As I am going to the US for my graduate study, I think about university rankings quite a bit.

Regarding the ranking methods, there are generally 2 approaches to rank CS departments – reputation-based and publication-based. The first type is exemplified by the famous USNews, Niche, Times Higher Education, QS, etc. A disappointing fact I heard is that these rankings are based on survey results from department heads of other universities, which seems to be quite subjective and easy to be gamed, thus unreliable IF I really want to assess the activeness of research in the schools (which a grad student needs). However, a reputable school definitely has its real benefits – much resources to research, bigger network of good researchers (including the peers), probably a more beautiful campus in a beautiful city, and eventually a nicer name on the student’s CV to get jobs after graduation. Well, that’s a lot to weigh. But they are hard to get in. And more importantly, maybe the actual advisor that the student works with is not “good” (which is subjective to the student). The second type is represented by CSRankings.org, which features a pretty fair way to rank universities based on their publication output in top CS conferences. I used this the most. But I am still wondering why the website does not rank individual professors across institutions? Maybe they want to incorporate the “network benefit” found in the first method, where students should choose both an active research advisor and an active research program overall.

Putting methods aside, let’s think about why we have rankings and follow them in the first place. I recognize that these rankings sometimes overwhelm and obssess me. But when being mindful, I know that my ultimate happiness does not come solely from the rank (and the following benefits) of my university. For example, I can think of 2 professors that I know and are in very highly ranked schools. However, I will be miserable to work with them because one is simply working on topics that I am not too interested in (to put 5 years of my life working on), while the other seems to have a unhealthy rat-racing and money-making mindset. I am pretty much a blank paper about research visions and ideologies, so I want to go with professors that are aligned with my living values (which is not a blank paper, hopefully).

Finally, I do have the fetish of finding a ranking dimension to boost the rank of my institution. So, according the CS Rankings about the top US universities in NLP for the last 4 years, UT Dallas is ranked 23rd. So I will just say my school is in the top 30 of my field, based on some artificial and far-from-perfect metrics. Hopefully, in the coming years, I can be less and less obssessed with these metrics, focus on the-real-work with a good teacher.