In the Japanese graduate-school system, the lab you join decides almost everything that happens during your degree. Not the university name, not the program ranking, not the city — the specific lab, which means the specific professor, the specific group of senior students, and the specific funding situation that lab is in when you arrive. This guide is a 2027-stamped framework for evaluating a Japanese lab before you commit, with explicit signals to look for, red flags to walk away from, and a scorecard you can apply to any shortlist.
Why lab fit matters more than school name
Western applicants often think of graduate admissions like a university brand decision: get into the most prestigious place you can, then sort out the details. In Japan that logic fails. The Japanese graduate system is built around the kenkyushitsu (research lab), which is a semi-autonomous unit run by a single professor with effectively full authority over admissions, funding allocation, research direction, and your timeline to graduation. You do not apply to "the University of Tokyo Graduate School of Engineering" in any meaningful sense. You apply to one professor's lab inside that school. Once you are in, your daily reality is that lab — its people, its hours, its money, its publication culture — and the rest of the university barely touches you.
That changes the optimization. A supportive, well-funded lab at a solid mid-tier university (think Tohoku, Hokkaido, Kobe, TIT outside the absolute top group, or many strong national universities profiled in the best engineering universities in Japan beyond the Imperial Seven ) will produce better PhD outcomes than a hands-off, underfunded, or burned-out lab at a top-three name. The CV line "PhD, University of Tokyo" looks great until a hiring committee realizes you have one workshop paper and a thin recommendation letter from a professor who barely remembers you. The CV line "PhD, Tohoku University" with three first-author journal papers and a personalized letter from a professor who co-authored every one of them wins almost every comparison.
So the right question is not "which Japanese university should I aim for?" but "which lab is the right lab, and which universities does that imply?" To do that well, you need a structured evaluation. The next section is that structure.
The seven dimensions of lab fit
Every Japanese lab worth considering can be evaluated on seven dimensions. These are independent — a lab can score high on one and zero on another — and the goal of your shortlist process is to find labs that score acceptably on all seven, not labs that score perfectly on one.
1. Research direction
Does the lab actually do work in the area you want to do? Be specific. "Machine learning" is not a research direction; "self-supervised representation learning for medical imaging" is. Read the last two years of the lab's papers. If you cannot, after reading three abstracts, articulate the lab's research question in one sentence, the lab is either too broad (a service group running other people's projects) or too far from your interest. Both are bad.
2. English-friendliness
Can you actually function in this lab without near-native Japanese? Concrete signals: lab meetings in English (or bilingual), an English version of the lab page that is genuinely maintained and not a Google-translate stub, recent publications with international co-authors, and an explicit statement somewhere on the lab or department site about international student support. If you are arriving without strong Japanese, this dimension is not optional. See accepted into a Japanese lab without Japanese for the realistic picture.
3. Lab culture and work hours
How does this lab actually run? Is there a core-time expectation (e.g., everyone in the lab from 10am to 7pm)? Are weekends free or implicit work time? Do students take vacation? The lab page never tells you the truth on this, and the professor is the worst person to ask because they answer aspirationally. Ask current students. A healthy lab has students who can describe their week without flinching and who take real breaks. An unhealthy one has students who answer "we work as much as needed" and look tired in the photos.
4. Accepted-students history (especially international)
Look at the members page over time, ideally with the help of the Internet Archive Wayback Machine snapshots of the lab page from three to five years ago. A lab that has had at least one international Master's or PhD student per year for the past three cycles is a working pipeline. A lab that took one foreign student in 2019 and zero since is not. A lab that is one hundred percent Japanese for the past decade — even if the program brochure markets to international applicants — will be a hard place to land and harder to thrive in.
5. Recent publications and impact
What has the lab actually produced in the last two years, and where? Top-tier venues for the field, Q1 journals, conferences with selective acceptance rates — those signal a productive lab. Publications only in domestic Japanese journals, or only in lab-internal proceedings, signal a lab that has receded from the international conversation. Neither is automatically a deal-breaker, but the latter requires a very good reason for you to join.
6. Funding situation
Does the lab have active grants? In Japan, the dominant funding source is JSPS KAKENHI, which runs in three- to five-year cycles. A KAKENHI Kiban-A or Kiban-S grant starting in 2025 or 2026 means the lab is funded through your degree. A lab whose last visible grant ended in 2023 is in a precarious position — student travel, equipment, and PhD funding all depend on grant continuity. KAKENHI grant histories are public on the KAKEN database (search the professor's name).
7. Post-graduation outcomes
Where do this lab's graduates end up? Postdocs at international universities? Industry research roles at Sony, Hitachi, Toyota, NTT, or international tech companies? Tenure track in Japan? An alumni page that lists clear destinations is the strongest signal you can find. A lab whose last three PhD graduates went into roles unconnected to research is not the lab to do a research PhD with.
Reading lab pages critically
Japanese lab pages range from minimal Japanese-only single-page sites to elaborate bilingual portals with research videos and student blogs. Most fall somewhere in the middle. Regardless of presentation quality, the same five sections carry the information you need: the members or people page, the publications or research output page, the research or projects overview, the news or updates feed, and (for the better-organized labs) an alumni or past members page.
Read in this order. Members first, because it answers the international-friendliness and team-size questions in thirty seconds. Publications second, because the most recent five papers tell you what the lab actually does today, regardless of what the research overview claims it does. Research overview third, for context. News fourth, for trajectory — recent grant awards, conference best-paper awards, faculty moves are all signal-rich. Alumni last, but read carefully; this page is the closest the lab will come to telling you what your career outcome will be.
A lab page that is missing one of these sections is not automatically disqualifying — many strong labs have weak websites — but it does mean you have to do more detective work elsewhere. For more on how Japanese labs are organized internally, see inside the Japanese lab system .
Reading the publication record
The lab's own publications page is curated and incomplete. Always cross-check against external sources. The three most useful are:
- OpenAlex (openalex.org) — free, comprehensive, lets you filter a researcher's papers by year, venue, and citation count. Best for a quick five-year productivity scan.
- Google Scholar — most professors have a profile. Citation counts matter less than the trajectory of recent papers. Ignore the h-index in isolation; look at "papers in the last three years" instead.
- The lab's own publication page — gives you author order, which the external databases sometimes miss. In Japanese labs, last-author or corresponding-author position usually means the professor; first author is usually the student doing the work. A pattern of student-first / professor-last is the healthy default.
Good signals from a publication scan:
- Two to six peer-reviewed papers per year over the last three years, in venues that match the field's reputation tier.
- Students appearing as first authors regularly. If every paper from the lab in the last three years has the professor as first author and "et al." for everyone else, the lab is not training students to lead projects.
- Recent papers in international venues (NeurIPS, ICML, CVPR, ACL, Nature family, PNAS, IEEE flagship conferences, top ACS / RSC journals depending on field) rather than only in domestic Japanese venues.
- Co-authorship with other groups, including international ones — signals the lab is plugged into the global conversation rather than self-isolated.
Bad signals: the lab's most recent paper is from 2022 or earlier (productivity has collapsed); every paper is single-author by the professor (the professor is not mentoring); the publication record is dominated by Japanese-language proceedings of regional conferences (the lab has retreated from international visibility); first authors are almost always the same one or two senior students with nothing from the rest of the team (the lab has hierarchy problems).
The "accepts international students" question
This is the single most important practical question for non-Japanese applicants, and it is the one professors are least likely to answer honestly upfront. They tend to say "yes, we welcome international students" because it is the polite default, even if the lab in practice has not had a foreign student in eight years. You have to verify with external evidence.
Concrete indicators that a lab actually accepts and supports international students:
- At least one current international Master's or PhD student listed on the members page, not in a "visiting student" or "intern" category.
- A visible history of international graduates in the alumni page (or in old Wayback snapshots of the members page).
- An English version of the lab website that is genuinely current — recent news items in English, not just an English version of the homepage that has not been touched since 2019.
- Lab seminars or research meetings explicitly described as conducted in English (or bilingually). Some labs note this on the seminars page.
- Recent papers with international co-authors, particularly with the professor as last author and an international student as first.
- Faculty profile mentions teaching English-medium courses, or the department belongs to a recognized English-medium graduate program — see English-taught Master's in Japan 2027 for which programs are credible.
If a lab has none of these signals, the polite "we welcome international applications" line on the website is marketing copy. Move on regardless of how interesting the research is — being the first international student in a lab that has never had one is a research-killing experiment you do not want to run during your PhD.
Evaluating the professor
Once a lab passes the structural filters, the evaluation narrows to the professor. Four sub-dimensions matter.
Career trajectory
Is the professor early-career, mid-career, or senior? Each has trade-offs. Early-career (assistant or associate professor, less than ten years in faculty) — usually highly motivated, hands-on with students, smaller lab, but less established funding and smaller network. Mid-career (full professor, ten to twenty years in) — typically the best balance of mentoring engagement, funding stability, and network. Senior (professor with thirty years tenure, near retirement) — strong reputation but often hands-off, lab may be in wind-down mode, retirement risk during your PhD is real. For 2027 PhD start, mid-career is the safest default.
Supervisor reputation in the field
A professor's recommendation letter and network become your career capital after you graduate. A professor who is invited to speak at top international conferences, who organizes workshops, who serves on program committees of the field's best venues, has a network you will benefit from. Look at their CV (often on the lab page or researchmap.jp): plenary talks, editorial board memberships, organized workshops are all good signals. See what Japanese professors look for in international applicants to understand the other side of this match — what makes you attractive to those professors.
Recent students' outcomes
Where did the last three to five PhD students from this lab end up? An alumni page makes this easy. If the lab does not publish one, ask the professor directly during first-contact correspondence. A professor who can name where their last five graduates went, and whose answer includes a mix of academia (postdocs, faculty positions) and industry research, has a healthy track record. A professor who answers vaguely or whose graduates went into unrelated roles has a problem.
Hands-on vs hands-off style
How often does this professor meet with students? Do they read drafts? Do they co-author? In Japan, hands-on supervisors meet weekly or biweekly with each student individually, plus a weekly group meeting; they read full paper drafts; their name appears as last author on student papers. Hands-off supervisors meet only on request, delegate mentorship to senior students, and may not co-author every paper. For international students arriving in their first year of Japan-based research, hands-on is dramatically better. The hands-off model rewards confident domestic students with formed research directions; it punishes incoming international students figuring out a new system.
Kohai and senpai: the lab dynamics that decide your daily life
Japanese labs run on the senpai (senior) / kohai (junior) relationship in parallel to the professor relationship. Senior students teach juniors how the lab equipment works, debug experiments, explain unwritten rules ("we always clean the lab on Friday afternoon"), and translate the bureaucratic logistics of being a graduate student in Japan. In return, when you become a senpai, you do the same for incoming juniors. This system is informal, unpaid, and absolutely central to how labs function.
A healthy senpai/kohai dynamic looks like: senior students who are visibly available to juniors, group lunches or lab dinners that mix years, mentorship pairings that are explicit (some labs formally assign each new student a senpai mentor for the first six months). An unhealthy one looks like: senior students who eat alone, no cross-year interaction visible during a lab visit, juniors who say "I usually ask the professor directly" because senpai mentorship has broken down. The latter pattern is correlated with high attrition.
During a lab visit, this is the single most important thing to observe. The professor will be polite to you; current senior students will tell you the truth. Watch how the students interact with each other when the professor is not in the room.
Lab visit tactics
A lab visit — even a virtual one — turns abstract evaluation into concrete signal. If you can manage a campus visit before applying, do it. If not, request a video tour and a chance to talk to current students. Both are common requests in 2027 and most receptive professors will accommodate.
What to do during an in-person visit:
- Spend at least half a day in the lab, not just an hour with the professor. Sit at a desk if offered, attend a group meeting if one is happening, eat lunch with students.
- Ask current PhD students three questions: "How often do you meet one-on-one with the professor?", "Roughly what hours do most students keep?", and "What was the hardest thing about your first six months in this lab?". The third question gets the most honest answers.
- Ask one international student (if any are present) what they wish they had known before joining.
- Look at the physical lab. Tidy or chaotic? Equipment well-maintained or visibly broken? Whiteboards in active use? Students at their desks during the workday?
- After the visit, ask the professor for a fifteen-minute one-on-one without students present, and use it to ask about funding, your specific research direction fit, and the timeline to graduation.
For a virtual visit, the priorities shift: insist on at least one separate video call with current students that the professor does not attend. Without that, you will get only the polished version. Most professors will agree if you ask politely.
Red flags that should end the conversation
Some signals are serious enough that they should remove a lab from your shortlist regardless of how interesting the research is. The list below is not exhaustive but covers the patterns that recur.
- Zero international students in the past five years. The risk of being the experiment is too high.
- Single-author papers from the professor. A professor whose recent papers list only their own name is not co-authoring with students. That style does not produce trained researchers; it produces students who do administrative work for someone else's papers.
- "Core time" plus weekend expectation. Some labs explicitly post "core time 9–9, weekends as needed." If you see this on the lab page, believe it.
- Recent professor career change. A professor who moved institutions in the last twelve months is rebuilding a lab from scratch and may not have funding or equipment in place. Wait two years if you want to join their new lab.
- No visible funding. No KAKENHI grants in the kaken.nii.ac.jp database for the past three years, no industry collaboration on the lab page, no equipment grants — the lab is financially marginal and you will feel it.
- Imminent retirement. Japanese national university professors retire at sixty-five (sometimes extended to sixty-eight). If your prospective supervisor is sixty-three when you arrive, your PhD ends with an emeritus advisor and an awkward institutional transition.
- High recent attrition. If the members page shows three students listed last year and two of them are no longer on the current page, ask what happened. The answer is sometimes innocent (graduated early, internship abroad). Sometimes it is not.
- Bench-fee or hidden-cost requests in correspondence. A well-funded Japanese lab does not ask incoming students to bring their own laptop, pay for their own conference travel, or contribute to lab equipment. If those requests appear early, the funding signal is bad.
The trade-off: prestige vs support
The single hardest decision most international applicants face is choosing between a prestigious-but-hands-off lab at a top-five Japanese university and a supportive-but-mid-tier lab at a strong national university. Both are real options for serious applicants, and both come with predictable trade-offs.
| Dimension | Prestigious + hands-off (top-5) | Supportive + mid-tier (solid national) |
|---|---|---|
| CV brand value | High | Medium |
| Professor mentorship | Low — meetings monthly or rarer | High — weekly meetings, draft reviews |
| First-author publications | Depends entirely on you | Higher and more reliable |
| Recommendation letter strength | Often thin and template-feeling | Personalized and substantive |
| Network exposure | Strong — lab attracts visiting researchers | Moderate — depends on professor's invitations |
| Time-to-degree risk | Higher — easy to drift | Lower — supervisor pushes timeline |
| Best fit for | Self-directed students with prior research | First-time PhD, especially international |
For most international applicants without prior independent research experience, the supportive mid-tier lab is the better choice. The top-name lab works only if you arrive already capable of running a research project alone, with a formed direction and the discipline to push it without external pressure. If you are not certain you are that person, you are not. Pick the supportive lab. The PhD in Japan: funding, duration, and English-taught options guide covers the broader funding picture that affects this trade-off.
The decision framework: a scorecard
Apply this scorecard to every lab on your shortlist before sending the first email. Each dimension scores from 0 to 3. Total scores below 14 should be removed. Scores between 14 and 17 are workable. Scores 18 and above are the labs to prioritize.
| Dimension | 0 — disqualifying | 1 — weak | 2 — solid | 3 — excellent |
|---|---|---|---|---|
| Research direction match | Different field | Adjacent | Clearly overlapping | Direct fit on a specific paper |
| English-friendliness | No English at all | English page, no recent updates | Maintained English site, some bilingual seminars | English-medium lab, international co-authors |
| International student history | None in 5 years | One in last 5 years | One per year for 3 years | Continuous pipeline, alumni in good roles |
| Recent publications | Nothing since 2022 | Domestic venues only | 2–4 international papers per year | 5+ at top venues, students as first authors |
| Funding (KAKENHI etc.) | No active grants | One small grant ending soon | Active mid-size grant | Multiple active grants, industry collaboration |
| Mentorship style | Hands-off, single-author papers | Quarterly meetings | Biweekly meetings, co-authors student papers | Weekly meetings, reads full drafts, co-authors |
| Alumni outcomes | Unknown / unrelated roles | Mixed, mostly domestic industry | Some research roles, some academia | Clear track record, named alumni in research |
Use this honestly. The temptation is to score every dimension at 2 because the lab "seems fine." Force yourself to find the evidence — alumni pages, KAKEN grants, OpenAlex publication lists — before assigning a number above 1. A lab with a 12 score that you scored as 16 because you were optimistic will become an 11 in reality once you arrive, and by then the decision is locked.
Practical workflow for building your shortlist
Putting it all together, the workflow that produces good shortlists in 2027 is:
- Read fifteen to twenty recent papers in your research area. Note all Japan-affiliated corresponding authors.
- For each candidate professor, do a 30-minute lab-page review. Score the seven dimensions on the scorecard above. Discard scores under 14.
- For surviving candidates, deepen the research: KAKEN grant history, OpenAlex five-year publication list, alumni page, Wayback snapshots of the members page. Re-score.
- Cross-check that the university has a viable application path for your situation. See MEXT University Recommendation 2027 if you are pursuing MEXT, kenkyusei vs direct Master's application if you are weighing the path, Computer Science Master's in Japan or studying AI / ML in Japan for field-specific routes.
- Send first-contact emails. See how to email a Japanese professor for the email itself, and the application timeline for Japanese graduate schools for when to send.
- For professors who reply positively, request a video call with current students and (if possible) a campus visit. Re-score after each interaction.
- Make the final decision on the basis of the scorecard plus the visit signal — not on the basis of the university name.
Your shortlist should ideally include four to seven labs after this filtering, spread across two or three universities. That gives you redundancy if some professors are not accepting students this cycle without forcing you to apply to labs you have not properly evaluated.
Practical resources to use during evaluation
The tools that actually move the evaluation forward in 2027:
- KAKEN database — search a professor's name to see their grant history. Shows funding stability.
- researchmap.jp — Japan-specific researcher profile platform; many professors maintain a profile with full publication and grant history in one place.
- OpenAlex — free publication database, better for trajectory analysis than Google Scholar's raw citation count.
- Wayback Machine — historical snapshots of lab members pages. Lets you see who has actually passed through, not just who is currently listed.
- Our internal directories — the universities directory and the professors directory — for cross-checking shortlists against published program profiles.
- For language preparation while you build your shortlist, the JLPT N3 study hub is the right level for most lab-adjacent reading; you do not need fluent Japanese to read a lab members page, but N3 makes you noticeably faster.
Bottom line
Lab fit is the dominant variable in a Japanese graduate degree. School name is a decorative add-on. Spend the time before you apply to evaluate labs across the seven dimensions in this guide, score them honestly, and let the shortlist that survives drive your university choices rather than the other way around. The applicants who thrive in Japanese graduate programs in 2027 are not the ones who optimized for the most prestigious admit; they are the ones who picked a supportive professor with active funding, a working international-student pipeline, and a publication record that proves the lab still produces research. That combination almost never lives at the place with the most famous brand name. It lives at the place where a specific mid-career professor reads every draft their students send and writes a recommendation letter five years later that gets them a postdoc. Do the evaluation now, send the emails on the timeline in the application timeline guide , build your scholarship plan from the MEXT 2027 complete guide , prepare for the recommendation conversation with recommendation letter for Japanese grad school , and treat the lab choice itself as the most important research project you will do this year.