Next: Computer Science Course Descriptions
Up: Graduate Degree Programs in
Previous: Master's Program
Subsections
The Ph.D. requirements were recently redesigned to encourage students to
get involved with research early in their studies. Continuation beyond
the first two years of the Ph.D. program requires successful completion
of the breadth and the research requirements. The Ph.D. requirements
consist of six components:
Students during their first two
years in the Ph.D. program must pass a qualifying exam
or receive a ``quals pass'' in the designated course in
four of the following six subject areas:
- 1.
- Operating Systems (Computer Science 210)
- 2.
- Architecture (Computer Science 220)
- 3.
- Algorithms (Computer Science 230)
- 4.
- Models and Complexity (Computer Science 240)
- 5.
- Numerical Analysis (Computer Science 250)
- 6.
- Artificial Intelligence (Computer Science 270)
The designated course for each subject area is listed
in parentheses following each subject area. More
advanced courses may be substituted with the approval
of the Director of Graduate Studies. A quals pass in the
AI area can also be earned in Computer Science 271. The subject areas
must include at least one of the first two areas 1
and 2 and at least one of the next two areas 3 and 4.
The qualifying exam in each area is
given once a year, typically immediately before the
beginning of the semester in which the designated
course is offered. The qualifying exams for areas 2,
3, and 6 are offered in late August, and the qualifying
exams for areas 1, 4, and 5 are offered in mid-January.
Students must complete a two-semester research project
under faculty supervision during their first two years
in the Ph.D. program. The goal of a research project
is to get the Ph.D. student excited about research
early in the program and to give faculty feedback on
the student's research potential. The project is
judged on the student's ability to dissect problems,
propose solutions, and analyze critically.
-
All first-year Ph.D. students must attend and participate in a special
research seminar course Computer Science 300 during the first
semester to help in their planning for the research
project. The course, which may be taken for credit, consists of descriptions by
faculty members of their research interests,
instruction in research strategies and technical
presentation, and practice talks by the first-year
students.
-
Each Ph.D. student must find a faculty member to
supervise the research project by March 15 of the first
year. During the following month, the student must
turn in a two-page written proposal and make a
15-minute presentation of the proposed project. The
results of the project are presented in a 30-minute
presentation during April of the second year,
accompanied by a research report. The report itself
must be written by the student alone, and if there is
joint work as part of the research, the student's
contributions should be specified.
-
The faculty member who supervises the research project
need not be the student's subsequent dissertation
advisor or be in the student's area of ultimate
concentration. On the other hand, in some cases the
project may lead to an eventual Ph.D. thesis.
-
In unusual cases, when there is a minor problem with
the presentation or work, the student can be given a
one-month extension.
Ph.D. students must complete at least four 200-level or higher-level
regular![[*]](http://www.cs.duke.edu/icons/latex2html/icons.gif/foot_motif.gif)
computer science courses beyond the designated courses for the
four subject areas passed in the breadth requirement.
At most two of the courses can
be outside the department, but only if they pertain directly to computer
science or to the Ph.D. program of the student. In addition to the four
courses, each student must take two courses
in a related, but non-computer science field, not counting any
course used to satisfy other requirements. The Director of
Graduate Studies should be consulted to determine if the proposed
field and courses are considered ``related''.
Students must earn at least a G on each course applicable
to the program, with a combined average no less than midway between E
and G.
Courses can be taken at the University of North
Carolina and at North Carolina State University through
inter-institutional registration. Such courses require
the permission of the Director of Graduate Studies and
that the student is registered for an equal number of
course units at Duke.
Each student is required
to have at least one semester of teaching experience, most
commonly in the form of a teaching assistantship.
Each Ph.D. student must present a public thesis
proposal, typically during the third year, accompanied
by a 15-30 page document addressing the motivation and
goals of the research, the state of the art in the
literature, the results already achieved, and the plan
for completion. The thesis committee will review the
presentation and question the student on the proposed
research and the particular area of concentration.
Each Ph.D. student must complete a doctoral
dissertation and defend it publicly before the thesis
committee. The full dissertation must be delivered to
the committee four weeks prior to the defense; minor
modifications suggested by the committee can be
incorporated after the thesis defense.
Full-time Ph.D. students are provided tuition and stipend support through
fellowships, research assistantships, and teaching assistantships.
Opportunities for assistantships are occasionally available to Master's
students.
Admission to the Department is highly competitive.
The admissions process for the Ph.D. program is entirely separate from
that of the Master's program. Students may apply to both programs.
Potential applicants can request further information or answers to
questions by sending email to dgs@cs.duke.edu.
Detailed information about the Department and its
programs is also available electronically via the World
Wide Web at http://www.cs.duke.edu/.
Next: Computer Science Course Descriptions
Up: Graduate Degree Programs in
Previous: Master's Program
Duke Department of Computer Science