PhD Requirements

Applicants who already have a Masters in Computer Science or closely related field may apply directly to the PhD program.

Those with a Bachelors in Computer Science or closely related field who intend to complete the PhD may apply to the MS/PhD track. A letter of intent to continue for the PhD must be submitted by the sixth week of classes in the last semester of the Masters program. This letter should name the advisor, briefly describe the research area, and include a list of proposed qualifying exam topics. Students will be notified of acceptance into the PhD program by the twelfth week of classes.

Students who entered intending only to complete a Masters may apply to the PhD program, but the process is more formal since the initial application was not for the PhD. Such students should submit an essay similar to the one required for admission with a Masters from outside Tufts. In addition, the student should provide the names and email addresses of three individuals who can serve as references. One should be the proposed PhD advisor, one should be another Tufts Computer Science faculty member who would be willing to serve on the dissertation committee, and the third can either be a Tufts Computer Science faculty member, or someone qualified to serve as the "outside member" on the PhD dissertation committee. Agreeing to serve as a reference doesn't commit the person to actually being on the committee, since the student's research may take a different direction. In this case the application deadline is the same as for students applying from outside, and students will be notified of acceptance along with outside students.

The department differentiates between admission to the Ph.D. program and Ph.D. candidacy. No students are accepted as formal doctoral candidates until they have exhibited merit in a qualifying examination and have identified a faculty member who has agreed to be their dissertation supervisor. Doctoral candidates are expected to plan a program of research under the direction of their dissertation supervisor and with the guidance of a faculty committee. Upon completion of this research, the candidate must prepare and publicly defend their dissertation.

Typical areas available for dissertations in computer science include: Algorithms, Artificial Intelligence, CAD and VLSI Design, Computational Biology, Computational Geometry, Human Computer Interaction, Machine Learning, Network Administration, Parallel Computing, Software Engineering, Theory of Computation, Visualization.

Students wishing to obtain a Ph.D. degree in Computer Science at Tufts must successfully complete several requirements, explained in greater detail below. Students should pass the qualifying exams before beginning significant work on the dissertation. The community/residence requirement may be satisfied concurrently with the others. To ensure that all students are making satisfactory progress toward the Ph.D., the Computer Science faculty will conduct an annual review of all students.

Students should earn 20 credits for the Ph.D. degree. At least two credits must be regular 100-200 level courses. The rest can be earned by multiple registrations for Comp 297, Comp298, or by taking other courses approved by the student's advisor. Students should normally register for at least two credits per semester consisting of some combination of regular courses and research credits. Full-time Ph.D. students should also register for Comp 502 during each term of full-time study, including both summer sessions.

This requirement is fulfilled by attending at least 50% of the weekly departmental seminars in each of four semesters. This is the minimal requirement--it is hoped that most students will attend a larger percentage of seminars for a longer period of time.

A student qualifies to begin research on their selected topic by completing the following (usually in the given order):

• Obtaining core competence
• Completing a preliminary research project
• Giving an oral presentation on this project
• Passing a written qualifying exam
• Passing an oral qualifying exam

The examination committee (see below) will make the final determination that these requirements have been fulfilled. A typical student should be able to satisfy the first three requirements while obtaining the Master's degree. In any case, the student should meet with the members of the examination committee well before the examinations to make sure that preparation for these requirements has been adequate.

Core competence: Each Ph.D. candidate is expected to have competence at the level of an excellent undergraduate in the following core areas:

• Computer Architecture and Assembly Language
• Programming Languages
• Data Structures and Algorithms
• Theory of Computation

In most cases, the candidate will have satisfied these requirements through course work in their Bachelor's or Master's degree, before beginning the Ph.D. program. The examiners at the oral part of the qualifying exam may ask questions on these areas.

Each Ph.D. candidate is also expected to have competence and experience in software implementation. The candidate must have completed an implementation project with documentation. In most cases, this will be a project completed as part of their previous degree work, such as an M.S. project or a final project in an implementation-oriented course.

Preliminary project: Each Ph.D. candidate is expected to have competence in execution and presentation of research, demonstrated by completing a research project. In most cases, this will be their M.S. project or thesis, whether completed at Tufts or elsewhere, or possibly a project in a regular course or an independent study. This is intended to start the student toward the path of selecting a research topic and a research advisor. The project should include writing of a report or thesis (which may consist of original research, but may also be a critical survey, etc.), giving a thorough review of the relevant literature, and listing some open problems that could lead to a Ph.D. dissertation. The examiners at the oral part of the qualifying exam will make the final determination whether this project demonstrates the student's readiness to begin Ph.D. research.

Oral presentation: The student will make a public oral presentation of at least 45 minutes on this research project. The public presentation must be advertised to the entire department at least one week prior to the presentation. Part of the talk should outline promising research directions. For a student entering the Tufts Ph.D. program with a Masters degree from elsewhere, the project itself will typically be work from their Masters program, but they must give a new presentation of that work at Tufts.

Written qualifying exam: Each Ph.D. candidate is expected to have advanced competence in three sub-areas of computer science. Two of these sub-areas should be closely related areas that together span the areas of the student's proposed research. The third sub-area must be further removed: if a student will conduct research in applied computer science, two of the sub-areas will be applied computer science and one will be theoretical, and vice versa. Some areas (algorithms, graph theory, computational geometry) will usually be theoretical and others (human-computer interaction, virtual reality, visualization) will usually be applied, but many areas (artificial intelligence, distributed computing, databases) could be either theoretical or applied depending on the focus.

Each Ph.D. student will be examined by a committee of three faculty members: the proposed dissertation advisor, who will serve as chair; a faculty member in a closely-related field selected by the Ph.D. student; and a faculty member in a somewhat removed field chosen jointly by the proposed dissertation advisor and the director of graduate studies. The Ph.D. student will contract with each of these faculty members as to the expected depth and breadth of the sub-area to be examined. These three faculty members will meet with the student to discuss and decide on a syllabus of topics for the examination. Ph.D. students will have four hours to complete a written exam with an in-depth question or questions written by each of the examiners. At the end of the four hours the student submits his or her responses but keeps a copy of the exam questions in what becomes the equivalent of a "take home exam." The student may consult published print materials but may not consult other human beings as he or she prepares to answer the follow-up questions pertaining to the written examination at the beginning of the oral qualifying exam described below.

Oral qualifying exam: Within two weeks of the written exam, they will take an oral exam covering the core areas, the research talk, and the advanced areas, but with an emphasis on the advanced areas.

In particular, the oral exam will begin with questions pertaining to the student's responses on the written examination. It is expected that the student will have independently both reviewed his or her responses and developed revisions and/or enhancements during the interviewing time. The student may have consulted published print materials but must not have consulted other human beings. The decision to pass or fail the student will be made by the examination committee that administered the written and oral exams, based on the following factors:

• score on the written (advanced area) exam
• performance on the oral (both core and advanced) exam
• evidence of ability to do thesis-level research

If a student fails this examination, then the student may try again within the next two semesters. A student is allowed at most two attempts to satisfy this requirement. For the second attempt, the committee will decide which parts of the exam (oral presentation, written qualifying exam, oral qualifying exam) need to be retaken. The examination is usually taken in the first year, but can be deferred at the discretion of the proposed dissertation advisor and the director of graduate studies.

Within six months of the successful completion of the qualifying examination, the student should submit a thesis prospectus approved by a full member of the Computer Science faculty who will serve as the thesis advisor. If the Ph.D. thesis advisor is not a tenure-track Computer Science faculty member, the student must have a second advisor who is a member of the department, and the student must keep the second advisor well-informed about his/her research progress. The student will do research under the direction of the advisor, write a dissertation about that work, have the thesis read by a committee, and defend the work in an oral presentation to the committee and all who are interested. The student's advisor will propose a committee to the faculty for approval. The thesis committee, chaired by the advisor, will include at least three Computer Science faculty members, one member from another department at Tufts, and one additional person referred to as the "outside member". This outside member cannot be a member of the Computer Science faculty, should be from outside the university whenever possible, and is expected to be a recognized authority on the subject of the dissertation. On the one hand, the outside member is included in the committee in order to provide an objective and disinterested evaluation of the student's work. On the other hand, once the committee has been formed, the outside member is expected to participate in giving the student the assistance and feedback necessary to assure that the dissertation meets the appropriate standard.

As a general rule, Ph.D. dissertations should be extended studies that go well beyond the scope of individual scholarly articles. They are expected to present a broad review of relevant literature and theory, to study extensively the problem posed, and to place the results in a larger intellectual/research context. On occasion, the dissertation may substitute two to four less extended, original studies on closely related problems. Such a dissertation must be presented as a single document, must have a common general introduction and literature review, must have appropriate connecting matter, and must have a general conclusion relating the results of the separate studies.

The Computer Science faculty collectively conducts an annual review of all doctoral students/candidates. Students/candidates will be asked to provide their academic advisor/oral examination chairperson/thesis director (whichever is most appropriate) with a summary of their progress, achievements, and difficulties in the preceding calendar year to be used in the review. The advisor and/or department chair will report to the student, in writing, the suggestions resulting from the annual review.