The reason is really pretty simple - AI/ML is an engineering discipline, and, as such, is not something that LACs are good at, because, they are not engineering schools, except HMC. I mean, you wouldn’t expect LACs to have that many offerings in mechanical engineering as well.
It’s like expecting Olin to have a good selection of advanced courses in social sciences or in Cell and molecular biology.
Is he an undergraduate? If so, unless he’s just doing basic stuff, the sort of things that are generally done by administrators, this can be a problem.
Of course I do remember when my wife was on her department’s grad committee as a grad students. They first culled more than half based on GPA and GRE scores, and then organized them by degree, country of origin, etc. It was (and still is) a top program in CS, so they got a huge number of applications.
Yes and no. many departments get funding for grad students from the colleges or university. Grad students in many programs are supported in large part by TAships. In other programs, RAships provide the support, in which case faculty need to have money for these.
In general, the money for RAs comes from research grants.
Like the fact that they cannot hire faculty fast enough to keep up with the demand. The limitations on accepting PhD students in anything connected to AI is having PhD advisors.
I agree, perhaps partially. Few top LACs (other than HMC, Swarthmore and few others?) offer engineering, and even fewer offer mechanical engineering (other than the military academies?), but most of them now offer CS, which includes AI/ML. I also don’t think of AI/ML, and CS in general, as an engineering discipline, but a cross between engineering and math. And math is a traditional liberal arts discipline.
He is a grad student (first-year PhD candidate). Few, if any, of the top CS schools require GRE scores anymore. GPA may have been a filter, but it wasn’t material by the time he got the applications. He was tasked to look at the applicants holistically (their background, their academic preparation, their research interests and publications, their SoP, their recommendations, etc.) He was doing something not unlike what an AO does in undergraduate admissions. He could reject or recommend an applicant, but not accept her/him.
BTW, nearly all the top programs in AI/ML interview their final candidates (often by the same professors whom the applicants identified in their applications and who would sponsor them if they’re admitted). That’s the final step/check.
The only undergrad school producing a higher percentage of students who go on to get PhDs than Harvey Mudd is CalTech. They aren’t in humanities or social science.
Harvey Mudd, Swarthmore, Williams, Carleton, and Pomona all produce a higher percentage of PhDs (per enrollment) in Computer Science than Princeton.
In Math and Statistics, Princeton is third behind CalTech and Harvey Mudd, but Swarthmore, Pomona, Reed, and Williams are in the top 10.
The idea that LACs don’t adequately prepare students for PhDs is being drastically overplayed in this thread, as it often is on CC.
I do believe that LACs “on average” prepare the students better for grad programs. The only Caveat “could be” the quality of the program the students go to from an LAC vis-a-vis R1s.
Congratulations! It’s not easy getting into a good PhD program in CS. My wife is in CS, actually in AI/ML (and in Big Data), which is why I am a bit more familiar with the inner workings, and the difficulties landing good faculty.
It sound like what my wife got stuck with many long years ago… I remember her “spreadsheets” (on actual pieces of paper) with all of that info. I think that their rules didn’t allow any individual to reject anybody, just put their yes or no vote, but their applicants had already been culled by GPA, by the staff (back then departments had many more office staff)
It’s been a few decades, but I remember her piles of paper with tables drawn on then on the livingroom floor.
But if you look at Reed it is the highest PhD placement only compared to Caltech, I doubt Reed students are as high quality as Harvey Mudd. Same is true for some other even lower ranked LACs as well. For example: Wooster, Kalamazoo, Juniata and many others. Btw many top colleges like Williams, Amherst etc. are not top PhD producers.
One possibility is that students from these colleges do not get jobs so they are forced to do PhDs. Little but far fetched but it is possible. It would be good to have a statistical backing to the target Universities for PhD.
Many LACs do have CS departments and CS majors (note also that many larger universities offer CS as a non-engineering major), but many of the LAC CS departments are small and spread thin offering the bare minimum upper level CS courses to make a CS major.
I chose those schools because the posters above who are suggesting that LACs inadequately prepare PhD students are both using Princeton as their baseline, so IMO looking at top LACs makes the comparison more apt.
Not sure to what you are referring here. Adjusted for number of students, and looking at all PhDs, Williams is 8th, ahead of every R1 institution except for CalTech and MIT. Amherst is 20th, ahead of every R1 institution except those two, plus Chicago, Princeton, Yale and Olin.
This isn’t to say that top LACs are a better option, but the opinion that they inadequate is, IMO, overstated.
People who decide in high school that they want to consider a PhD eventually make informed choices of where they want to go. I am sure LACs are adequate for some. Not for others. We wouldn’t know as much as to what they would know about their own circumstances. If someone says a LAC is inadequate for their needs,’it probably is. And vice versa.
Isn’t the prevailing religion on CC that you should let the kid decide everything? Even 8th grade kids …
That is pretty misleading. Part of the UC’s mission is to admit people of different backgrounds in California, specifically lower income, since UCs can’t use race. Lower income households tend to have lower test scores, the private colleges ahead of UCB have more flexibility especially when it comes to how wealthy they want their class to be.
