U.S. Computer Science Delegation Visits China
Summary
In the spring of 2006, an NSF-sponsored delegation of CS scientists (mostly School Deans and Department Chairs) visited various CS research centers and departments in China and met with peers. The purpose of the visit was to improve our knowledge of CS research in China and to establish a dialogue with our peers.
The trip included visits to: Beijing—IBM China Research Lab, Microsoft Research Asia, Institute for Computing Technology (ICT, an institute of the Chinese Academy of Sciences), Peking University, and Tsinghua University; Nanjing—Southeast University and Nanjing University; Xi’an—Northwestern Polytechnic University and Xi’an Jiaotong University; Shanghai—Shanghai Jiaotong University; and Suzhou—Suzhou University. In addition, a one-day U.S.-China Computer Science Leadership Summit, held at Beihang University in Beijing, provided an opportunity to discuss issues of common interest.
For more details about the trip, including a list of delegation members, see: http://dimacs.rutgers.edu/Workshops/China/. In the near future, DIMACS will issue a report on the facts, impressions, and recommendations gathered by the delegation. This article summarizes the impressions of one trip participant; the opinions expressed here are his own.
Background
China is a huge country whose economy is progressing by leaps and bounds. The high savings rate in China supports large investments that quickly modernize China’s infrastructure. In addition, the huge Chinese market attracts large foreign investments. As standards of living improve, the demand for higher education increases. China now has more than 15 million students and more than a million graduate students in the higher education system; it has the second highest graduate student population in the world, second only to the United States. Although new, modern campuses are growing quickly everywhere, the university system still has problems coping with the large demand. The competition for entering the most prestigious campuses is fierce.
As standards of living continue to increase in China, its economy is moving away from cheap manufacturing and into advanced technology and services. The government is attempting to accelerate this shift as it recognizes the limitations of cheap manufacturing as a source of wealth for China: an economic model based on a cheap workforce cannot reduce income inequalities and lift China’s countryside out of poverty; service jobs can be created with a smaller investment per job than manufacturing jobs; and high-tech manufacturing and services can reduce the strong dependency of the increase in China’s GDP on the increase in its energy consumption and pollution generation. China is keen to develop home-grown technologies in order to compete in “high-tech” with little dependence on foreign knowledge; it wants to educate a workforce that will support a sophisticated modern economy. Government investments in Computer Science and Technology (CST) research and education reflect these priorities; sophisticated IT is seen as essential to advanced manufacturing and services.
The Chinese Academic System
The current academic system in China is of recent, post-Cultural Revolution, origin: academic degrees (Bachelor’s, Master’s, and Ph.D.’s) were officially established in 1981. Chinese graduate students started studying abroad, and research exchanges with foreign countries were re-established at about the same time.
The higher education system is largely funded by the state, with top universities funded mainly by the central government and second-tier universities having more local government funding. The funding reflects national and local priorities. For example, the government recently funded the creation of software engineering schools (at the Bachelor’s and Master’s level) in order to educate the workforce of software engineers and programmers, which is needed to support China’s push into software services. The government funds projects at several universities and research centers that develop commercial microprocessors for desktop systems, microcontrollers, or DSP chips for TV sets. These projects are clearly aimed at providing home-brewed technologies for China’s push into consumer electronics. ICT is also developing video encoding algorithms to replace current standards and avoid the payment of royalties to foreign companies.
China’s services increasingly depend on private funding, and universities have not escaped this trend: an increasing fraction of university income comes from student fees and industrial ventures.
Admission into universities is by national entrance exams. These exams are taken at the same time all over China, and success in the exam largely dictates a student’s future. Students are working very hard to prepare and parents are investing large sums to provide good tutoring for the exams. A complex system of quotas provides different priorities to students both within and outside the university’s catchment area and for certain minorities. Admission will also depend on the subject of study chosen by the student. As a result, a student may choose an area of study to get into a top university, rather than studying in a preferred area and be assigned to a lesser university. Some universities are charging significant fees for students they accept above their quotas. Increasingly top universities are also establishing private or quasi-private schools, often in collaboration with foreign universities, leveraging their reputation and charging higher tuition in popular study areas like business (CS studies have not yet been affected by this trend).
Science parks are attached to many top universities. Technology developed in academia is used to start private companies located at these campuses. The relationship between such companies and academia can be very tight: the company may be partially owned by the university and by faculty at the university; research projects at the university are closely aligned with the company’s product line; and faculty and students work for the company. Since government funding of research is explicitly aimed at creating a local high-tech industry, and state and local governments often invest in economic ventures, the commingling of private enterprise and public research seems to be better accepted in China than in the United States.
Chinese faculty hold three ranks: lecturer, associate professor, and full professor. Only a subset of the full professors can supervise Ph.D. students. Younger faculty in the top universities are likely to have credentials similar to those faculty at U.S. universities; however, not all faculty hold doctorates. This is especially true of older faculty, whose careers were impeded by the Cultural Revolution, or faculty at less prestigious universities. The fact that advising is restricted to a small group of full professors encourages a hierarchical system where younger faculty and students work in large research groups that are headed by full professors who can advise. A senior professor may advise a dozen or more students with the help of the junior faculty in the group. The group leader also bears major responsibility for obtaining funding for the group—it seems that research grants are normally held by senior faculty.
The number of colleges and universities with degree programs in Computer Science and Technology (CST) has grown to more than 70; more than half of these are less than 10 years old. However, there are only 15 universities and one research institute that have been authorized by the Ministry of Education to offer Ph.D. degrees in CST. In addition, there are another 28 universities that can offer Ph.D.s in sub-disciplines of CST, and another 132 universities that offer Master’s degrees. Industrial research centers, such as Microsoft Research Asia (MSRA), also participate in higher education. MSRA is a “postdoctoral station”—that is, an institution that offers postdoctoral positions (considered a necessary step in the education of a Chinese academic). In addition, MSRA has several joint Ph.D. programs with universities.
The demand for CS graduates with doctoral degrees seems strong, and is boosted by the very large number of industrial research centers established by international companies in China. We were told of a drop in starting salaries for programmers, but no drop in enrollments at top universities.
CS research in China is broadly similar to research in the United States; however, there seems to be less emphasis on theory and foundations and more emphasis on system implementation.
Research and education funding in China is significantly lower than in the United States: China spends a lower fraction of its GDP on education than the United States (<4% vs. 5.9%). The budgets of research funding agencies are smaller: for example, the 2005 budget of the National Natural Science Foundation of China (NSFC) was 3.5 billion RMB, or about $430M. However, central government funding is also focused on a smaller number of top universities; these universities often have an up-to-date equipment infrastructure. For example, we visited several groups that work on (standard cell) microprocessor design; each had expensive testing equipment and up-to-date CAD tools. A significant fraction of funding comes from projects that in the United States would be considered development rather than research: university research groups, including faculty and students, will work on software infrastructure projects for the government. This has the obvious disadvantage of reducing the amount of time young faculty and graduate students can spend on publishable research. On the other hand, students receive practical training that may be missing in the United States, and faculty may access data from deployed infrastructure or from realistic software development projects that is harder to obtain in this country.
Our Chinese peers are very keen to raise the quality of research in their departments and to encourage publications. Universities attempt to do so by imposing procedures that evaluate and promote faculty according to their number of publications in international journals that are indexed by SCI or EI. (This is a problem for a discipline such as CS, where conference publications are more important than journal articles.) Also, Ph.D. and Master’s theses are now reviewed anonymously unless the content appeared in an international journal. The government has provided quality improvement funds to top universities that are used to hold international conferences, to support Chinese faculty in attending conferences abroad, and to attract world-renowned, foreign-trained scientists to China. While there have been a few such high profile hires (e.g., Andrew Chi-chi Yao at Tsinghua University), the total number is small.
Conclusions
China’s rapid evolution ensures that any report from this country will soon be obsolete; it is clear that China is gearing up to become a powerhouse in science and technology, and the U.S. research community should closely observe its rapid evolution.
Many opportunities exist for collaboration between the United States and China in CS: the Chinese universities we visited are interested in joint educational programs, academic exchanges, and joint research projects. Both NSF (who opened an office in Beijing during our visit) and NSFC are keen to encourage such collaborations. Language is not a major obstacle for the younger generation; the differences between the Chinese and American models, while creating some obstacles, provide unique opportunities because of their complementary strengths.
Marc Snir (snir [at] uiuc.edu) is Professor and Head of the CS Department at the University of Illinois, Urbana Champaign, and a member of the CRA board of directors.
