Doing a Machine Learning PhD While Working in Japan

Commenters discussed the experience of doing a machine-learning PhD while working in Japan, with the author (mcyc) saying the route worked well for them, recommending people work before and during a PhD in CS, and noting a possible end to the MEXT university track. Several responders praised the practical perspective and asked about the author’s earlier choice of Korea; sbinnee warned that Korean PhDs can take eight-plus years based on anecdotes, while sstim attributed long Korean timelines to grant-driven busywork rather than being universal. gregjw suggested ML is relatively safe from automation, and sashank_1509 countered that AI automation of research is a bigger concern than geography.

Opinion split on whether Japan’s system really “pushes” three-year graduations. tkgally commended the writeup but pushed back that three-year targets vary by program and field, describing a language-and-information-depart­ment culture where cross-disciplinary committees, lengthy revision cycles, and visa or funding constraints often make three years unrealistic. Others criticized the article’s academic tone - rrr_oh_man and monkpit attacked hedging writing - while LtWorf and suddenlybananas mocked or questioned some reactions. Overall, discussion ranged from practical logistics and national differences to worries about writing style and future automation.