Doctor of Philosophy in Computer Science

The Doctor of Philosophy (Ph.D.) is the highest degree offered in any discipline. It is intended for those who will pursue research in the field. The degree is usually required to teach computer science at the university level, and for employment in prestigious industrial research laboratories. Students may enter the doctoral program with a bachelor's degree or with a master's degree, and must have adequate background in computer science. Since completion of the degree requires presentation of a dissertation comprising original research, the time required is pre-determined. Four to six years of full-time work are typical, less if the student enters with a master's degree in CS.

There is extensive information on doctoral requirements in the PSU Bulletin, and students are responsible for familiarizing themselves with that information. Departmental requirements are given below.


Admission

Admission to the Ph.D. program is very competitive, and a limited number of openings are available each year. Two factors, beyond the student's qualifications, influence admission to the Ph.D. program: availability of a Ph.D. adviser and of financial aid. Ph.D. applicants submit the same admission information required for M.S. applicants, with the exception of having a more detailed statement of purpose, and marking "Ph.D." on the Computer Science Departmental application.

An applicant who is not at PSU must communicate his/her own interests very well to make a successful Ph.D. application, because admission will be given only to those for whom potential advisers exist in the department. The student can identify potential advisers in his/her personal statement accompanying the application and can communicate directly with faculty about research topics. Ideally, one of the student's recommendation letters would come from a prospective adviser with whom a solid contact has been established.

Return to Top

Advising

Upon being accepted as a graduate student, you will be assigned a faculty adviser. Doctoral students are also assigned additional faculty members to make up a preliminary advisory committee.

The selection of a Ph.D. research adviser is a process extending over several academic terms whereby the faculty member and student get to know each other and until both are certain that they can work together. The student will usually take courses from the faculty member, and they will discuss potential research topics.

When you begin research work on a Ph.D. dissertation, you should change your adviser to the person who will supervise your dissertation. Your adviser must be a full-time faculty member, and he or she must agree to the change. There are PSU regulations about advising duties (particularly for doctoral degrees). Make sure that any change has been properly recorded and processed by the Computer Science Office.

Return to Top

Assistantships

A majority of the departmental teaching and research assistantships support Ph.D. students. A newly entered student is most likely to be given a teaching assistantship, while someone who has been at PSU and established a good connection with a potential adviser is more likely to be given a research assistantship. First year Ph.D. students who are admitted with financial support are typically supported as a TA for up to 4 quarters. These students are then asked to apply to serve as a TA for up to 3 additional quarters. Continued support for Ph.D. students depends on their making satisfactory progress in the program. Students are encouraged to find a research adviser and research assistantship sometime during their second year.

Return to Top

Course Requirements

Ph.D. students must complete an approved program of 90 graduate-level credits.

  • 18 credits are required core courses(Core courses must be completed by the end of the second year):

CS 581 Theory of Computation
CS 584 Algorithm Design and Analysis
CS 558 Programming Languages
CS 533 Concepts of Operating Systems
CS 538 Computer Architecture or ECE 585 and 586
CS 510 Scholarship Skills

  • 27 elective credits from the focus areas or other graduate courses:

Programming languages
Systems and networking
Databases
Software engineering
Theory of computing

  • 18 credits of 601 research or elective credits.
  • 27 credits are dissertation research credits (CS 603).

Return to Top

Course Restrictions

All regular courses must be graded (that is, not P/NP) and passed with a grade of B or better.

Research, projects, internships, and reading & conference credits may be graded P/NP or with a letter grade. These credits are set up independently by the student and a faculty member. These credits must be passed with a P or a grade of B or better.

Students must maintain a 3.5 GPA in core course work, with a minimum grade of B. For the elective courses:

  • These courses should be used to strengthen the program in the student's research area.
  • CS 501 (Research) and CS 509 (Graduate Assistant Practicum) may not be counted.
  • A maximum of 6 credits of CS 505 (Reading and Conference) and CS 506 (Special Projects) may be counted.

Your CS adviser must approve the courses in your program. Students and their adviser¡¯s will fill out a PhD Plan of Study after the student successfully completes 9 graduate level credits, no later than 6 months before their Research Proficiency Exam.

Return to Top

Candidacy

To be admitted to Ph.D. candidacy a student must:

  • First, pass the comprehensive exam.
  • Second, present a dissertation proposal describing the proposed research, to be approved by the dissertation committee.

Return to Top

Comprehensive Exam

The Comprehensive Examination is a university requirement for the Ph.D. degree. In the CS department, it takes the form of a Research Proficiency Examination (RPE), covering original research performed by the student. The exam consists of a written paper, a public oral presentation, and a public question and answer session. Passing this exam is required before a Ph.D. student can prepare and defend a dissertation proposal. The exam is normally taken in the Spring term of the student's second year.

Purposes:

The RPE serves several purposes and meets the needs of several constituencies:

  • For the university, it is the comprehensive exam that is required of all Ph.D. students.
  • For the faculty, the exam provides an opportunity to assess the potential of the student to undertake guided research, including the ability to present research results in written and oral forms. The exam provides a strong indication of whether the student can successfully conduct research ultimately leading to a Ph.D.
  • For students, it provides the experience of conducting independent research and working with a particular research advisor, of learning to present that research in the way that is customary in our field, and of getting feedback from their peers and from other faculty on their performance.
  • For the student body, it helps build a learning community by encouraging student collaboration and mentoring.
  • For all members of the department, it provides a forum for sharing information about ongoing research.
The Exam:

The RPE consists of a written paper and oral presentation describing original research conducted by the student under the supervision of a faculty advisor. The paper and presentation should demonstrate that the student has identified a novel and worthwhile research problem; mastered the relevant previous work in the area; proposed a plausible approach to the problem; and, where feasible, obtained initial results.

As a basic guideline, the RPE work should be of a format and standard that would be suitable for acceptance to a reputable conference in the research area, except that:

  • Its original results may be more preliminary than a conference would require;
  • The paper and talk should be accessible to the entire CS community, i.e., a broader audience than for a typical research conference.

The research reported by the RPE will typically be joint work between the student and the faculty advisor. The paper and talk, however, should be the student's own work.

Evaluation Criteria:

The paper and oral presentation will be evaluated according to the following criteria.

Both the paper and the oral presentation should:

  • Be well-organized and tell a coherent "story."
  • Summarize the problem, approach, and results obtained.
  • Describe the problem addressed by the research, and explain why this problem is interesting and important within the context of the research area.
  • Describe the research methodology applied.
  • Describe the research results obtained.
  • Explicitly characterize the original contributions of the student.
  • Describe the relationship with relevant existing work (not merely summarizing the other work).
The paper should:
  • Be largely accessible to a general CS audience. (There can be deep technical parts that require specialized knowledge, but the remainder of the paper should summarize these specialized parts in more accessible form.)
  • Be written in mechanically sound English.
  • Make appropriate use of diagrams.
  • Fit within the suggested range of number of pages (as described below).
The oral presentation should:
  • Be audible and understandable.
  • Be accessible to a general CS audience.
  • Fit within the allowed time period.
In addition, when receiving questions after the talk, the student should demonstrate the ability to:
  • Grasp reasonable questions promptly
  • Give succinct and correct answers
  • Engage in dialogue with the questioner to clarify questions where necessary
Exam Mechanics:

The exam is normally conducted once a year at the end of Spring quarter. The date of the exam will be on Friday of the 7th week of Spring Term. The exam is normally taken in the Spring term of a student's second year. Well-prepared students, such as those who enter the program already holding an MS, may take the exam in their first year. Students who enter the program in Winter term will normally be treated as if they had entered in the previous Fall term; those who enter the program in Spring term will normally be treated as if they had entered in the next Fall term. Students can delay the exam beyond the Spring of their second year only by special permission of the graduate committee; such permission should only be granted in exceptional circumstances (e.g., serious illness).

Students are strongly urged to identify an RPE advisor and begin working towards the RPE during the Fall quarter previous to the exam. Students intending to take the exam must register with the Graduate Coordinator by the end of Winter quarter. Abstracts (of no more than 250 words) must be submitted to the Graduate Coordinator five weeks prior to oral presentation. The research paper is due two weeks prior to the presentations. Faculty will determine and communicate the outcomes of the exam within one week after the presentations.

Research papers should be in LNCS conference format, with a target length of 15 pages and a strict maximum length of 20 pages. Papers must be submitted electronically, in pdf format, to the Graduate Coordinator.

Oral presentations will be given on one or two adjacent days. Talks are open to the entire CS department community. Each talk should be timed to take 30 minutes. Talks will be scheduled one hour apart to allow ample time for questions from the audience.

The possible outcomes of the exam are: pass, retake oral, retake written, retake both, fail with no retake.

If the paper must be rewritten or the presentation must be repeated, this must normally occur by the end of the first month of the following Fall term. Only one retake attempt is permitted.

Evaluation Mechanism:

Each student taking the RPE is assigned a committee of six faculty, who are responsible for reading the paper, attending the oral presentation, and meeting to determine the exam outcome. This committee includes the faculty member who supervised the student's RPE research. It must also include at least two members from outside the student's research area. The committee is selected by the Graduate Coordinator in consultation with the student's RPE advisor and with the approval of the Chair of the Graduate Committee. The Graduate Coordinator will attempt to distribute committee memberships pseudo-randomly to limit the load on any one faculty member.

Although each student committee has decision-making power on the outcome of that student's RPE, the faculty as a whole are encouraged to be involved in the RPE process for all students by attending all the oral presentations. Moreover, to keep the decision-making process as transparent and uniform as possible, the discussions and decisions of all the RPE committees will take place during a single meeting at which all faculty members are welcome. Faculty members who attend a student's talk but are not on that student's committee can provide input to the committee's discussion during this meeting.

It is hoped and expected that committees will reach decisions by consensus. However, if voting is required, at least five positive votes are required to pass (either of) the oral or written components. A "fail with no retake allowed" also requires at least five votes.

The exam outcome will be conveyed in a letter from the department chair to the student. In addition, the student's research advisor will convey a set of faculty comments to the student to help him/her improve in future research, writing, and/or oral presentation.

If a "retake" of the paper or presentation is required, the student committee will meet again to reach a decision. (In exceptional circumstances, such as a change in RPE topic, changes in the committee membership may be made if approved by the Chair of the Graduate Committee.) As in the initial case, all faculty are encouraged to attend the second presentation and are welcome to attend the committee's decision meeting. Five positive committee votes are again required to pass either component; no second retake is possible.

Additional Issues:

Students who have already completed and defended an MS thesis may do an RPE paper and talk covering the same research topic. However, students may not submit their MS thesis for their RPE. If a considerable time has elapsed since completion of the MS, it may make more sense for the student to do their RPE on a different (though perhaps related) piece of work.

Similarly, students who have published a conference or journal paper may submit an RPE on the topic of the paper. The RPE paper may be based on the published paper, but might require modification to address a broad CS audience and to remove any portions that were written by co-authors.

There is no assumption that the faculty member advising the student's RPE work will become the student's dissertation adviser, or that the RPE topic will become the student's dissertation topic.

RPE forms: Several forms may need to be completed in the course of doing the RPE. These include:
RPE schedule for 2008: Please direct any questions to the Graduate Coordinator.
  • March 1: Last day to register. The advisory committees will be confirmed by the end of Winter term.
  • April 11: Student abstracts are due.
  • April 28: RPE committees will be confirmed by April 28, and students informed of who the members are.
  • May 2: RPE papers are due. GC will post them on the faculty Twiki, but students should also give individual copies to their committee members.
  • May 16: RPE presentations.
  • May 23: Students will be informed of results.

Return to Top

Dissertation Proposal

After passing the comprehensive exam, a student prepares a written proposal describing his or her dissertation research topic, in consultation with the chair of the advisory committee. (It may be appropriate for the student to first find a new advisory committee chair, if the research area has shifted since admission.) The proposal:

  • Identifies the general research field (e.g. networks) and presents an annotated bibliography of literature relevant to the particular dissertation topic (e.g. verification of network routing protocols).
  • Describes the research problem.
  • Describes the proposed approach to solving the problem, including specific methods to be tried.
  • Gives a timeline for completing the work.

The proposal is circulated among the dissertation committee, and, when they agree that it is satisfactory, the student publically presents the proposal. The presentation begins with a public talk, then the dissertation committee questions the student in private.
Students making satisfactory progress are expected to have a dissertation proposal approved by the end of the third year. The abstract and date for the defense must be submitted to the Graduate Coordinator at least two weeks in advance of the date of the defense.


Return to Top

Dissertation

Doctoral students must complete a dissertation comprising original research work, and defend it at a final oral examination. It is expected that the work will be of a quality to merit publication in refereed journals or conferences. The abstract and date for the defense must be submitted to the Graduate Coordinator at least two weeks in advance of the date of the defense. After the dissertation is in its final form, the student should submit a copy in pdf format to the webmaster at cs.pdx.edu for posting.

Dissertation Guidelines

The University has a set of Dissertation Guidelines that describe the format and content requirements for the dissertation.