March 2, 2017, the opening ceremony for China’s National Engineering Laboratory of Deep Learning Technology was held on the Baidu Campus, which is also considered to be the home of “China brain” (or officially known as Baidu Brain). This is a day which marks artificial intelligence as a important strategic deployment in China’s national policy.
Sun Wei, Deputy Director of the National Development and Reform Commission of Technology, and Robin Li, the Chairman and CEO of Baidu, as well as representatives from the Chinese Academy of Sciences and Tsinghua University, gave speeches at the opening ceremony and witnessed the brand new laboratory unveil.
According to Robin, leading the project is both an encouragement and responsibility for Baidu at the same time. The laboratory is a major milestone for China’s AI industry, “we don’t just lead national research, but also international research, so that researchers can see the lab as the end of a pilgrimage. In other words, we are the national team of deep learning—— not the Chinese national football (soccer) team, but aiming to be like China women’s national volleyball team”.
Figure 1: Robin Li , Baidu Chairman and Chief Executive Officer
Artificial intelligence is the Core of Baidu’s Core
Personally, I am very pleased to witness the opening ceremony of the national engineering laboratory, not only because Baidu is taking lead, but also because it had been a personal dream. –Robin Li
During the 2015 annual CPPCC meeting, Robin made the “China brain” proposal. This is a plan to build the world’s largest deep learning platform under support of the Chinese government. It will have the largest number of servers and the best algorithms to facilitate deep learning research for Chinese developers, entrepreneurs, researchers, with guaranteed infrastructural environment in software and hardware. Two years later, the National Engineering Laboratory is in construction under the lead of Baidu. Robin says he is “very content” with the way things worked out.
Robin also mentioned that, Baidu have been applying deep learning technology to image search for five to six years now. The company also founded the Institute of Deep Learning (IDL) four years ago, made the “China Brain” proposal two years ago, and is now handling the construction of the National Engineering Lab.
“AI is the core of Baidu’s core business, and we are very serious about it.”
In the last half decade, Baidu invested heavily in AI. In the past two years, their financial investment into research was 20 billion CYN. “We don’t offer the highest salary competing with China’s other top 500 enterprises, but our investment in research and development is definitely the highest, and most of our research input goes to AI.”
Andrew Ng, Yuanqing Lin, Qi Lu and other domain experts had joined Baidu successively. The company is a magnate to talents because of its generous support for research. After Qi Lu joined, Baidu made many high-profile changes. As the Baidu IDL laboratory director Yuanqing Lin addressed, “Qi Lu is a highly efficient person”. Baidu recently abolished its medical division and initiated a division for its personal robot assistant Dumi, which reported directly to Qi Lu. It also acquired AI company Ravens Technology, formed an autonomous driving commercial group, and now heads the construction of National Engineering Lab, indicating its determination to lead the AI community.
Behind the First National Engineering Laboratory in Artificial Intelligence
It is reported that the National Engineering Laboratory of Deep Learning and Application will build 4 levels of deep learning engineering platform, such as large-scale industrialized AI products, artificial intelligence application technology, artificial intelligence dispatch technology, and an overall technical architecture including 7 intensive research focus.
According to Lin Yuanqing, the director of Baidu Deep Learning Laboratory, “In the future, as to operation of the national lab, Baidu will provide software, equipment, and talent support to its development.”
Lin said that the laboratory will have an edge in auditory, visual perception, and natural language understanding, which are the three basic areas of artificial intelligence. Especially for transition problems from simple to complex scenario, including multi-agent image comprehension, speech recognition in noisy environment, and multi-round dialogue interaction.
At the same time, the lab will also focus on researching large-scale intensive computing, mass data training, computer vision, biometric identification, complex environment perception, new human-computer interaction, machine translation, standardized services, and deep learning intellectual property research, to build platforms, issue patents, standardized products, simulate environment, improve the national capacity for technical support, strengthen the cooperation between industry and research, and cultivate talents.
In the future, the lab will focus on providing better support to China’s artificial intelligence infrastructure, promoting the comprehensive development China’s deep learning technology and application in research and development.
It is reported that the daily request for Baidu brain’s speech synthesis function exceeded 250 million, the accuracy rate of its speech recognition is 97%, and for face recognition 99.7%. For two consecutive years, Baidu is on top of MIT Technology Review’s list for the global top ten breakthroughs in technology, ahead of the rest of Chinese companies. Now, the company leads the “national team” of deep learning and gains more authority in the field.
AI Industry Continues to Fight Over Talents
On August 26, 2016, China’s National Development and Reform Commission issued an “Special Notice for an Organization to Declare Innovation in Internet+ Industry”, which includes the construction of 3 national laboratories and the specifics of AI technology in Internet Plus. Deep learning is also marked to be the most important research focus.
According to Professor Zhu Jun from Tsinghua University’s Department of Computer Science and Technology, “AI will continue to be a hot topic, and the shortage of AI talent will remain to be a problem. Companies like Baidu, Alibaba and other enterprises are employing students from Tsinghua and easily paying them by the millions. This will bring up the cost for enterprises to employ AI talents, and cuts the funding for Research & Development. Of course, the final cost will be loaded onto users.”
Needless to say, the fight over talent is heating up: research and industry are placing talents as their top priority, while tech giants and start-ups are destined to battle over them in the future.
Nowadays, China is sprinting into the era of artificial intelligence at full speed. Private enterprises such as Baidu are in charge of building the national engineering laboratory. And this is sign of fair competition. Baidu, a company founded in the internet era, will have to continuously seize its upcoming opportunity in AI.
Original Article from Synced China http://www.jiqizhixin.com/article/2384 | Localized by Synced Global Team: Meghan Han, Rita Chen