Tsinghua's Approach to Military-Civil Fusion in Artificial Intelligence
What does military-civil fusion (军民融合) in AI look like in action? I’ve translated an article by a vice president of Tsinghua University, often characterized as "China's MIT," which describes its commitment to supporting China’s national strategy for military-civil fusion, while advancing an "AI superpower strategy" (人工智能强国战略).
The Road of Military-Civil Fusion for Artificial Intelligence Development
You Zheng (尤政), Vice President of Tsinghua University
June 8, 2018
At present, artificial intelligence (AI) is a strategic technology that will lead the future and is also one of the fastest growing fields of technology today. The development of AI will have a major and even disruptive impact on social and economic development and on people’s lives and production. Colleges and universities are points linking the first productive forces for science and technology, the first resources for human talent, and the first impetus for innovation. In the process of implementing an AI superpower strategy (人工智能强国战略), (we) should fully give play to our advantages in personnel cultivation and in science and technology research.
Tsinghua University is one of the earliest units to systematically undertake research on AI technologies. In 1978, the Department of Computer Science established the “Artificial Intelligence and Intelligent Control” teaching and research group, and recruited the first batch of artificial intelligence master students. Tsinghua University has a number of AI research bases, including the State Key Laboratory of Intelligent Technology and Systems (智能技术与系统”国家重点实验室) established in 1990 and the Intelligent Microsystems Ministry of Education Key Laboratory of the (智能微系统教育部重点实验室), and several AI teams. Take the team of academician Zhang Bo (张钹) as an example; they have achieved globally influential results in the field of general AI (通用人工智能) as represented by natural language understanding. This team undertook the “AI Theories and Crux Technologies for Future Human-Machine Cooperative (Combat) Operations” (“面向未来人机协同作战的人工智能理论与关键技术) project with total funding of over 100 million RMB [about $15 million] from the CMC Science and Technology Commission National Defense Frontier Innovation Special Zone (国防前沿创新特区). Their research results in Bayesian memory learning theories and methods have attracted great attention in the field of international AI. (They) advanced the achievement of innovative results deep learning adversarial attack and defense (对抗性攻防) theories and algorithms for AI safety/security, winning all three championships in an international competition convened by Google. The team member, associate professor Zhu Jun (朱军), was selected among the MIT Technology Review’s “35 Innovators under 35” in China, and chosen by IEEE Intelligent Systems for “AI’s 10 to Watch.” It can be said that Tsinghua University possesses a very good foundation and accumulation (of research) in the field of AI, and with regard to serving the AI superpower strategy, is duty bound.
In accordance with central requirements, Tsinghua University will closely integrate the national strategy of military-civilian integration and the AI superpower strategy. Tsinghua University was entrusted by the CMC Science and Technology Commission to take responsibility to construct the High-End Laboratory for Military Intelligence (军事智能高端实验室). With regard to basic theories and core technologies, military intelligence and general AI possess commonalities. Therefore, Tsinghua University regards the construction of the High-End Laboratory for Military Intelligence as the core starting point for serving the AI superpower strategy.
In the process of constructing the High-End Laboratory for Military Intelligence (军事智能高端实验室), Tsinghua University adheres to the concept of gripping with both hands basic research and applied research. Premier Li Keqiang has repeatedly emphasized the importance of basic research. Without basic research acting as a support, it is difficult to sustain the development of high technologies, and it is also difficult to have real competitiveness. Therefore, the consolidation of basic research is a responsibility that colleges and universities should take on in the nation’s AI strategy. At the same time, AI development in China has very large dividends in application scenarios, and applied technology research is also very important. Therefore, Tsinghua University insists on basic research as a support in applied technology research in AI talent training and scientific research innovation, with military requirements as a guide, promoting the development of basic AI research.
You Zheng at a national military-civil fusion exhibition
Basic research refers primarily to the research on basic theory and methods of AI, and focusing research on core foundational issues in the AI discipline. It is the crux that we can, with leading international universities, must shoulder in AI research. Tsinghua University cuts into core AI research from two angles. The first perspective takes brain science as the main cut. The research focuses on brain science research and forward-looking exploration of brain-inspired AI models. In this regard, the layout is not focused on supporting mainstream applications of real-world AI, but focuses on a future active layouts. This kind of research has a strong interdisciplinary character. To this end, Tsinghua established a physical laboratory, the “Tsinghua Brain and Intelligence Laboratory” (清华脑与智能实验室) last year to promote interdisciplinary research between AI and brain science. The second perspective takes computing as the main cut, focusing on targeting in-depth research of the difficulties facing the current basic theoretical framework of AI. In AI, ultimate competitiveness depends on the innovation capabilities in the above basic research. According to the official data of the U.S. Computer Science Rankings, Tsinghua University has ranked second in the world in the number of high-level papers published in the field of AI in the past decade, indicating that Tsinghua University has already entered into the world’s advanced ranks in AI basic theory and methods.
Applied technology research refers primarily to the major national requirements, especially military requirements, with research and development of key application technologies, promoting the military-civil fusion in China’s AI domain. Tsinghua has already achieved rich results in research on AI application technologies. For example, (Tsinghua) with Tencent and Sogou has already undertaken more than ten years of strategic cooperation, with a large number of research results applied to Tencent’s WeChat and Sogou’s search engine systems, resulting in favorable economic benefits. Tsinghua University also pays very much attention to “hard bones” (啃硬骨头) in applied technology research, focusing on perception technology and supporting real smart systems. Tsinghua also established a school-level interdisciplinary research center, the “Tsinghua University Intelligent Unmanned Systems Research Center,” to further promote research and development in AI applications technology.
Tsinghua has also made useful explorations in institutional mechanisms. To promote cross-disciplinary research in the field of AI, the university established an interdisciplinary cross-disciplinary research small group, issuing several documents supporting interdisciplinary intersections. For example, in the a cross-curriculum teacher part-time system, a professor of the Department of Computer Science and Technology can take a part-time position in the School of Life (Sciences), and an interdisciplinary cross-discipline assessment system is also in place, which will play an important role in cultivating interdisciplinary scientific research talent in the field of AI. The school has just approved the “Computational Legal Studies” Master of Engineering program in the School of Law, which is also an attempt to integrate the school’s AI and liberal arts so as to try a brand-new specialty direction for the subject.
Finally, it needs to be emphasized, while seeing the development AI application technologies flourishing in China, we must certainly pay more attention to basic research in AI. Colleges and universities should settle their minds (定下心来), sitting on the cold bench in the hot field of AI, and daring [to take on] hard bones, especially in actually supporting military applications of AI. Only in this way can we ensure that our country becomes a veritable AI superpower.