Be cautious with university rankings. Universities can be assessed by research, student satisfaction, teaching quality, cost, accessibility, or by specific fields. Some excel in computer science, others in medicine or the humanities.
A single overall ranking is therefore meaningless - look instead for the measure that matches your priorities. For instance, for research impact in computer science, see: https://csrankings.org/
Similarly watch out for ultra-specific rankings that are used to dupe you into thinking the school excels at something that isn't a real category. My state alma mater managed to rank #2 (to HBS of course) in "international business" for years but really has a worthless MBA program. Pure marketing.
>Durham University improved by 30 places year-on-year in its students’ evaluation of teaching quality, which was the main driver in securing its third place in the overall university league table.
Yeah this is worthless. "Elite" colleges in the US value student evaluations of teaching VERY VERY highly in things like tenure review. Well, guess what kind of grades professors give almost every single student at those colleges... It's a race to the bottom. Students aren't experts in teaching; they're rarely even experts in learning coming out of modern secondary education.
This is a cool site. For anyone who doesn't feel like clicking, the top overall is Carnegie Mellon. There are three from China in the top 10, the other seven are American. ETH Zurich is the first outside of China/USA at number 12.
However if you select only AI, Carnegie Mellon drops to 3rd and only two of the top ten are outside Asia (mostly China but also National University of Singapore and KAIST in South Korea).
But honestly, why would you activate the "hype" category? I can go to my local community college if I want to hear people say "my AI is alive" or "you should always optimize locally."
That being said... CMU always had a decent program (as did UIUC.) They came down in my estimation when they started granting trade school degrees. In the old days, you had to take a class on parsing and foundations of computing to graduate. You could talk to a grad who knew what a recursive descent parser was and what it's drawbacks were. And there's a chance they could understand the basics of Turing's paper(s).
There's a maxim I read somewhere that undergrad was supposed to change the way you thought more than teaching you facts (and maybe skills.) What I liked about the old system is you were taught a concept (functional decomposition, for instance) and then concrete examples were given to support the concepts. Now it seems like students are taught where to download python packages and enough unix commands to type in example code from online forums.
CS pedagogy in the states is a joke, but I guess if I want someone to write a PHP plugin for wordpress, I know where to go.
Research impact correlates to size, which may be directly opposed to getting a good education as smaller institutions may have more resources per student.
Not true. It quantifies publications in top conferences. These are highly competitive, and you cannot just publish any junk there. If the site were to measure ANY publications, then it would be indeed skewed, as you suggested.
I'm not refering to junk papers by quantity, but that not all papers at the same venue are equal, and some fields take longer. For example a logic/programming-language paper often takes much longer to theorize than an AI paper.
Even within the same field should a paper that speeds up LLM's by 10x be equal to one that improves object detection accuracy in a specific dataset by 2%, or a dataset paper? All could be published at NeurIPS and all are good papers, but that does not make them equal in impact.
A better example might be that a hypothetical paper curing some cancer through some magic technology, or a differnt one detecting cancer better using AI could both target a similar venue (as evidenced by various nature puplications), but are vastly different in impact.
US News rankings are garbage based in no small part on opinion surveys and famously manipulated year over year.
Though I strongly disagree with their choice of conferences, probably the best regarded ranking of computer science schools is CSRankings.org (https://csrankings.org/)
A single overall ranking is therefore meaningless - look instead for the measure that matches your priorities. For instance, for research impact in computer science, see: https://csrankings.org/