)
Time to do something
But many analysts say Chinese companies are still heavily dependent on U.S.-made artificial intelligence servers to develop algorithms and conduct deep learning training, and are uncompetitive in the core areas of strategic technology.
Earlier, a study by Jefferies, China’s largest supplier of surveillance cameras, said that its biggest surveillance camera supplier, Hikvision, could face a shortage of supplies from the US. It will be possible to find alternatives to key components from the United States, such as cameras and storage devices, from domestic and Japanese suppliers. However, Hikvision still relies heavily on Intel and Nvidia’s CPUs and GPUs to drive AI servers.
“Nvidia’s GPU outperforms other solutions in deep learning training, thanks to its CUDA toolkit and libraries,” analysts Rex Wu and Lydia Lin said in a Jefferies report. “CUDA provides high-performance support and built-in artificial intelligence capabilities.
Kuang Kaim, an AI engineer based in Shanghai at Diannei Biotechnology, says cpu and gpu are essential for deep learning, which requires processing a lot of data on hardware. “They (CPUs and GPUs) are like roads. No matter what kind of car you drive, it must be on the road. ”
Li Ke, president of Dahua Technology, another big Chinese surveillance camera maker, said on a conference call that about 10 per cent of the company’s components were supplied from North America.
“For those ‘hard to replace’ parts such as cpu, gpu, FPGA and analog equipment, we also have a replacement strategy, ” Li said, adding that the company is currently in stock for more than a year to ensure it can adapt to any shocks.
“We’re going to redesign our products, ” said Huang Fanghong, secretary of the board of directors at Hikvision. If necessary, we will make our own chips. “Whether this is feasible is debatable. Although China can make alternatives to Intel and Nvidia, they are not yet up to the level of US companies.
Many Chinese enterprises have made good progress in self-made of artificial intelligence chips. Alibaba is producing its own AI chip and processor “Xuan Tie 910”, Huawei released the AI processor series, Tsinghua University team released the brain-like computing chip “Tianjic”, pi2star Technology released the Sticker series of artificial intelligence chips …
Dai, of Forrester Research, also points out to Chinese companies’ investments in AI hardware, noting that “however, the feasibility of chipset replacement depends on many other factors, such as Tapeout success rate, server compatibility, software interoperability, and workload applicability.” ”
In addition, China’s development in artificial intelligence has some other shortcomings.
“China is ahead in two areas,” said Eric Schmidt, a technology consultant at Alphabet, Google’s parent company. One is the area of face recognition monitoring. The other is financial technology. That doesn’t mean they’re leading the way in artificial intelligence as a whole. ”
A poll found that China was a “fast follower” but that “the best and most original papers are still emerging in the West,” Schmidt said.