Balancing Sponsored Vs. Thesis Research

One of the things that has helped me procrastinate on moving forward on my thesis has been splitting time between the research which funds my assistantship and the work that will get me graduated.

I want to be clear–the sponsorship was not at all an obstacle; it was of course essential to funding my education. However, the times that I had uncertainty about my thesis work, it was an easy fallback to switch and concentrate on sponsored work. I’m writing to hopefully help some folks avoid the same trap and not have to learn their lesson through experience.

1. If you can, use your sponsored research as your thesis topic. This is maybe the most obvious answer. As I mentioned in my previous post, I didn’t think my sponsored project was going to have enough novelty to turn into a thesis. If it had, it could have drastically simplified things; it’s worth the effort to find at least some overlap.  Don’t throw out the rest of these nuggets, though, as there is almost always a tension between the sponsor’s objectives and the student’s individual objectives.

2. Set self-deadlines in ‘rhythm’ with sponsor’s deadlines. Usually, some of the most intense times in sponsored research come in the days and weeks before a deadline. Even the most organized people tend to cram a bit more in to meet a deadline, and this means that thesis efforts are going to suffer. However, if you’ve planned carefully, you can give yourself deadlines that allow you to work intensely on one and put the other on the backburner, and vice versa.

3. Double dip. This is hand in glove with #1, but focus initial efforts on both projects to payoff for both. For example, spending the effort to learn a mathematical theorem or algorithm that applies to both projects.

4. Work in chunks. There is a measurable cost in time and productivity when you switch from one project to another.  In theory, then, it would be most efficient to work on focus on a single project until completion, then switch.  Obviously that’s not realistic; so the next best thing you can do is to force yourself to work in chunks. The size of the chunk depends upon recurring requirements; for example if you have to present updated results weekly, you’ll want to make sure you have a chunk of time (2 or 3 days?) to get those results.

5. Leave something undone.  If you are working on multiple lines of effort, you inevitably have to change your focus from one to the other every so often.  Something that’s helped me is, when you step away from working on one, leave something easy undone. When you come back to that effort, you’ll know exactly what to do next, and it will get your mind primed to think about the project.  I’ve found that if I don’t do this, it takes a long time for me to get back on track again when I switch.


How I Finished My PhD Thesis Proposal (Finally)

About a month ago, I finished my PhD thesis proposal; it was a long time in coming. I wanted to share with whomever was interested what it took for me to finally get it done, and what I would do differently if I had it to do over.

“Piled Higher and Deeper” by Jorge Cham

In the fall of 2010, I passed my PhD qualification examinations, and had been a graduate student for nearly two years at that point. I was working on obstacle avoidance for helicopters as part of my research assistantship, and I assumed that obstacle avoidance would form the basis of my thesis work, but I wasn’t quite sure how, or if the work I was doing was going to be sufficiently unique. At any rate, I wasn’t terribly concerned because I had just finished my quals, and up to that point, that was all I cared about.┬áThat was mistake number one.

LESSONS: 1) They make a big deal about qualifiers, and they are important; but the thesis is THE thing that defines a PhD. 2) Start on your thesis from the very beginning. I don’t expect that I would have made a lot of progress on my proposal prior to quals, though I certainly could have made some, and had a plan for how to proceed immediately after passing the exam.

Through 2011, I continued studying obstacle avoidance and didn’t make time for much else, research-wise, even though it was becoming clearer to me that the field was saturated. Thankfully, I was able to make progress on a few ideas that resulted in a journal paper. The good news was that I improved my system for keeping up with publications, the bad news was it was going to hard to come up with something unique for a PhD thesis. By the end of 2011, I started rooting around for another topic to focus my PhD research upon, and in early 2012, I thought up an idea that–upon review with my advisor–seemed to fit the bill well.

LESSONS: 1) Be systematic about keeping up with publications. Use RSS or automated email lists on the journal’s web page to find out about new articles. Store and index the useful articles with a program like Mendeley. 2) Reading a lot of papers in your field (and some in OTHER fields) can be exactly the kind of thing to get the creative juices flowing. 3) Completing a survey paper in your general area of interest, as I’ve recommended before, can be a big, big help.

At this point, one would think, ‘Great! You’ve got an idea, now run with it!’. The truth was, I didn’t quite know how to start or how high the bar was to submit a proposal. I spent the rest of the year working on sponsored research (collaborative autonomy) and procrastinating on thesis work because it seemed so overwhelming. However in early 2013, I saw a notification of a conference that I wanted to attend and made up my mind that I would produce a paper for that conference using my thesis topic. Having that motivation and a deadline helped me really move.

LESSONS: 1) Approach preparing your thesis proposal like an engineering problem: break it down to smaller pieces (like the literature survey, preliminary work, experimentation, and draft writing) and attack one at a time. 2) Prioritize working on toy problems that can quickly test your basic ideas, discard those that don’t work, and move on to the promising ones. Ask yourself, “What’s the simplest experiment I could do to (in)validate my hypothesis?“3) Find ways that you can publish preliminary and intermediate results to keep you moving forward.

Having done the work in MATLAB for that conference paper helped me identify the issues at hand that needed to be solved. In the fall of 2013, I worked on another sponsored project where I needed to troubleshoot code for an air transport simulator, which required a lot of effort to learn since I was not familiar with object-oriented programming (OOP). Shortly after, I started to develop from scratch a C++/OOP framework for my algorithms and experiments–first because it was faster, and second because it would need to be integrated into GUST for eventual flight testing. In April 2014 it dawned on me that, despite a number of preliminary results available, I had still not “put pen to paper” and actually written anything. So I dove in and started to write. I wrote the literature survey first, then a couple of chapters on my general approach to the problem, then got some more results and wrote about them, then finally finished drafting that sucker near the end of the summer.

“Piled Higher and Deeper” by Jorge Cham

LESSONS: 1) Working on other research projects, even unrelated to your thesis topic, can teach you important elements necessary for your own research. For me it was understanding OOP, which is all but essential for the work I’m doing. 2) Start with a shell or template of the proposal (preferably in LaTeX) and start up front by writing as you go, a paragraph a day if that’s what it takes. Don’t wait until you know exactly what you’re going to say, because it’s all going to be edited over and over again.

“Piled Higher and Deeper” by Jorge Cham

I presented that draft proposal to my advisor and research lab in Fall 2014, and got feedback that I would need a few more results to be ready to propose (sorry to be so generic!). It took me a couple of months to troubleshoot and figure it out but I got them and re-wrote the necessary sections of the thesis proposal. I contacted the committee members in January of this year and sent them my (gulp) 120-page document. I had gotten so knee-deep in my own writing that I forgot that four busy professors and one busy research engineer would not likely have time to slog through my verbosity, and one of them called me on it. He said, “I just need to know enough to understand your idea, and that there is a gap in the literature that this will uniquely fill. It could even be just 3-4 pages.” Tail between my legs, I cranked out a roughly 10-page version of the proposal in a much more digestible conference paper format to satisfy his request.

LESSONS: 1) There’s a presentation that an advisor gave his students years ago, that explained that their research would dominate their view but in reality it is just a little bump on the sphere of human knowledge. Don’t forget that–a proposal, and even a thesis, are not an ad nauseum diary of your research experience. They are communications to your committee about your work. Be concise and brief, though be ready to give more information if asked. 2) A big mental problem that I ran into (and that partly led to that monstrosity of a proposal document) was constantly wondering if I had done enough, that I was proposing enough work. Ask your advisor up front if she can provide you with 3 good examples of thesis proposals and ideally the corresponding theses. Then find out what made them good, and keep this as a measuring stick for your own. 3) Once you have your idea and have researched the field, come up with very specific thesis objectives and vet them with your advisor. It’s tempting to go big, but there is a balance between satiating your curiosity and keeping the scope manageable.

Remember that the objective of the proposal is quite simple. There are, of course, differences between universities and departments, and there are both written and unwritten norms you have to conform with, but basically, you want to:

  • show the committee you are capable of doing the work,
  • demonstrate that the proposed research can make a unique contribution to the field,
  • demonstrate you have a reasonable chance to support the hypothesis/that the proposed solution will succeed (supported by your demonstrated competence and by similar efforts in the literature),
  • set a contract for a certain amount of work,
  • and confirm that the committee has the expertise to truly assist.

In retrospect, I probably could have gotten this thing done a year or more earlier. While I’m satisfied with my experience in graduate school, I want to share some of these lessons with others who could benefit from them. Good luck.