Ph.D. Degree Requirements
To receive a PhD in Computer Science at NYU, a student must:
- Satisfy a breadth requirement
- Satisfy a depth requirement
- Satisfy a teaching requirement
- Write and defend a thesis proposal
- Write and defend a PhD thesis
- Satisfy general NYU degree requirements
1. Breadth requirements
The breadth requirement form is availabe on the forms page for PhD students.
Rationale: The breadth requirement is designed to ensure competence across three broad areas of computer science: theory, systems, and applications. Within theory, the topic of algorithms is a requirement for every student.
Every student must complete requirements (1a), (1b), (1c), and (1d) by May 15 of the second year of PhD study in order for support to be continued.
Every student must receive a grade of A or A- on the final examination in the Honors Algorithms course. Students may take the final exam without being enrolled in the course.
The syllabus and final exam for every offering of the Honors Algorithms course will be determined by a committee of faculty members who routinely teach this class.
This requirement can be satisfied in two ways. Either:
- the student receives an A or A- in an approved course in systems listed in Appendix .
- the student has received an A or A- in a similar PhD-level systems course at another university with standards comparable to NYU's. This determination will be made by the Director of Graduate Studies (DGS). In this case, the student is required to work on a medium-size or larger software project at NYU This project can be part of coursework or the student's research. A brief report on the project must be accepted by the DGS.
This requirements is satisfied in one of three ways. Either:
- the student receives an A or A- in an approved applications course listed in Appendix ,
- the student passes a departmental exam in one of the subjects, if an exam is offered, or
- the student has received an A or A- in a similar PhD-level applications course at another university with standards comparable to NYU's. This determination will be made by the DGS.
(1d) Free choice
The student must either:
- receive an A or A- in an approved course in theory listed in Appendix .
- receive an A or A- in an additional course from the courses that can be used to satisfy requirements (1b) or (1c). This course cannot coincide with the courses used to satisfy (1b) and (1c) or
- have received an A or A- in a similar PhD-level course at another university with standards comparable to NYU's, substantially different from the courses used to satisfy requirements 1b and 1c. This determination will be made by the DGS.
Once a student has passed all courses and exams necessary to satisfy the breadth requirement, the student must inform his or her academic advisor in writing, specifying how each of parts 1a, 1b, 1c, and 1d has been satisfied. The academic advisor verifies that the breadth requirement rules were followed and forwards the information to the DGS.
The classes that can be used to satisfy breadth requirements will be reviewed regularly by the graduate curriculum committee; The graduate curriculum committee proposes the changes to the faculty for approval. Current list of approved classes can be found in the appendix.
The following standards will be maintained:
(a) Classes must be at the PhD level, i.e., more advanced than undergraduate or MS-level classes.
(b) The classes and exams must be rigorous and stable. Examples of inappropriate classes include those in which students are traditionally not differentially evaluated (e.g., all students receive As or "pass"), courses whose content completely changes from year to year, and courses in which grades are based on attendance or making a presentation of someone else's paper, rather than on tests and homework assignments.
(c) Acceptable systems classes will involve substantial software development.
2. Depth requirement
The depth requirement forms are availabe on the forms page for PhD students.
No later than May 15 of the second year of PhD study, each student must be involved in a research project under the guidance of a faculty research advisor. It is the responsibility of each student to find a faculty advisor and a research project, and to inform the DGS about his/her choice of advisor. Students must inform the DGS if they change the research advisor.
Students are required to form a depth exam committee and have the committee, an exam topic and a tentative date approved by the Director of Graduate Studies by the end of the first semester of their second year of studies, This exam may be taken no more than twice.
A DQE is administered by a committee of at least three faculty members, nominated by the student and his/her research advisor, and approved by the DGS. Each DQE will have a designated chair who is not the student's research advisor. If the research advisor is not a member of the committee, the research advisor must write a letter evaluating student's progress, and send it to the DQE committee members before the exam.
The DQE committee will define a syllabus, which must be approved by the DGS, and make the syllabus available to the student no later than two weeks before the exam.
The DQE has two parts:
(2a) An examination to demonstrate the student's knowledge of the research area. The scope of this exam should be similar to a typical PhD-level special topics course. It should not be as broad as an introductory course nor as narrow as a thesis. Examples of suitable topics are "Type theory in programming languages", "Probabilistic algorithms", "Computational learning theory", "3-D modeling", "Semidefinite programming", and "Low-power computing". Topics such as "Databases" or "Programming languages" would be too broad; topics such as "Voronoi diagrams" or "Tail-recursion optimization" would be too narrow. This exam may be oral or written, at the discretion of the committee. The requirement is that it seriously test the student's knowledge of a research area as distinct from the student's research accomplishments.
(2b) An oral presentation of the student's research accomplishments. A student is expected to have conducted original research by the time of the exam. This research may have have been carried out independently or in collaboration with faculty, research staff, or other students. Students are encouraged, but not required, to have publication-worthy results by the time of the exam. It is not sufficient for a student to present a survey of previous work in an area or a reimplementation of algorithms, techniques, or systems developed by others.
The committee, by majority vote, gives a separate grade for (2a) and (2b) as one of "PhD Pass", "MS Pass", or "Fail." A PhD pass on both parts must be achieved for support to be continued beyond the second year. A student who receives a "PhD Pass" on only one part of the exam may request permission from the committee to retake only the other part of the exam.
If a student has passed the DQE and then changes his/her area of research, the student need not retake the DQE.
3. Teaching requirement
By the end of the third year of study, each student must have served as a section leader of at least one course in the department. Courses on related topics outside the department may also be used to satisfy this requirement subject to approval by the DGS. The student must also participate in the department's teacher training session at or prior to the semester in which they teach. In certain circumstances, the DGS may allow the student to satisfy this requirement by serving as a course assistant or as a grader. These exceptions will be determined by the DGS based on the availability of suitable recitations.
4. Thesis proposal and presentation
Students are required to form a thesis proposal committee and have the committee and a tentative date for the thesis proposal presentation approved by the Chair and the Director of Graduate Studies by the end of the first semester of their third year of studies.
When a student is ready to start work on the PhD thesis, the student must (i) select, with the approval of his/her research advisor and the DGS, a thesis reading committee, and (ii) submit a written thesis proposal to the committee.
The student and the student's research advisor suggest the composition of the thesis reading committee for approval by the DGS and Department Chair. The committee must include at least three members. All changes to the composition of the committee must be approved by the DGS and the Chair. The committee members can be regular computer science faculty, faculty from other departments, or individuals of like standing from outside the University. At least one member of the reading committee must be regular Computer Science faculty.
The thesis proposal should include:
- a description of the research topic
- an explanation of how the research will advance the state of the art, and
- a tentative research plan
After the thesis reading committee has approved the thesis proposal, the student should schedule a thesis proposal presentation and notify the Program Adminisitrator once this has been finalized. This presentation should be announced to the faculty by the Program Administrator,PhD Program, at least one week before it occurs. The presentation may or may not be open to people other than faculty, at the discretion of the research advisor.
Substantial subsequent changes to the thesis topic must be approved by the thesis reading committee. The proposal must be defended no later than May 15 of the third year of studies.
5. Thesis and thesis defense
The final step in the PhD program is the student's defense of his/her PhD thesis. The procedures to be followed for the thesis defense can be found on the Dissertation Defense Checklist .
6. General NYU requirements
Students must end the semester in which they take their fifth class and all subsequent semesters with a GPA of 3.5 or higher. Note that the departmental requirement in this case is more stringent than the GSAS requirement (cumulative GPA of at least 3.0).
In addition the following general GSAS requirements have to be satisfied:
- Students must complete three years of full-time study beyond the undergraduate degree, at least one year of which must be in residence at the GSAS.
- Students must complete 72 points of graduate credit including at least 32 points for courses taken at the GSAS. At any time, students must have successfully completed 66 percent of credits attempted while at NYU, not including the current semester. Courses with grades of I, W, and F are not considered successfully completed.
- Time Limit. All requirements for the doctoral degree must be completed no later than ten years from the initial date of matriculation. However, if the student transfers credit from classes taken as part of a previously earned master's degree, then the time limit is seven years.
Other GSAS and NYU requirements can be found in the GSAS Bulletin.
7. Academic standing
To be in good academic standing a student must meet the deadlines for all requirements specified in this document and maintain the required minimal GPA. For supported students, maintaining good academic standing is a condition of the guaranteed support. If a student fails to maintain good academic standing, his or her support may be discontinued, and the student may be removed from the program. A student's academic standing is determined by the DGS each semester. The PhD admissions and financial aid committee decides in which cases support is discontinued. In most cases, a student will be removed from the program when his or her support is discontinued for failure to maintain good academic standing.
The following courses can be used to satisfy the breadth requirements:
- CSCI-GA.2243 High Performance Computer Architecture
- CSCI-GA.2434 Advanced Database Systems
- CSCI-GA.2620 Networks and Mobile Systems
- CSCI-GA.2621 Distributed Systems
- CSCI-GA.3110 Honors Programming Languages
- CSCI-GA.3130 Honors Compilers
- CSCI-GA.3140 Abstract Interpretation
- CSCI-GA.3250 Honors Operating Systems
- CSCI-GA.2270 Computer Graphics
- CSCI-GA.2271 Computer Vision
- CSCI-GA.2560 Artificial Intelligence
- CSCI-GA.2565 Machine Learning
- CSCI-GA.2566 Foundations of Machine Learning
- CSCI-GA.2567 Machine Learning and Computational Statistics
- CSCI-GA.2572 Deep Learning
- CSCI-GA.2590 Natural Language Processing
NOTE: Only one of these classes can be counted for breadth requirements (either Applications or Free Choice). Machine Learning emphasizes applications, and Foundations of Machine Learning emphasizes theoretical aspects of machine learning, although both include theoretical and practical components. Please familiarize yourself with class requirements and consult your academic advisor before choosing one of these classes.
1d. Free choice
Any of the courses listed under 1b and 1c, or any of the following courses: