Grad Student Profile: Ke Yi

From the Spring 2006 issue of Threads

Ke Yi

After a highly successful undergraduate experience at China’s premier Tsinghua University, "Ke “Kevin” Yi sought out the opportunity to realize his research aspirations in a program that would provide him with an exceptional faculty and computing environment. In the fall of 2001 it was the good fortune of Duke CS to welcome this promising student into our scholarly community.

It did not take the faculty long to recognize the outstanding potential Yi possessed for working out solutions to complex mathematical problems, but then one should not have been surprised given the many honors and prizes he had achieved as an undergraduate. Among some of these were his earning a Silver Medal in the 9th International Olympiad in Informatics, which earned him exempted exam entrance into Tsinghua University, a Meritorious Winner in the 2000 Mathematical Contest in Modeling, and the Tsinghua Top-Grade Scholarship which was awarded to only 5 out of 12,000+ undergraduate students that year.

As early as his first year at Duke, Yi was conducting research work with Professor Jeffrey Vitter, which led to a paper that appeared at (the distributed systems conference) PODC’02. Starting from his second year, he began to work with Professors Pankaj Agarwal and Lars Arge on I/O-efficient algorithms, namely algorithms that can handle data sets much larger than physical memory. Traditional algorithms often become horrendously slow when they run out of memory, if not bail out completely. During the past three and half years, Yi has developed I/O-efficient algorithms for a number of important problems, many of which arise from databases and terrain analysis: range searching, XML labeling, terrain modeling, flow computation, etc. Recently, he has been working on an algorithm that automatically extracts important topological features from a terrain, and his new I/O-efficient algorithm turns out to be more than 100 times faster than the existing method!

Yi has always maintained a strong interest in databases, and has been actively conducting database research in collaboration with Professor Jun Yang and his group. The research projects span a number of different topics: query processing, XML, view maintenance, data streams, etc. Although not included in his thesis, Yi has greatly broadened his horizon from these projects, and is delighted to see how algorithm design could change the way data is managed in real systems. Apart from academic research, Yi also spent two summers at IBM Watson Research Center and AT&T Labs - Research, respectively, where he worked with industrial researchers and got a chance to apply his expertise in algorithms to real-world problems.

Yi, now in his fifth year and preparing his way toward his dissertation defense and degree completion, has a list of impressive accomplishments and award recognitions that attest to the exceptional gift this student has for conducting research and making an outstanding contribution to his discipline.

Return to Profiles main page