Doing a Machine Learning PhD While Working in Japan (tokyodev.com)

49 pointsby pwim18 hours ago16 comments

drnick1 10 minutes ago

While I would certainly consider jobs abroad, I think that not getting your PhD in the US is a huge risk, especially as an American who might eventually want to return to the States. American PhDs basically trade a substantial premium and are employable in academia worldwide, while the converse is exceedingly rare. Most foreign academics working in the US got their PhDs here.

mcyc 2 hours ago

I am the author of this. Happy to discuss!

One thing to note is that, after I submitted the final draft of this article, the Japanese government ended the MEXT university track or is at least planning to. There was a memo, but the status is not clear.

Anyway, I had a great time doing my PhD this way, and generally I recommend that people work before a PhD and try to work during it (if you are in CS). It gives you a lot of freedom and, at least for me, it is very helpful to have more than one pipeline of work at a time so that when one slows down (as it inevitably does in research), you can draw on the other.

sbinnee an hour ago

Congrats! I am surprised that you first chose Korea for PhD. As myself being Korean, I hear a lot of horrendous stories about PhD students in Korea. They tend to spend for more than +8 years. I know a few people who went through their PhDs and certainly it took +8 years for all of them. I did my PhD by the way in France and it took me 4 years.

Did you know that it could take so long when you first started your PhD in Korea? Could this be one of your reasons to move to Japan?

sstim 44 minutes ago

FWIW, the "extended" Korean PhD is an artifact of grant projects that need people to do the busy work (usually reports), and it seemed to be uncommon to have foreign students thrown into that pool. Scholarship max durations are also a limiting factor.

One aspect worth calling out as a huge plus of Japanese national institutions which the author didn't disclose is funding for compute is much better than what one would get from a Korean institution (e.g. the author's school runs one of the biggest GPU clusters in Japan).

tkgally 2 hours ago

Excellent article. I have worked for twenty years at another university in Tokyo and taught and supervised many non-Japanese graduate students, but I still learned things from the author's perspective.

One small pushback: The author writes, "the Japanese PhD system pushes you to graduate in three years." This applies in some programs and some fields, but is by no means universal. In the graduate program where I used to teach (I am now emeritus and no longer supervising students), the steps leading up to the Ph.D. and the expectations for the thesis made it practically impossible for students to finish in three years, especially if they continued to work part-time. This created particular challenges for foreign students on MEXT scholarships who had to work out other ways to support themselves and renew their visas while they continued to write their theses.

mcyc 2 hours ago

Thank you! And, re: 3 years, that's good to know, I should have been a bit more precise.

What field were you supervising?

3 years in CS is, imo, a little tight but not bad. But I always wondered how people in other fields (stem and otherwise) did it. CS simply doesn't have to deal with the extraneous factors that come with, e.g., biology or climatology.

tkgally 2 hours ago

I was affiliated with a graduate department in language and information sciences, and the fields of the faculty and students ranged from literature, history, and philosophy to theoretical linguistics and natural language processing. I don’t recall any students doing hardcore CS, but it wouldn’t have been inconceivable, especially if there was an NLP connection. But the academic culture of the department leaned more toward the humanities and social sciences than to science and engineering.

One of the factors making the finish-in-three-years target very difficult to meet is that the thesis committees would include faculty from different fields and backgrounds, often with differing expectations for what a Ph.D. thesis should include. It was not unusual for a committee to ask a student to make a variety of substantial changes and additions to a thesis draft before submitting. That revision process could take six months or longer; because of the schedule for graduate coursework and defenses, that alone would push the student beyond the three-year target.

Add in all of the other stumbling blocks that might come up in graduate school—family issues, financial problems, an advisor who retires or gets ill, etc.—and three years was no longer a realistic goal in that department.

rrr_oh_man 4 hours ago

This sounds SO.MUCH. like the type of hedging academic writing I despise

  Doing a PhD is difficult and isolating, which is exacerbated by being an international student in an unfamiliar country. And so, while I overall very much enjoyed my time there and have freely recommended it to many people, the decision to pursue a PhD in Japan should always be considered carefully.

(Watch https://news.ycombinator.com/item?id=49879250 why it's bad)

LtWorf 4 hours ago

If you can't detail why it's bad in any shorter form than a 90 minutes video, perhaps it's not that bad.

monkpit 2 hours ago

Agentless passive, hedging, zombie nouns. It’s the type of writing that’s meant to meet a word count.

amemi 16 minutes ago

But that does not seem to be related to academics, as GP called it.

Perhaps it is the contrast of "freely recommending" and "always should be considered carefully"?

Or taking a larger leap: the awkwardness of saying how hard (isolating) it is to do a PhD in an unfamiliar country, then saying that you enjoyed it, but others should consider it carefully?

sashank_1509 5 hours ago

I would worry more about AI automating your PhD than whether you did it in Japan or EU

suddenlybananas 5 hours ago

What's the point of this comment? Did you even read the article?

rrr_oh_man 4 hours ago

psychosis

msdz 4 hours ago

No, you don’t get it. The until now completely un-recursive not-self-improving training runs will automate themselves away any day now, and ML PhD’s will become worthless!!

gregjw an hour ago

Surely considering the field he got the PhD in, hes in the safest field for job security regarding this.