LaTeXiT and ghostscript 9.27

update As of ghostscript version 9.50 (possibly earlier), this is no longer issue, so you can unpin and brew upgrade to your heart’s content.


A couple weeks ago, I started my day with a usual brew upgrade, then got to work on an upcoming paper deadline… only to find that LaTeXiT broke, and everything would be half-rendered: symptom

It didn’t take long to find the culprit: ghostscript was a likely candidate, so I checked what brew had upgraded, and as expected Ghostscript had been upgraded to 9.27. However, it took about an hour to figure out how to actually roll back to 9.26, as a lot of previous methods to do this in homebrew have been deprecated.

I finally managed to fix this by directly adjusting the ghostscript formula. I finally fixed it, and figured this might be useful for others. You can roll things back with:

  1. go to the homebrew-core folder: cd /usr/local/Homebrew/Library/Taps/homebrew/homebrew-core
  2. find the commit for ghostscript 9.26: git log master -- Formula/ghostscript.rb
  3. check it out: git checkout 6ec0c1a03ad789b62
  4. uninstall ghostscript brew uninstall ghostscript
  5. reinstall it making sure not to touch anything else:HOMEBREW_NO_AUTO_UPDATE=1 brew install ghostscript
  6. pin ghostscript so that it won’t be upgraded anymore: brew pin ghostscript

You can now continue your life!

In the meantime I’ve pinged Pierre Chachatelier, the maintainer of LaTeXiT, and he suggested also an alternative: install a patched version of ghostscript. I’m lazy, so I like letting hombrew do everything (also meaning, I didn’t test this), but if you need ghostscript 9.27 for whatever reason, there’s this.

JSPS fellowship, final interview

In a previous post I documented the paper application for the JSPS scholarship fellowship. A number of applicants are accepted outright after this step, however a number are required to give a final interview in the form of a 4-minute powerpoint presentation. I was in this second set, and here is my advice for this final part.


TL;DR

  • Read the JSPS ‘About us’ page. This will give you a good idea of what they are looking for.

  • Impact and Originality are two key messages that should be emphasized.

  • Independent Thought. Which parts of the project are your own ideas, from your own initiative (as opposed to your professor’s)? This is another key point that should be clear.

  • Add Japanese Subtitles. Or alternatively make the slides completely in Japanese. You’re allowed to present in English, but having Japanese on the slides will make it easier for the audience to find the thread again if they lose track for a moment.

  • Get feedback from a wide audience. The panel will be a mixed crowd (from your general field), so you need to make sure your presentation is both understandable and appealing to a wide audience. Skype with old buddies from back home, ask other PhDs at your uni, hell, go knock on the door of some professors and ask them too. Not all of them will indulge you, but the ones who do will probably give you some really good feedback.


The story

The TL;DR pretty much covers it and most of the reasoning is explained in this post, so I won’t go into too many details. Of course being able to give a solid and compelling presentation is key, but that isn’t specific to the JSPS interview. So I will focus here on the Q&A section. Most of these questions are rather standard, so you should prepare and practice them. With some practice, you can often answer questions in a way that emphasizes things that you think make you a strong candidate but aren’t directly asked for (i.e. I wanted to point out that I have had a lot of international experience, which I think is one of the things that sets me apart from the average Japanese student).

Q&A

Q1: Why do you want to study in Japan?
This is a very standard question, and a perfect excuse to talk about something more important. I quickly mentioned that I like Japanese/Asian culture, then explained the main reason was specifically to join the Ishiguro lab. I had met prof. Ishiguro a year earlier during a summer-school, we got along very well and I really liked both the research topics his lab had as well as the approach they take. Then I took the opportunity to explain that his lab was very much focused on hardware development, whereas my background is more heavy on optimization and optimal control, so this provided a lot of synergy: I would be bringing new know-how into the lab, and at the same time there was a lot of new stuff for me to learn. Lo and behold, the very next question was…

Q2: What is your contribution, and what is from the lab? Actually, this was phrased as “Well, actually you already mentioned this in the previous question, but …”. Having known that this was important to them, I managed to preemptively cover a good part of this question already! This meant that now I could go a little deeper into the details (I had some back-up slides to help illustrate) without losing the audience (with only 6 minutes Q&A, I didn’t want to spend too much time on a single question. Being able to answer concisely is important). They seemed quite impressed and very satisfied.

Q3: What do you expect to be most challenging? This is actually a very standard question that I hadn’t thought of and prepared for. Fortunately, I had been working on my topic for a while already and was able to give a good answer, but you should definitely be ready for this one.

Q4: What are your plans after you PhD? “Do you already have plans to stay with your current lab?”. Reading between the lines… will you stay in Japan? This is a question I was expecting and I was worried might count negatively, since I doubt I will stay in Japan. At the same time, I wanted to be completely honest. I answered that I plan to stay in academia, which means I will have to be flexible and go somewhere that has an interesting project and also funding, so I don’t know what I will do next. I then added (which perhaps wasn’t necessary) that this meant probably I would leave Japan, also because I think it is important to move around in order to continue learning.
Since JSPS is of course interested in fostering the Japanese scientific community, this was a risky statement… but I got the fellowship, so I think they appreciate the honesty and also know that even if you leave the country, it is still likely that you will indirectly help by being a good contact outside of Japan. Regardless, my advice is in any case to be honest and keep your integrity.

Q5: Energy is a big challenge for field robotics, how does your topic relate to this?
This was the last question! A somewhat technical question, but not in the specific details. Also relatively easy to answer, since my research is quite closely related to this problem (or rather, this problem is very related to my research :P).

I have heard from friends that they had more specific technical questions. In general I had the impression that the questions were prepared beforehand: expect that they have read your application well and know exactly what they want to double-check before recommending you. This also means that if you’ve changed your research plan from your original application (indeed, about half a year passed between the paper submission and the interview, so this is possible if not likely), you should be ready to explain what/why you changed the plan.

JSPS fellowship, paper application

I’m a grad student at this lab in Japan, and after many trials and tribulations I managed to get the JSPS DC 1 scholarship fellowship. This fellowship is rather competitive, and I felt that due to lack of English information as well as many cultural differences, the application is somewhat challenging and confusing for foreigners. This post is meant to give some suggestions and hopefully help other would-be applicants.


TL;DR

  • Read the JSPS ‘About us’ page. This will give you a good idea of what they are looking for.

  • Start Early! Making a strong application will take several iterations. The good thing is, after this you should also already have a good road-map for your PhD.

  • Get help for Japanese. I wouldn’t write the entire application in Japanese unless you can actually write it yourself. But I do think it helps a lot to have a short mini-summary in Japanese at the beginning of each section.

  • Keep it Simple! The panel of reviewers should be from your general field (unless you apply to the interdisciplinary section), but not necessarily from your specialization. Remember, make it simple and clear, but don’t dumb it down!

  • Get feedback from different people. This is the best way to find out what parts of your explanations are too field-specific, which parts are unclear and which parts are actually interesting to a wide audience.

  • What is your Story? What is the story that links your master’s thesis/project to your PhD research? In Japan it is generally assumed that they are strongly related, so think carefully about it.

  • List of Publications is really important. Not so much advice as a general warning.


The Story

Some general information

Let’s start with some ‘admin’ stuff. There are 2 tracks, DC1 and DC2, with the same application procedure and deadlines. The difference is that DC2 is for people who already in their first year of PhD studies and therefore only covers 2 years (in Japan the standard PhD is 3 years), whereas you apply for the DC1 before starting. You do have to apply through your university(the university will then submit your application to JSPS), as well as know which lab you will be in and what your project will be: you will need to describe in great detail what your research will be. This description can be in English, however the rest of the application must be done in Japanese. This means it will be very difficult if not practically impossible to apply without actually being in your lab in Japan.
Start Early! The application deadline is usually sometime in April, but you need to pre-apply through your univisty a couple months earlier, and in particular you will need plenty of time to prepare the application. I spent close to 2 months focused mostly on this; I started sneaking in sessions of programming optimization exercises as a way to let off steam, it was that bad!
This fellowship is really good for Japan: not only is the financial support good, you’ll get some funding for your research and I’ve gotten the impression that getting it actually boosts the rep both for the recipient but also for the lab and the school. Presumably because of this, Tohoku U has special sessions where you bring your first drafts to be checked by a “mock” panel of various professors to get feedback and help improve student’s chances. So check if there is similar support at your uni. Note that chances are the announcement of this support will be in Japanese(I wouldn’t have known about this if not for the Japanese PhD in my lab who also applied), and the support itself might well be in Japanese too. Use it anyway. Your reviewers will/should be a mixed panel of people from different fields (should be your general field, but not your specific field), and they will be Japanese. One of the hardest parts for me was to understand and adapt what is actually impressive (and what isn’t) in Japan, and getting real honest feedback is critical and will probably surprise you.

Research Description

The biggest part of the application will be a description of your research. This is structured as follows:

  1. Research Description so far (i.e. master thesis): 1.5 pages
  2. Research Topic (i.e. your PhD)
    • Background, challenge, approach in relation to part 1.: 0.5 page
    • Purpose and Contents of the research: 1 page
    • Originality and Impact: 0.5 page
    • Time plan: 0.5 page
    • Legal issues (if for example you conduct experiments on animals, or take surveys or whatnot…): 0.5 page

It is important at this point to understand that Japanese academic careers work very differently from the west: just as people usually join and be loyal to the same company for life (Shūshin-Koyō or ‘lifetime-employment’), students are generally not encouraged to move between labs but often join a lab very early (here at Tohoku, usually at the end of the bachelor degree) and stay with that lab throughout their undergrad, grad and post-grad studies.
This is so commonplace that it is assumed that you will do your PhD not only in the same lab you did your masters-project on, but on the same project. This is why the description of the research topic (part 2) is in relation to part 1 (the Japanese explanation actually asks how your master project led you to the new topic, etc.). This is generally not the case for foreigners (at least it wasn’t for me), as I finished my studies in Switzerland before coming to Japan. Luckily (and this is actually the reason I decided to come this lab in Japan specifically), my previous project was closely related to my new topic, so I could weave a nice story. In anycase, this is the take-home message: when explaining your master project make sure to highlight and emphasize the parts that help explain what your PhD research will be about. After all it’s 1.5 pages at your disposal to convince the audience about funding your PhD.

A very important point is the language. You are allowed to write this section in English, however I strongly advise you to get someone to help translate at least parts into Japanese. During one of the Tohoku U feedback sessions (see above), I was told “You should use more figures and less text. A reviewer who sees this much English might not bother reading it thoroughly.”. This is a hard truth; you should not assume all your reviewers are completely comfortable with English, and if it is difficult to understand your ideas they will be less convincing. In my case, in addition to ensuring the explanations were simple and well-illustrated (which is a good thing regardless of the language/circumstance), a lab-mate graciously helped me translate the entire “Personal Motivation” section, and for the technical explanations my professor translated a short “mini-abstract” paragraph for each section. I am very convinced this greatly helped my chances.

Finally, when you explain your project you are supposed to emphasize the originality in two ways: both the originality of the research with respect to the field itself, and also which parts are really your own ideas and not the brainchild of your professor/lab. JSPS is actively trying to foster independent thinkers, and as the standard in Japan is for students to continue working on the same project as they transition from undergrad to grad studies, this is something they actively check for and it should be clearly explained.

If you are collaborating with other labs (especially internationally), you probably want to mention that. This is all very much conjecture, but I have the impression that although a large part of Japanese society is still very isolated in their world-view and aren’t very interested in or aware of what happens off the island, I feel that the upper echelons have identified the importance of opening up and made it a priority. This might be a bonus point. It will probably depend a bit on who actually reads it.

List of Publications

This part was a bit of a shocker for me. In Europe (at least in Switzerland), it’s rather unusual for a master student to have any publications, and quite impressive if you have just one. This is because the education system is completely different, and you aren’t really a “graduate” until you finish your masters (bachelors degree is 3 years, and everybody does a masters).
In Japan (and from what I understand, it is similar in the US) however, the situation is different: it is quite common to enter the industry after a bachelors, which lasts 4 years.
So when I went to the first feedback session, having listed my 1 conference proceeding + 1 conference paper, I was feeling pretty confident and even cocky. Only to have a prof frown at the list and ask “Can’t you fill out all this white-space a little more?”.
You see, what I had not considered is that once a student joins a lab in Japan, he will publish regularly in domestic conferences, which have relaxed or even no review. It’s a different system, with a focus on giving young students practice and experience, rather than original research. Indeed from what I’ve seen, it is about equally unusual for a graduating Japanese student to have a paper at an international conference. This notwithstanding, most Japanese students will have a very impressive-looking list of perhaps a dozen publications and achievements, and this is one of the most important criteria for selection. The good news is that international publications are given more points, however it’s still difficult to make up for the quantity that your competition will have.
So if you have the luxury of spending a semester or two at the lab before starting your PhD, you can make it clear to your prof that you plan on staying there, and participate in some of these domestic conferences. From what I’ve seen, it is common practice to send specific students who will continue as PhDs to more conferences, specifically for this reason.

Personal Motivation and Achievements

This section is more standard, and separated in 2 parts. The first is a free-form regular motivation letter, and in the second you are encouraged to list actual accomplishments and things you did that lend weight to what you claimed in the first part. As an example, I wrote that I believe active communication (‘science communication’) of research to people from other disciplines as well as the general public is essential. To back it up, I explained I’ve taught robot-programming workshops for elementary-grade children, and also got a poster award at a biology conference (my background is in engineering).
If you can, I advise to get a friend to translate this part into Japanese. Since it doesn’t need to be so technically precise, I think it makes more sense that this part is in Japanese (I didn’t feel comfortable not writing the technical part myself, hence in English).