Familiarizing oneself with state-of-the-art at the beginning of a PhD

Question

From my research internship experiences (my previous University didn't really focus on research much) and what my current advisers told me, a general flow of a PhD is like any other long(er)-time project:

• familiarize oneself with state-of-the-art on the subject

• generate your own ideas (by trial-and-error) and integrate with current approaches

  (with this phase becoming a lot fuzzier the more advanced your "project" is)

• write it up for the world to know.

As a fresh PhD student, I'm currently in the middle of familiarizing myself with the state of the art, following the advice of many older students ("be a brave soldier in the beginning and do and read everything your advisers throw/send/e-mail your way"). And I do understand the importance of it (in fact, more often than not, I love it). But, it does give one an impression of self-uselessness sometimes (I have a talk with myself every few weeks or so to remind myself of my motivation and resolve the "uselessness" issue).

So, my question is: Typically, how much time would a fresh PhD student spend on going through state-of-the-art at the beginning of his/her PhD?

And some sub-questions:

• is it expected/typical to produce some kind of output (articles?) during this period?
• what kind of output is expected at the end of this period?
• what would be some indicators that this period is ending which a student himself can notice

---

In the end, just to provide some context: I'm doing a PhD in Europe, and we have a limit of 3 years for a PhD (sometimes extended for up to 6 months) and I'm studying Computer Science.

Answer 1

Your question likely indicates that you need to work on your communication with your advisor. Hasn't he told you what is expect and how you are progressing?

I will try and keep this answer focused on the question, but I apologize if it strays.

Some indicators that your are becoming familiar with the field:

• When you talk to your supervisor you are familiar with some of the references mention and names start to mean something to you. Better familiarity is when this holds when you talk to colleagues and go to talks and seminars
• When you have read/glanced at most of the references in articles that you read. Better is when the most exciting thing about reading new literature is finding a reference to something you didn't know about

As far as output, ideally during the course of your dissertation you should become familiar enough with the relevant literature to write a literature review article. You should feel like you could write a review at the end of the familiarization stage. Actually writing a full review is probably a bit premature, since you want the review to tie in with your eventual dissertation. I would suggest that a useful output is a dissertation proposal with a strong literature review based motivation. This is not a publishable output, but it is tangible.

Answer 2

I don't think there is a clear transition between “familiarizing oneself with the subject” and “being an expert in this topic, and generate new ideas and approaches”. The transition is gradual, and it's called “the PhD”.

However, there are ways to quantify this evolution. For example, ask yourself: in a discussion with your advisor and a few other experts on a topic related to your PhD, how able are you to make useful comments and suggestions? How often do you come to your advisor saying “I have tried to do X because I read about it and I think it can apply to my issue”?

Regarding the written “output” of the beginning of a PhD, it pretty much depends on you and your advisor. Mostly, the output is knowledge in your mind, but it can also be useful to make notes (both for yourself and your advisor) on each subtopic you discover. For myself, as an advisor, I ask students to contribute bibliographic notes (written or orally) every now and then, for me and other students of the group who work on related topics.

Answer 3

I'm hesitant to make broad generalizations across disciplines, but I've seen this pattern a few times in a number of fields, so for what it's worth:

• You can expend to spend the better part of a year (or more) familiarizing yourself with the field itself.

• You can expect to spend another better part of a year (or more) familiarizing yourself with the state-of-the-art in that field.

• You can expect to spend the rest of your life staying up to date on what's being done.

• While you're doing all this, you will also be implementing some of the research you're reading about, doing coursework, teaching, grading, writing grants, and doing actual research. This will take time away from simply reading up on stuff.

Answer 4

Initially, you should really just seek to understand. Thousands (or tens of thousands, or hundreds of thousands) of brilliant people have been working on problems for many years. If you're going to contribute, hadn't you better understand what some of them did, and why? So read, think, ask questions. Your initial goal is not to produce anything except familiarity with the core concepts and techniques. Or, to put it another way, you should be producing ideas and questions, hopefully to talk about with another lab member or your mentor. If you skip this phase, you'll either be slavishly following your mentor's directions without understanding why; duplicating work that's already well-established, probably (though not necessarily!) in an inferior way; or working on something that is fun for you perhaps, but not particularly relevant for anything. (Many Ph.D. students, in my experience, do end up falling afoul of one or more of these.)

Now, you ought not expect at the end of this period that you'll have as keen a grasp of the field and directions as your mentor; the point is to start getting the perspective you need to understand why, for instance, your mentor suggests using a bayesian analysis of job-completion times to help with load balancing.

At this point--unless your mentor is highly concerned about this--I wouldn't fret too much about producing papers. Get yourself in a position where you can produce good papers. Once you're there, then work out the next part of the plan (you'll very likely find that it depends heavily on what you're planning to do, which you won't know until you understand the literature and appropriate techniques).

Answer 5

My advice: don't read everything otherwise you will never finish! Finishing your PhD should be your goal.

Become familiar with the dominant themes in your field and continually think how your PhD fits in the scheme of knowledge.

Focus on what is relevant to your research.

Answer 6

The problem with the general outline you describe is that it's top down, and doesn't reflect the bottom up nature of doing research. In other words, while it's useful to familiar with the state of the art, you shouldn't necessarily start by assimilating the state of the art in the field. That's too overwhelming !

The best approach early on is to start small, and work on something concrete. Doing a Ph.D is really like doing an apprenticeship to hone the craft of doing research. Starting with assimilating the state of the art is akin to reading lots of books on music theory before touching a piano.

So start with a small problem. Try different approaches. Talk to lots of people. Read papers that might have ideas to help you with the problem. Talk to lots of people again. Try more ideas. Solve a piece of the problem. Discover it's been solved before ! Realize that you've been able to recreate someone else's original research. Realize that this is a good thing.

There will come a time when you suddenly realize that you're familiar with most of the state of the art. And that you've been adding to it.

And then it's time to graduate :).