Stanford Engineering
Computer Science
Engineering

PhD | Qualifying Examination

Overview

The qualifying examination tests a student's depth of knowledge and familiarity in their area of specialization. Qualifying exams are generally offered in all areas covered by the written comprehensive exam. It is possible for a student to request a qualifying exam in an area not already offered, such as one that cuts across current divisions. The feasibility of this request is determined on a case-by-case basis by the PhD program committee. A student should pass a qualifying exam no later than the end of their third year.

A student may take the qualifying exams only twice. In some cases a conditional pass is awarded. When the designated conditions have been met (such as CAing for a certain class, taking a course, or reading additional material in a specific area), the student is credited with the pass. If a student fails the qualifying exam a second time, the PhD program committee is contacted because its an indication that the student is not "making reasonable progress". This is cause for dismissal by default from the PhD program. The qualifying exams are a University requirement and are taken very seriously. Therefore, sufficient time and in-depth preparation must be given to the quals area that the student chooses, to ensure success.

The format of the qualifying exams varies from year-to-year and area-to-area, depending on the faculty member or quals chair in charge of each specific exam. Examples are in-class written exams, "take-home" written exams, oral exams, written assignments and/or a combination of the above. The quals chair administers the exams and the results must be submitted to the PhD program officer, as they will enter the information into the University's Axess (PeopleSoft) and Departmental database systems. Passing the qualifying exam certifies that the student is ready to begin dissertation work in the chosen area. If a student wishes to do dissertation work in an area other than their qualifying exam area, the student's advisor and/or the faculty in the new area will determine whether an additional exam is required.

Information about the Qualifying Examination

The student's advisor needs to email phdstudentservices@cs.stanford.edu (and cc faculty who were on the Quals committee) the qual results.

  • Artificial Intelligence (Jiajun Wu)
    • The candidate student must form a committee of 3 faculty members.  A committee needs to have (at least) 2 core AI faculty on it.  Upon request, we can consider having 1 core AI and (at least) 1 AI-affiliated faculty.  In all cases, at least 1 core AI faculty must be present.
    • The student is asked to prepare a 30-minute presentation on a research project the student is working on.
    • The student supplies to each committee member a short report summarizing the student’s research project and a list of references that is related to such a project. Report and list of references are due to the committee members 3 days before the exam.
    • The exam is one hour long and it is divided in two parts:
      • During the first half hour the student presents the research project.
      • The second half hour comprises a 30min QA session related to the research project by the committee. During such sessions committee members can (but are not necessarily committed to) ask questions related to any of the papers in the list of references. This gives the opportunity to committee members to assess general mastery of the area the student is working on.
  • Statistical Machine Learning (Percy Liang)
  • Natural Language Processing (Dan Jurafsky)
  • Biomedicine and Health (Anshul Kundaje)
    • The student must form a committee with 3 members.
      • The candidate’s advisor/s should be a member/s.
      • At least one member must be a Stanford CS faculty.
      • Two members must be working in Computational Biology.
      • One member will be non-computational from an affected field of biomedicine.
      • At least two members must be doing work directly relevant to the candidate’s work.
    • The exam should take 60-90 minutes. The candidate should prepare:
      • 30 minutes presentation on their research.
      • 30 minutes presentation on 3 papers which are jointly picked by the quals committee and the student, relating to the student’s current and future research directions.
    • After the exam has been taken, the candidate will email the CS PhD Student Services Admin, cc’ing all members of their quals committee, with the exam’s outcome.
  • HCI (Michael Bernstein)
  • InfoQual (Jure Leskovec)
  • Physiqual (Ron Fedkiw)
  • Systems Quals (Subhasish Mitra)
    • Form a panel of 3 professors (CS systems faculty). Select 3-4 papers, in consultation with the panel, in an area not identical to your thesis work for you to read, review and synthesize over a period of 3 weeks. Depending on the panel's advice, you may need to execute a small implementation project. For example, a project might answer a related research question, reproduce or compare results in a novel setting, or quantitatively investigate the implications of certain design decisions.
    • The exam has a written and an oral component. Three weeks after selecting the papers, turn in a 5-10 page report (not counting references) as well as pointers to any software or hardware artifacts created as part of the project (if any). Approximately one week after submitting the report, make an oral presentation to the panel, followed by questions.
  • Theory Qualifying Exam Overview (Moses Charikar)
    • Analysis of Algorithms
    • Form a panel of three professors, select 3-4 papers in an area related (but usually not identical) to your thesis work for you to read, review and synthesize over a period of a month (30 days). Write a report on your review/synthesis, give it to the committee, and also make an oral presentation to the committee, followed by questions.
  • Visual Computing (Doug James)
    • The goal of the visual computing qual is for students to demonstrate mastery of recent academic work in an area of visual computing relevant to their research. By general mastery we mean that the student can articulate and motivate key problems in the chosen area, can understand and clearly teach others about state-of-the-art solutions, and can critique the field’s leading solutions.
    • The student must form a committee of three faculty members, where at least two are faculty members actively performing research in the area of visual computing.
    • The student and the committee should agree on an active visual computing research area, and the student must make a “lecture” that summarizes and characterizes the state-of-the-field in this area. The lecture should involve the presentation of key ideas from at least three papers, focusing on both the problems those works are trying to solve, the key insights of the proposed technical solutions, and also how the papers relate. The student should get approval for the papers list from the committee prior to the qual. As part of the prepared lecture, the student should attempt to identify and discuss potential “next steps” research in the area. The committee will be interested in the student’s opinions about the strengths and limitations of the discussed papers, and how future work (e.g., ongoing research by the presenting student) might improve or build on them. In other words, we want to hear well-reasoned takes from the student under examination.
    • The exam period is one hour long and it is divided into two parts:
      • During the first ~45 minutes, the student will present a lecture on the topic of interest and any relevant background. The committee may intersperse with questions as appropriate.
      • The remaining 15 minutes comprises a Q&A session where committee members can ask questions related to the lecture and any of the chosen papers. This gives the committee an opportunity to assess the student’s general mastery of the research area chosen.