CSE #1 in Published Papers at Top Cryptography Conferences
According to an authoritative study by professors from MIT and the University of Maryland, the Computer Science and Engineering (CSE) department at UC San Diego has the #5-ranked program in theoretical computer science in the country. The new ranking of computer-science departments is based on the number of papers accepted to major algorithms and theory conferences and weighted according to whether the conference is considered a Rank 1 venue or Rank 2 (with Rank 3 and unranked conferences not taken into consideration).
Buried in the data was more good news: CSE faculty at UC San Diego had more papers accepted to the top conferences in cryptography than any other U.S. university. They had 43 papers in Advances in Cryptography (CRYPTO), and 36 papers in the European Conference on Cryptography (Eurocrypt). The data were updated as of July 2014.
“This ranking is based purely on conference publications, which more closely reflects both the field's own internal valuation and the evaluation of the university system,” notes CSE Prof. Mihir Bellare (at left). “But one should take any ranking with a pinch of salt, since publication quantity may be loosely related to quality and decisions taken by conference program committees are not perfect.”
For all conferences in theoretical computer science (including cryptography), UC San Diego came in #5 after MIT, Carnegie Mellon, UC Berkeley, and Princeton. While the latter four institutions were in the top-5 in U.S. News’ ranking for theory programs in computer science, UC San Diego ranked #14.
“It is heartening to see a young department outperform much older established departments and schools in computer science. The results reaffirm our strategy of hiring excellent core faculty and creating a collaborative environment,” says CSE Chair Rajesh Gupta. “Pursuing this strategy over the past decade, we have built a strong research program in cryptography, computer security, programming languages, bioinformatics and computer architecture. Our ongoing efforts will push frontiers in data sciences and intersections of computing to societal infrastructure and human health.”
In detailing their methodology, MIT professor Erik Demaine and University of Maryland professor Mohammad T. Hajiaghayi noted that the best measure of productivity in the computer-science community and a large fraction (if not all) of jobs, awards, grants, products, etc., in computer science are based on how prolific the researcher is at top conferences.
"Due to our belief on lack of transparency and well-defined measures in methods used by U.S. News to rank CS departments in theoretical computer science (and in general)," wrote Demaine and Hajiaghayi, who undertook the study as part of their Big Dynamic Network Data (BigDnD) project. They attempted to provide "a ranking based on a real and measurable method for top 50 U.S. universities."
To decide which algorithms and theory conferences should be considered Rank 1 or Rank 2, the authors used a standard list produced by Georgia Tech, and then deleted all Rank 3 and unranked conferences. The university was given a full point for each publication in one of the eleven Rank 1 conferences, and half a point for each Rank 2 conference publication.