15

An incoming graduate student is typically required to help out as a TA, take courses both related to their research and required by the program, as well as to start reading up on the research topics in his field. Unless he has worked in the same research area prior to joining grad school, it is unlikely he would get any publishable results very soon after joining grad school.

To avoid being depressed by this apparent lack of results in the first few semesters, it would be helpful to know what the adviser/admissions committee expects out of him in that time, and by when he would be expected to start having publishable results?

I'm interested in the answers related to Theoretical CS, but as always, I believe it would be applicable to any grad student as well - so it should not be specific to this field.

9

This wil vary significantly according to advisor, but I'd say a typical plan is:

  • Years 1-2: coursework, begin research
  • Summer of year 2: Small research publications (in my field, 2-4 page conference proceedings, small steps)
  • Year 3: Get some real research done, more small papers
  • Year 4+: ~1/2 paper a year, ish

However, the variance may be too great for this to be meaningful. I have a friend who published 17 papers during his 6 year graduate student tenure, and I have a number of friends who published zero peer-reviewed papers during my grad school tenure. Take the numbers with a grain of salt.

  • Ummm...I understand that unless you have some published results, the grad committee/adviser would not be letting you graduate with a Ph.D (understandably!) - so, isn't no publications a deal-breaker for getting a Ph.D? – TCSGrad Feb 19 '12 at 19:01
  • @shan23 - Actually, no. In my program, a publication was not a pre-requisite for graduation. (Oddly enough, many other students in my lab were from another graduate program in the same university, where a publication was required... confusing indeed!) – eykanal Feb 19 '12 at 19:09
  • @shan23: As always there's wide variation by field/country/institution, but in many places a PhD student isn't expected to have any publications prior to the thesis itself, and since it can easily take 2+ years for a paper to go from submission to publication, there's no expectation that any of the thesis has actually been published by the time of graduation. – Henry Feb 19 '12 at 20:10
  • 4
    Most (top-20 US) computer science PhD programs have a de facto publication requirement for PhD students. In my group (in theoretical computer science), students don't even propose a thesis until they've published at least two papers on the topic. – JeffE Feb 20 '12 at 9:41
  • 2
    +1 for "It depends" and +1 more for "the variance may be too great for this to be meaningful". – JeffE Feb 20 '12 at 9:41
5

In France, the PhD is funded for 3 years generally. In our university (Paris-Sud university), PhD students meet a committee composed of the adviser, the head of the lab and the head of the doctoral school each year. In TCS, if you have not published (or have a paper close to be published - that is submitted) during the second year; this yields a big red flag. So, it means that it is expected that some results are ready for publication during the third semester of the thesis (even in a small workshop).

Edit : PhD students are doing a 6 months internship before entering the PhD, with their PhD advisor mostly, so in fact the research is done on 3,5 years, and often 4 without much troubles.

  • Do posters count as published results? Also, some (or many, depending on the field) workshops do not publish proceedings - would that still be considered as a publication? – TCSGrad Feb 19 '12 at 19:04
  • 1
    publication = serious proceedings and a serious PC. For instance, a workshop as SWAT will probably be considered OK. We do not count posters as publication except in the case of conferences where all papers have the same number of pages in the proc., but some are presented as posters, other as talks. – Sylvain Peyronnet Feb 19 '12 at 19:09
  • 1
    The French (and European systems generally) assume a master's already in hand; therefore, French PhD students enter around Year 3 of the outline provided by @shan23 in his question. – aeismail Feb 19 '12 at 20:53
  • @aeismail This is indeed an important thing to know! – Sylvain Peyronnet Feb 19 '12 at 20:55
  • 1
    To add to Sylvain's comment: Very few theoretical computer science conferences have poster sessions, so that's not a real option anyway. – JeffE Feb 20 '12 at 9:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.