美国计算研究会怒怼USNews全球大学计算机排名

　　美东时间2017年11月11日，代表美国200多家计算机研究机构与学会的美国计算研究会(Computing Research Association, CRA)发表公开声明，抗议US News不久前发布的全球大学计算机科学排名。

　　在2018US News全球大学计算机排名中，清华大学计算机高居世界第一、UT Austin、南洋理工分别榜眼和探花。前十名之中，只有五所美国大学，而亚洲高校也占据了半壁江山。特别值得一提的是，中国的清华、华科、和浙大都进入了全球十强。这一世界排名新格局，显然令美国学界不太适应。在公开声明中，CRA要求业界人士直接忽略这一“荒谬”的排名。

　　CRA成立于 1972年, 由200多家活跃于计算机领域的北美研究机构组成，包括各个高校的计算机系，政府、工业界、和大学的实验室和研究中心，以及各大相关学会组织，如AAAI, ACM, CACS/AIC, IEEE Computer Society, SIAM, USENIX等，其目标是通过政府、业界、和学界的紧密合作，促进计算机科学领域的研究创新。

　　US News的这个排名，显然动了CRA的奶酪。在公开信中，CRA声称，US News排名方法在计算机领域存在严重问题，如只纳入期刊论文、而没有考虑在计算机领域更具影响的会议论文。所谓的研究声誉，对于横跨半球的两个学校而言，也像苹果和梨子一样，不具可比性。此外，排名数据也不公开和透明。因此，CRA论断，这个排名就是胡扯：

　　Anyone with knowledge of CS research will see these rankings for what they are — nonsense— and ignore them. But others may be seriously misled.

　　据称，CRA曾就排名与US News多次交涉，排名推出之后甚至要求US News撤稿，但US News并未理会。

　　以下为CRA公开声明全文。

　　CRA Statement on US News and World Report Rankings of Computer Science Universities

　　To the Computing Research Community:

　　The latest US News and World Report (USN&WR) ranking of Computer Science (CS) at global universities does a grave disserviceto USN&WR readers and to CS departments all over the world. Last week, we respectfully asked the ranking be withdrawn. Unfortunately USN&WR declined.

　　The methodology used — rankings based on journal publications collected by Web of Science — ignores conference publications and as a consequence does not accurately reflect how research is disseminated in the CS community or how faculty receive recognition or have impact. Furthermore, the list of venues is not public. So while some may debate the soundness of any bibliometric-based rankings, there will be no debate about the flaws in the rankings USN&WR has published; the methodology makes inferences from the wrong data without transparency and, consequently, it arrives at an absurd ranking.

　　Another important factor in USN&WR rankings is reputation. However, reputational rankings in a large, mostly disconnected community, arguably are problematic. It is unreasonable to expect that departments half-way around the world will have anything close to an accurate assessment of each other, given that they are speaking a different language, have different value systems, different levels of resources, and different goals.

　　Anyone with knowledge of CS research will see these rankings for what they are — nonsense — and ignore them. But others may be seriously misled.

　　CRA hosted a discussion which included Robert Morse, USN&WR lead on rankings, at Snowbird 2016, our biennial meeting. Among the many items discussed, there was a robust discussion of rankings using bibliometrics and the various issues involved. In particular, we outlined the problems with using a data source that does not index the conferences in which the most impactful and highly cited peer reviewed research publications appears. CRA offered to partner to create a ranking scheme that would be credible to the computing research community, but he declined to work with either CRA or the broader community. On November 1st, after the global rankings were published, we contacted Morse pointing out the problems with the methodology used and asking that these Global Rankings based on flawed source data be withdrawn, but he did not respond.

　　We urge the community to ignore the USN&WR rankings of Computer Science.

　　Sincerely,

　　Susan Davidson, Chair CRA Board of Directors, Weiss Professor, University of Pennsylvania

　　Andrew P. Bernat, CRA Executive Director

　　Carla Brodley, Dean, Northeastern University

　　Laura Haas, Dean, University of Massachusetts Amherst

　　H.V. Jagadish, Bernard A. Galler Collegiate Professor, University of Michigan

　　Kathryn S. McKinley, Google

　　Mario Nascimento, Chair and Professor, University of Alberta

　　Fred Schneider, Chair and Samuel B. Eckert Professor, Cornell University