Expectations for a research student at a Japanese university

Question

I have recently decided to enter a grad school in one of the Japanese universities (namely, TiTech). It seems that the general recommendation is to start as a non-degree research student first and prepare for entrance exam while being a research student. After reading this question, I still don't understand what is expected from a research student. On one hand, it is a first step before doing a masters degree, on the other - I find the explanation very confusing. After all, the word "reasearch" in "research student" is there for a reason, right?

My question is - do I have to come up with a research proposal before contacting my potential academic supervisor? I have found a professor doing a research in field I am interested in (inductive and abductive logic programming in AI), but the problem is that this is not something you can get experience with in undergraduate studies. It doesn't come up in an undergraduate curriculum, so I have only a general understanding of the topic. Starting a novel research (if that's what research student should do) without some background will be very hard. Yet I cannot get this background unless I get at least an Ms in AI first.

This is especially worrying given my background. I graduated from my university 10 years ago and was employed all this time as a software engineer. I have a good GPA (4.99 out of 5), and I lived and worked in Japan for 6 years and got a JLPT N2 certificate. However my bachelor thesis was about image processing, not about AI, and in any case it is not particularly relevant after so many years in industry.

Answer 1

I was a research student at the University of Tokyo for one year.

As the answer to the other question states, a research student has literally no responsibilities whatsoever. The apparent point is to allow foreign students time to study for the entrance examination for the formal graduate program. They might give you as little as four months (roughly) or as much as a year before the exam is held.

Your advisor may or may not recommend that you sit in on classes and learn how the university works. If you are in the sciences, there's a good chance that you will be assigned to a research team in preparation for entering the college. If you are in the humanities, it's likely that your only responsibility will be to try to pass the examination. If you are already familiar with your field, you will have an extreme amount of free time.

"Research student" is actually kind of a misnomer. During my pre-acceptance year I did independent research on my own initiative (mostly because I was bored) and had two papers accepted for publication, but my advisor did not ask what I was doing, and my office did not know I had published papers until after I was accepted as a masters student.

Answer 2

It seems that the general recommendation is to start as a non-degree research student first and prepare for entrance exam while being a research student.

This recommendation is for people who go to Japan on a Japanese government scholarship, and is due to the fact that the scholarship only covers your airfare when you go to Japan after having been admitted to your desired program. Thus, if you want to directly enroll in a masters (or doctoral) program, when you go to Japan to take the entrance exam you do so at your own expense, whereas if you start as a research student there is normally no need to go to Japan beforehand.

In your case, since you're probably not going to Japan on a government scholarship, coming as a research student first does not seem really necessary, unless you really need some time to prepare for the entrance exam to your desired masters program. Keep in mind however that without a scholarship, enrolling as a research student does cost money.

As I said in the linked question, all the details of what you will do as a research student are to be agreed with your supervisor. Of course, no reasonable supervisor would expect you to be able to do original research straight out of undergrad...