October 16, 2003 - last edited June 18, 2011
This web page explains how our department's Ph.D. program is structured. It also gives additional details of department and university practices, beyond the formal Ph.D. requirements outlined in the course catalog. Substantive policy changes will be reflected on this page, and major ones will be announced by email.
JHU is run by its faculty, not by bureaucrats, so requirements are rarely written in stone. If you have good reasons, you can petition the full CS faculty for an exception. Such petitions should be submitted in writing to the department chair. They will be considered at the next possible faculty meeting, or forwarded to the Graduate Board in the case of a university-level requirement. Note that deadlines will only be extended in unusual circumstances, such as serious medical problems or a change of advisor.
(For policies that originate above the department level, see the Whiting School of Engineering's graduate policies and procedures and the JHU Graduate Board's forms and regulations. The most important of those policies are explained on this page as well.)
Our Ph.D. program takes about 5 years, though the duration varies from student to student. Here is the basic timetable of requirements:
| Years 1-2
("qualifying requirements") |
8 courses
2 research projects |
| Year 3 | Graduate Board Oral Exam (GBO)
First meeting with thesis committee |
| Years 4 through n-1 | Annual meetings with thesis committee |
| Year n | Departmental seminar
Thesis defense |
These requirements are discussed below. The schedule is designed to help you keep moving along toward the Ph.D. You are expected to stick to it. In particular, you must not delay the third-year requirements, even if you wish you were better prepared for them.
Note that the Ph.D. qualifying requirements are a slightly stricter version of the M.S.E. requirements. Once you complete the courses and your first qualifying project, you are therefore entitled to an M.S.E. degree from JHU (upon request). Once you complete your second qualifying project and the GBO, you become a dissertation student and will get a small increase in stipend.
If you transfer into the Ph.D. program from the master's program, then the deadlines are ordinarily set as if you had been in the Ph.D. program all along (unless your advisor recommends an extension at the time of transfer).
If you spend one or more semesters as a part-time student, then the deadlines are extended accordingly, but you must still pass the qualifying requirements and the GBO within 4 calendar years of entering the program.
As a Ph.D. student, you will get to know many of the faculty. However, you will have special relationships with the following people:
Your advisor's job is to help you become a successful member of the academic community. He or she will guide your course selections and your research, give you career advice, and tell you when you are ready to defend your thesis. Most advisors also offer funding to their students, contingent on research productivity.
A faculty advisor will be assigned to you when you are accepted into the Ph.D. program. Most students keep their advisors until they graduate, but you may change advisors at any time. Any consenting JHU CS professor may serve as your advisor, including those with secondary, joint, or research-professor appointments in the Department of Computer Science. (Column D of this table shows more precisely who may serve as a thesis advisor or co-advisor at JHU.)
Your thesis committee's main job is to decide when to accept your Ph.D. thesis. Obviously, it's wise to keep them informed and seek their counsel while you are researching and writing the thesis. You will probably also approach them for letters of recommendation as you are finishing.
You must choose your thesis committee by the end of the third year (and as soon as possible after your GBO Exam), so that you can meet with them annually. The committee is chosen in consultation with your advisor and must include:
In other words, you need at least 3 Ph.D.'s, at least 2 CS faculty members (including your advisor), and at least 1 core CS faculty member.
Your committee members must agree to serve. In unusual circumstances you may change the membership of the committee, again in consultation with your advisor.
The graduate program coordinator, Cathy Thornton, can help you with all kinds of administrative and financial matters.
The graduate program chair, currently Prof. Jason Eisner, oversees the CS graduate program and CS graduate student life generally. If you feel you need to talk to someone outside your committee, confidentially or otherwise, try the graduate program chair (or the department chair). The graduate program chair also welcomes more general questions, comments, and concerns.
You'll attend the CS department seminars throughout your degree here. These seminars are research talks by guest speakers. The usual time is 10:30 on Thursday or Friday.
Virtually every CS department in the country has a seminar series like this. That is how CS grad students and faculty keep up to date with developments in other areas of computer science. Knowing what is going on outside your narrow area of CS will enable you to teach, to draw new connections in your research, to converse knowledgeably during job interviews, to write and review grant proposals, and so forth.
To emphasize the importance of the seminar series in your education, the faculty has given it a course number. You are required to register for 600.601 every fall and 600.602 every spring. These courses are graded pass/fail based solely on seminar attendance. Grading is done on the honor system, so please be honorable -- as well as courteous to our guest speakers -- by showing up for most of the seminars (at least 2/3).
Coursework will help you educate yourself in your research area and in CS more generally. You may take or audit courses as long as you are here, including courses in other departments. A few graduate courses are offered during the summer.
In addition to the department seminar (600.601-602), you must take 8 graduate courses within your first two years. These must include at least 6 core CS courses -- two each from the Analysis, Applications, and Systems areas. The other 2 may be from CS or a related department.
Some students prefer to get the 8-course requirement out of the way in the first year. However, 4 graduate courses per semester leaves little time for research and teaching. So other students prefer to spread the coursework out over three or four semesters. Your decision will depend on your funding situation, your personal preference, and your advisor's recommendation.
Every semester, your advisor must approve and sign your course registration. This continues to be the case after the first two years.
Other "course" requirements:
Advanced students should register for "Dissertation Research" (600.801-802) in semesters where they are not taking any regular courses.
During your degree, you must register every semester for "Computer Science Seminar" (600.601-602). If this conflicts with another course, please contact Cathy Thornton for a waiver.
All doctoral students (and certain master's students) are required to take the mini-course Responsible Conduct of Research (360.625), which lasts only 8.5 hours and is offered every summer, fall, intersession, and spring. You are expected to complete the course by the end of the first year. Failure to do so before the start of the fourth semester may result in a loss of funding.
How do I know if a CS course can be used as one of the 8 graduate courses? The eligible CS courses are those courses numbered 400 and above. Note that 400-level CS courses are aimed at a mixed grad/undergrad audience, whereas 600-level CS courses are specialized grad seminars. The rarely used "Graduate Research" (600.701-702) and "Independent Study" (600.809-810) may be taken for graduate credit and a letter grade, under a faculty member's supervision. Courses do not count unless they are taken for a grade; thus you can't count pass/fail seminars (most 700-level courses) or the required department seminar series (600.601-602). (Exception: 600.464/664 does count.) Also, courses are ordinarily 3 credits; a 1-credit course counts as only 1/3 of a course.
How do I know if a non-CS course can be used as one of the 8 graduate courses? Any graduate course offered by a full-time JHU program is eligible with advisor approval. Your advisor must agree that the course is relevant to your degree -- either to computer science generally, or to your specific program of study and research. Graduate level courses are those numbered 600 and above, and in some departments (including AMS), those numbered 400 and above.
How do I know if a course can be used as one of the 6 core CS courses, specifically to fulfill an Analysis, Applications, or Systems requirement? Here is the official list of eligible CS and non-CS courses. For CS courses, the designations are also given in the course catalog. Non-CS courses are listed only if they have substantial CS content; moreover, at most 1 non-CS course from this list can be counted toward the total of 6 core courses, and your advisor must approve. The list may be extended by the faculty from time to time at student request; such changes require a faculty vote.
Do I have to do well in the courses? You need at least a C- for a course to count, and your average grade for the 8 courses must be at least B+. What you learn will also help you in your GBO Exam, your research, and your future career. But ultimately, the world will judge you on your research, not your grades.
Can I fulfill any of these requirements using graduate courses taken elsewhere? Yes, if the courses have not been counted toward an undergraduate degree. They must be of comparable difficulty and appropriate for the requirements in question (e.g., Analysis/Applications/Systems), as attested by a syllabus, problem sets, or other course materials. You may apply up to 2 appropriate non-JHU courses toward the course requirements, with advisor approval. If you are willing to forego JHU's MSE degree (typically because you already earned a master's elsewhere), then you may apply more than 2 appropriate non-JHU courses toward the Ph.D. requirements, with the approval of the graduate program chair; this may include up to 4 appropriate courses from JHU's EP programs. In this case (where you reduce your coursework below 6 courses and forego the MSE degree), you may be asked to complete the other Ph.D. requirements sooner.
The Ph.D. degree is primarily a research degree, of which coursework is merely the foundation. Our program quickly gets you involved in research. From the JHU catalog:
A student must complete two projects, each under the supervision and with the written agreement of a different faculty member in the Department of Computer Science. Upon conclusion of each project, the student must write a "Project Report" describing the project in detail. This report will be a public document and will be kept on file in the department office.
The supervising faculty member must approve the project report. Departmental approval of a given project will be determined collectively by the faculty of the Department of Computer Science following the spring semester of each academic year. A factor taken into account in the departmental review of a project is the stated willingness of the supervising faculty member to enter the initial stages of a Ph.D. research advisor/advisee relationship with the student.
Students have at most two years to satisfy the project component of the Ph.D. qualifying requirements.
This requirement is essentially a chance to try out two prospective advisors -- and for them to try you out. It is wise to establish a comfortable working relationship before you embark on a long thesis project. If you have already settled on an advisor, then the second project could be a way to develop skills in some other area of computer science.
The requirement is also an opportunity for you to write some publishable research papers. In many subfields of CS, new Ph.D.s looking for jobs are expected to have published several papers already. This is an excellent chance for you to get started.
Even if the work does not immediately lead to a published paper -- for example, it is a coding project intended to enable future research -- you must still write it up as a project report. This ensures that you get some relatively early feedback about your writing.
It is your job to find faculty members who are willing to supervise you on projects of mutual interest. Usually you should take someone's graduate course before trying to do research with him or her. The idea for a project may come from you or from the faculty member. Note that faculty members may have varying ideas about the appropriate topic, scope and duration of a project, so you should discuss this at the start. If a qualifying project builds on a course project, the work done for course credit should not be double-counted.
In your third year, you will face a committee of 5 professors who will evaluate your readiness to do Ph.D. research. This Graduate Board Oral Examination (GBO) is a University examination, required of all doctoral students at JHU. (The Graduate Board is the committee that oversees all graduate programs at Hopkins.)
(Also in your third year, you must take the in-person mini-course AS.360.625 Responsible Conduct of Research. This is under 10 hours and is offered during the summer, fall, intersession, and spring sessions.)
In our department's tradition, the center of the GBO exam is a Preliminary Research Proposal that you write and present. This does not have to be a fully developed thesis proposal (although it could be, if you have progressed quickly). It should at least motivate some interesting research problem in the context of previous work, and sketch your possible approaches to solving it. Preliminary results are helpful but are not required.
Writing the proposal should be useful for you, and it will typically develop into the thesis topic. However, it is not a commitment to a topic. Its purpose is simply to focus the GBO exam. You must distribute it to all GBO examiners at least 2 weeks before the GBO. 8-10 pages is sufficient.
The GBO exam format is closed-door and rarely exceeds 2 hours. You will begin by presenting your Preliminary Research Proposal, and then the examiners are free to ask any questions they want. The department prefers that examiners focus on your readiness to do original research in the area of the Preliminary Research Proposal. We hope that their questions will focus on the technical substance of the proposal, your ability to discuss the broad area with clarity, flexibility and maturity, and your knowledge of subjects that are likely to come into your work. However, the examiners are not bound by our requests and may assess you in any way they choose. You are therefore advised to discuss expectations with them before the exam, if they are willing.
The department will ask your advisor to suggest some reasonable examiners (see below). It is not proper for you to choose your own examiners, though your advisor may discuss with you.
The committee will consist of least 3 inside and 2 outside examiners. Your advisor counts as an inside examiner; so do all faculty with primary appointments in JHU CS. Everyone else counts as an outside examiner. Outside examiners are intended to contribute valuable perspectives, and also to ensure that the department doesn't let its standards slip. At least one of the outside examiners must be an Associate, Full, or Emeritus Professor; the most senior outside examiner will serve as chair.
Usually, examiners must be tenure-track JHU faculty. However, the Graduate Board can approve scholars from outside JHU, or research faculty at JHU. The department must petition the Graduate Board 4 weeks in advance to authorize such persons. Authorization to serve on GBO committees lasts for 5 years.
It can be tricky to find outside examiners who know enough about your research area to ask useful questions. Your advisor may nominate professors of non-CS courses you've taken -- then at least you have one topic in common. However, do not feel that you have to take irrelevant courses simply to collect GBO examiners, especially since CS faculty from nearby universities can be approved as outside examiners.
Possible exam outcomes are described here. The most common ones are unconditional pass and conditional pass. In a conditional pass, the committee will require you to remedy some weakness in your preparation, e.g., by earning an A- or better in a particular course by a particular date.
It is the department's job -- not yours! -- to select your GBO committee and schedule the GBO. At least a month before you are to take the GBO (i.e., during your third year), you or your advisor should inform the CS graduate program coordinator, Cathy Thornton.
The hardest part of the GBO is finding an appropriate committee of 5 faculty examiners who are all free at the same time as you are. Fortunately, this is not your responsibility. It is handled by the department (i.e., Cathy together with the graduate program chair, Prof. Eisner).
Cathy will allow your advisor to suggest some appropriate examiners, and your advisor may in turn discuss this with you. In particular, you and your advisor should predict who will be on your thesis committee so that those faculty can be included on your GBO committee if possible.
Concretely, your advisor should send Cathy:
The names, ranks, and email addresses of at least 3 possible inside members in addition to himself/herself. (This includes an extra inside member, to be approved as an alternate in case someone cancels or doesn't show up.)
The names, ranks, and email addresses of at least 3 possible outside members. (This includes an extra outside member, to be approved as an alternate.) At least two of the outside members must be Associate, Full, or Emeritus Professors, so that someone can serve as chair.
If any of the above is not a tenure-track JHU faculty member, then your advisor should also send Cathy (1) that person's full CV, (2) a one-page summary of your research, and (3) an explanation of why that person's expertise is needed at your GBO or on your thesis committee. The department will combine these into a letter petitioning the Graduate Board for approval. (See here for details.) Note that approval takes 4 weeks. Of course, this is unnecessary if the examiner is already approved.
Cathy will ask for your availability so that she can begin scheduling the exam. The department will then nominate a panel of examiners in accordance with their availability, JHU regulations, and your advisor's suggestions. Remember that you do not pick the examiners, and under no circumstances should you offend the Graduate Board by filling out the GBO form yourself. Cathy sends this form to the Graduate Board for their approval, three weeks before the exam.
Once the exam is scheduled, Cathy will tell you who the examiners are so that you can send them your Preliminary Research Proposal.
You are required to meet formally with your thesis committee for a progress review at least once in each year, starting in the third year. Your annual progress letter from the department will be based on this meeting.
Your particular committee may also choose to set other requirements, including other meetings. A common requirement is that you write, present, and perhaps revise a thesis proposal -- an exercise that helps you focus your research and calibrate your expectations with the committee's.
It's often convenient to combine the third-year progress review with the GBO. The thesis committee can attend the GBO to hear your Preliminary Research Proposal, and then remain after the end of the GBO for additional discussion with you. This is straightforward to arrange when your thesis committee is a subset of your GBO committee. (Remember, the Graduate Board can authorize your non-JHU thesis committee members to serve as GBO examiners.) But even if some members of the thesis committee are not GBO examiners, they can still attend the GBO silently with permission of the GBO committee chair.
Sometime between your GBO and thesis defense, you must present your thesis work to the department in a 1-hour talk. This is primarily for the department's benefit -- everyone deserves to find out what you've been working on all those years.
Note: Some students use this requirement as a way to practice their job talk. Others use it as the first hour of their thesis defense.
The Ph.D. thesis, or dissertation, is the signal achievement of the Ph.D. degree. It is a large, careful, and substantive piece of of original work. Most computer science dissertations are 150-200 pages long, with hundreds of bibliographic references, and systematically investigate a set of ideas.
Your dissertation is presumably not the last piece of research you will ever publish, or even the most important. However, it may be one of the largest. Writing this document is a satisfying way to wrap up your graduate experience, but is itself a considerable creative act requiring plenty of time. You'll want to synthesize and explain several years of work (a process that may lead to new insights), and fill in the gaps.
Your advisor will help you decide when your thesis is essentially finished and ready to defend. You must give the thesis to your committee members at least 2 weeks before your scheduled defense date (and preferably earlier), so that they have time to read it carefully. Your defense date must also be publicly announced to the department.
The thesis defense is a public event, usually consisting of a 1-hour talk followed by questions from the committee and other audience members. Following the defense, the committee will decide what changes are required before they will sign off on the thesis.
Thesis committees almost always ask for changes, ranging from expository improvements to substantial further research. You can reduce this workload somewhat by consulting your committee frequently before the defense. But even so, you should plan for a month or more of hard work after the defense.
Your dissertation will be submitted to the JHU library and must follow certain formatting guidelines.
Time management can be tricky in the final year of the Ph.D. You may be applying and interviewing for jobs as you try to finish the research and write the thesis. And everything will take longer than you expect. So make sure to leave lots of slack in your schedule.
Ph.D. funding at JHU is pretty much the same as it is at other good CS departments in this country. No one is obliged to agree to fund you. But practically all of our Ph.D. students do receive funding for either 9 or 12 months per year.
Most students spend their first year or two funded by a teaching assistantship from the department (TA funding). Subsequently, students usually receive funding from their advisor in the form of a research assistantship (RA funding).
TA and RA funding are contingent on satisfactory progress toward the degree. Also, they may come with some strings attached. No matter how brilliant your research is, the department can't give you TA funding if you don't teach, and your advisor can't legally give you RA funding from a federal grant unless a reasonable amount of your work is related to the grant topic.
There are also other ways to get money:You may be able to win your own funding from some sort of fellowship. For example, if you are a U.S. citizen and have completed less than 12 months of full-time graduate study, you can apply for an NSF Graduate Research Fellowship.
If you can find a summer job doing CS research in industry, that can be good experience and pay well. (Although it might interrupt your progress toward graduation.) Of course, you cannot receive summer RA funding if you are also working a full-time job elsewhere.
The department hires course assistants (CAs) at an hourly rate. This could be a good way to make some money if you were not otherwise funded.
In general, you should discuss your plans with your advisor.
If you enjoy teaching and plan to become an academic, then you should aim to get some teaching experience during grad school.
You might ask a faculty member if you can give a couple of his or her lectures (especially if you are the TA). Be sure to ask for feedback. You could also help develop course materials, such as homework assignments.
More significantly, our department is willing to let Ph.D. students design and teach their own short courses. A short course is a 1-credit course that meets for a total of 12 hours (1 hour per week all semester, or 3 hours per week for 4 weeks). If you are interested, you should get a faculty sponsor and develop a course proposal for approval by the department chair. This should be done by the middle of the previous semester so that the course can be scheduled and listed in the course catalog.