Flight, Optimization, and Wind Laboratory
© 2024. All rights reserved.
Flight, Optimization, and Wind Laboratory
© 2024. All rights reserved.
This document provides an overview of your graduate experience following by milestones and advice.
Unlike course-based degrees (undergraduate, many masters programs) a PhD degree is primarily based on research and so timing is much more variable. For someone coming straight out of an undergraduate degree, 5 or more years is typical (or 3 or more years for someone coming out of a masters program). The below timeline is notional as individual circumstances vary widely. It also is based on someone coming straight out of an undergraduate degree as that is our most common scenario.
Be aware that this timeline varies from typical department norms of two classes (I’d recommend almost never taking two classes). For students coming straight out of an undergraduate program, I recommend front-loading courses for the following reasons. (1) Students often struggle to make the type of progress they planned for during semesters where they are trying to balance courses and research. This is especially true during that first year and half when the demands of TAing, qualifying exams, and graduate-level courses take much time and energy. (2) Students usually lack the background and experience to be productive until they progress further in their coursework. After that first-year and a half I recommend taking one (or zero) classes each semester until you’ve completed your classwork.
Also note that I’ve assumed a very linear and regular progression of conference and journal papers, but for many this progression is highly nonlinear with the bulk of the papers coming towards the end.
Year 0 (prior to graduate school starting)
Year 1 (TA Fellowship)
Year 2
Year 3
Year 4
Year 5
Every semester our lab engages in goal setting, reporting, and 1-on-1 reflection/planning meetings.
Each semester you will propose a deliverable, we will meet to discuss it, and together determine an agreed upon outcome. You will certainly have other goals, but for the purpose of this meeting we will only discuss one key deliverable. Your graduate student progress, via the BYU Graduate Studies Office, will be tied to satisfactory completion of this deliverable.
You should engage in regular planning and goal setting as you work towards your key deliverable.
At the beginning of the semester, and midway through the semester we will have reflective meetings, focused on high-level plans, goals, and ideas. We should also meet frequently through the semester for more technical discussions. I have a youcanbookme link that allows you to schedule those meetings whenever you wish.
This is less a report, but more a summary of what you accomplished that semester. A semester report should only require a minimal amount of effort, just assembling what you’ve already done. The summary should focus primarily on the key deliverable, but will also include things like: a document of derived theory, a literature review, a defended prospectus, passed qualifying exams, a submitted paper or revisions to a paper, a submitted abstract, a presentation, a post on social media, code that was developed or documented or registered, mentoring, TA’ing, etc. Include everything relevant.
We post our reports on Slack at the end of each semester. It is important to regularly look back at what we accomplish and to share that. Doing so helps us celebrate the wins, identify opportunities for improvements, and calibrate our ability to plan going forward. Not everything you report needs feedback, but for those things you want a close look at, be sure to note that.
This page overviews some milestones you will need to go through as a graduate student. I do not provide details on the logistics, those you can find in the department graduate handbook, rather I am imparting some advice I frequently repeat to new graduate students. These are my own views, not all faculty members will view the purposes of these milestones the same way.
My main advice is to choose courses you are interested in. You will learn more and have a more satisfying experiencing if you follow your interests. You may even find some ideas in an unrelated field that help you with your research. Of course, you will also need to take classes that are directly related to your research, but presumably you are interested in those anyway (otherwise you picked the wrong research topic). See Learning > BYU Courses for some suggestions.
See above on timing and number of class. Take your classes seriously, but remember that research is always a higher priority over classes. It can be too easy to linger in the relative safety of classes where there are often known answers and more immediate feedback from homework scores, but you have to prioritize the more important but much harder open-ended questions with no known solution and delayed feedback from research. This can be a hard transition for undergraduate students.
A PhD student might take one or zero classes some semesters to focus on research, or participate in an internship. There is more flexibility in your time (research will be the bottleneck) so you should rarely take 2 classes at a time (MS students need to take 2 each time).
If you are a PhD student you will almost certainly find a few interesting and useful classes that won’t count towards your degree (e.g., a 400-level CS class, a business class). You might consider taking them anyway (within reason). Although if they aren’t relevant to your research, you might save them until after you complete your program of study. If you are a MS student, unfortunately there just isn’t time, but fortunately one or two 400-level technical classes can count for you.
I’d also recommend taking advantage of the unique opportunity you have to take fun 0.5 credit classes like tennis, bowling, martial arts, etc. Good for the body and mind. (again, within reason)
A typical course registration looks like the following:
Taking 1 credit every semester/term should put you on track to complete the necessary thesis/dissertation credits. Some adaptation will be necessary if gone for a summer internship or if you are an international student (don’t register during spring/summer)
Your research will have some aspects that are outside your advisor’s areas of expertise. The purpose of a committee is to provide additional expertise in those areas and to provide broader feedback and direction within the same field. Select other faculty that would be helpful to you. You don’t need to meet with them regularly to report or anything like that. They are a resource to help you, so meet with them when they can help you with something. It’s ok if you don’t know them. As long as your topic crosses their expertise in some way, they will be happy to help. All of us serve on a wide variety of thesis/dissertation committees, in most cases with students that we did not know previously.
When you attend a conference please write a short summary (just a paragraph) for each talk you attend. This will help you learn more deeply (not just listen and leave) and by sharing your notes with the group after the conference you will be helping others find some relevant papers.
This is your main marker of progress. See the section on publishing. Most of your journal papers will be regular scientific articles, but some alternatives may be appropriate depending on your situation. Some alternatives are listed below. (1) Review papers. These are kind of like massive literature reviews (citations are well into the 100s), but more than just synthesizing the literature they require expertise and experience to provide over-arching insight and identify future directions. These types of papers are often well cited and impactful if an appropriate subject is identified. (2) Engineering note. This type has various names but is a publication that contains a small but impactful new development such that its scope requires only a short paper (typically about 1/4 the length of a regular paper). (3) Software paper. This is not necessarily a widely accepted form yet, but can be impactful nonetheless. Journal of Open Source Software is an example. The paper itself is very brief (just a few pages summary). The peer review is a similar process but the review is focused on the software (see criteria for JOSS here). This can be an appropriate option if you have a well-developed software package. (4) Educational paper. There are also few venues for this less conventional option, but Structural and Multidisciplinary Optimization is a good one. These papers focus on tutorials and detailed descriptions of methodologies that would be useful in an educational context, but are not already readily available in textbooks or other sources.
If you take your preparation seriously, this can be a transformative part of your education. A PhD student needs to demonstrate impactful and original research (primarily demonstrated through your dissertation), and needs to demonstrate a broad understand of fundamentals (primarily demonstrated through qualifying exams and coursework). The goal is not to pass the exams. The goal is to develop a solid understanding of fundamentals and develop experience and confidence in self-directed learning. You should be so well prepared that the exams are a mere formality. You will strengthen your ability to learn new things independently—an essential skill in an ever changing world. You will also find great satisfaction in your deepened level of understanding and greater confidence to move forward into new areas of learning. There will likely never be another opportunity in your life to dedicate your time and focus so completely to subject mastery in this way, so although it is hard be sure to take advantage of it!
Utilizing a TA Fellowship is highly effective preparation as teaching others is perhaps the best way to really learn the material. Typically, one would work as a TA in Fall and Winter, teaching in the areas of their two main exam topics, then spend a portion of the summer preparing for the math exam and reviewing all topics. You should then be well prepared to take the exams in the Fall (some who are well prepared elect to take the exams that first Winter to get them out of the way earlier).
I strongly recommend studying with a group. It can be helpful to take turns working problems out on the board in front of the group. A group provides feedback (essential so that you don’t unknowingly learn something incorrectly!), encouragement, and can help you gather a broader set of problems to test your understanding. During the semester of quals, if additional preparation is still needed, you might consider taking a lighter course load.
A prospectus is essentially a short proposal, describing the plan for your PhD research. You need to start your prospectus early, ideally a year in advance of the deadline. Roughly speaking you will spend a semester reviewing the literature (keep detailed notes that will be helpful in writing later), a semester iterating with me and others on your ideas, and a semester writing and defending (although these are not really discrete tasks but rather will blend together). Start with just four sentences describing your four paper topics that we can discuss and iterate on. Next, add a little more detail to each, still probably at the level of bullet points. Once we’ve iterated to the point of the solid plan then you will turn your attention to the writing part.
Writing a proposal is different from writing a scientific paper. A proposal is intended to be a persuasive argument (rather than a presentation of results and an illumination of their implications). You need to persuade the reader that the problems you are trying to solve are important, that you’ve carefully thought through how you will accomplish the work, and that you have the resources and training you need to succeed. You must have a solid plan, but don’t worry that it must be cast in stone. Like all research projects you should expect that there will be deviations. The purpose is to help you plan ahead and think about the bigger picture. Rather than think of this as a one-time document, you should think of it as a big step in what should be periodic reflection and replanning to align with what is most impactful and essential.
Make sure you explicitly answer the following questions in any proposal, as these are the criteria an evaluator will use to judge its merit. These questions come from George H. Heilmeier, a former DARPA director:
Other advice related to addressing these questions in your prospectus:
For a PhD Dissertation your focus should be on journal papers. Our guideline is that it takes four journal papers in a specific area to develop the requisite expertise. These four papers form the core of the dissertation, and often are inserted directly as four chapters (with annotations for where they were published). Alternatively, some people like to pull out common intro/conclusion content from the four papers to create a more cohesive narrative across the dissertation.
Surrounding these papers you will add a few new sections: 1) a unifying introduction, 2) a chapter or two after the introduction that describes fundamentals and background necessary for the rest of the dissertation, 3) a unifying conclusion, and 4) appendices that provide important details like lengthy derivations, model parameters, data sets, etc.
These new sections in the dissertation are different from their journal counterparts. For (1), the introduction, a journal paper is necessarily brief, because the audience is subject matter experts in the journal’s topic. However, for a dissertation the audience is general and you should assume they come in knowing nothing about your topic. Motivations, themes, key concepts, etc., need to be described.
For (2), the background section, a journal is space limited so for well-known methods we don’t reiterate all the details, but rather provide just enough info that one could still reasonably reproduce. In a dissertation one should generally goes into full and complete detail. A good rule of thumb is that the dissertation should stand alone (i.e., one shouldn’t have to refer to other references to figure out derivations or reproduce your results). You’ll still have references of course (lots), but anything key from those references should be in the document. This means that some of the background you add isn’t necessarily novel, but like a textbook, is repackaged in a way that adds clarity and connection to your topic. This is probably the most important contribution of your dissertation beyond the papers themselves. The main chapters are all available in journal papers elsewhere (and most people will go to the journal versions), but your dissertation contains excellent detail to help someone get up to speed in your area. This is something you can and should be writing throughout your years as a graduate student. It represents your unique acquired body of knowledge.
For (3), the conclusion, it should be comprehensive across all the papers, and should clearly highlight your unique contributions to the field.
You should have all necessary detail to reproduce your work. Some of that info is too lengthy or not appropriate in the main text, and so should be placed in an appendix. For big data sets, or code, hyperlinked archived data sources are usually preferable to typed out versions.
An MS Thesis is similar to described above, but with an expectation of one journal paper.
For all students you should write the thesis as you go. In other words, write out all the details of your derivations, etc. It is easier to distill from all the detail down to a more concise journal paper, rather than trying to go the other way (especially when significant time may have lapsed from when you performed the work).
Most graduate students underestimate the time required to conduct a proper job search. For a PhD student a job search might span two semesters (less for an MS student). Visits typically include presenting at a seminar and interviewing with people across multiple levels for a full day or two. Preparation is needed to deliver an effective presentation, to learn what you can about the organization/people beforehand, and to prepare relevant questions. The preparation, travel, and interviewing can take a better part of a week, and you’ll likely do this more than once.
You’ll want to apply to multiple places. Even if a certain group is interested in you, circumstances can change quickly. While leveraging your network is always helpful, this is especially important at the PhD level. In fact many PhD-level positions aren’t available from an internet search and are created specifically when the right person comes along. Receiving your first job offer is exciting, but don’t feel pressured to take it just because it is there. More than the name of the company/organization you will want to make sure there is a strong fit with the specific team you will be working with, including your direct supervisor(s). Those relationships will have a bigger impact on your working environment than the overall company/organization culture.
Knowledge/code transition is important. You’ll need to plan adequate time to document, train other students, integrate code with other lab code so that it can be used by others. As a PhD student you’re expected to leave behind a well-documented useful code package(s) that reflects the understanding and skill you’ve developed. If you’ve been following good practice, usually a couple weeks post dissertation is sufficient to finish documenting, testing, registering, etc. More time may be needed if your code is less well organized.
Depending on your funding source we may have some outstanding sponsor deliverables. Make sure to ask about any expectations in this area so that you can plan your time accordingly.
In short, don’t plan your graduation timing with only the defense/dissertation in mind. A robust job search, code packaging, and funding obligations are some important elements that we sometimes underestimate or don’t account for.
You need to submit your thesis/dissertation to your committee well before the defense (I believe it is a minimum of two weeks, but see the graduate handbook), and it needs to come to me for review 2 weeks before that. If you find things you need to fix after this date, do not send updates to your committee. Some of them will have started reading, and if only some members of your committee read the updates it will cause confusion and give the impression that your work is not ready. You will have an opportunity to make changes afterwords. Like journal reviews there will almost certainly be comments and suggestions for improvement.
I suggest preparing a handout with your slides printed on them (4 per page, double sided), one for each member of your committee. This will help your committee follow along and note their questions next to the slides. Put slide numbers on the slides. The committee will want to refer back to things and having numbers will make this much easier.
The notes on presentations are applicable to your defense. The only difference is a little more emphasis on clearly highlighting your contributions.