What Makes a Top-10 CS Department?

Policy, Research

I’ve just returned from a visit to KAIST, the Korea Advanced Institute of Science and Technology, in Daejeon, South Korea. As far as “big government bets” go, KAIST has been an unabashed success. KAIST alumni dominate the upper echelons of virtually all of Korea’s tech companies. It is no exaggeration to say that KAIST made the “economic miracle” of Korea possible.

My visit included a distinguished lecture, meetings with the president, provost, and deans, and sessions with faculty members. What I saw was a modern, western-style, leading-edge institution of higher education trying very hard to move up in the international pecking order. To do this, they have been rethinking virtually everything they do, including adopting American-style tenure processes, changing all instruction to English, aggressively expanding the faculty, and establishing 11 interdisciplinary research centers (each with its own building!).

The desire to “move up” seems to permeate every department, including KAIST’s CS Department. The CS Head, Key-Sun Choi, explained his belief that to be “top 10″, his department would need to establish fundamental new research terminology, develop more faculty with international distinctions such as ACM Fellowships and Turing Awards, and have a department head with a significant international research reputation.

Of course, I couldn’t help but think about CMU CS and note that we meet all of Dr. Choi’s criteria. (Well, I suppose the last criterion is debatable ;-) But is that all there is to it? I’m not sure. One of the most distinctive characteristics of US research universities is the international makeup of the faculty and student bodies. Some top US universities have more than half of their faculty and graduate students coming from other countries, and this represents a stark contrast between KAIST and CMU. This creates an international outlook and a diversity of views. But, more importantly, being so international enables CMU to draw from a much, much larger talent pool than KAIST currently does. And, in the end, what probably makes a research university great, more than any other thing, is the talent level of its faculty and students.

My sense is that KAIST is one of the world’s top research universities, but probably approaching a local maxima given that the faculty are drawn largely from the native population. In order to access a global talent pool, the Korean government may have to do more than increase KAIST’s funding, but also increase emphasis on basic (as opposed to purely applied) research and help make Daejeon an inviting place for expatriates.

Peter Lee @ March 18, 2009