Ph.D. Comp-II Research Proposals

The Comp-II Requirement

As part of your Ph.D. program, you will need to complete the Comprehensive-II requirement, "normally" around your sixth term. As of students who started Fall 2019 or later, it must be done by the end of your seventh term, or you will be required to withdraw from the program (save for a very limited set of extenuating circumstances). You will need to complete the following steps:

1. Informally decide on your research topic with your supervisor.
2. Write a research proposal; see below for details.
3. Assemble an Advisory Committee:
   1. Choose two Computer Science faculty members (other than your supervisor). Consult your supervisor for advice on selecting your committee.
   2. Ask the faculty members to be part of your committee via email. Include the title of your research proposal and some details about the content (e.g., the abstract). Wait for both faculty members to confirm that they will be part of the committee.
   3. Send an email to the Director of Graduate Studies asking them to approve your committee. You should include the names of the committee members and, at a minimum, the title of your research proposal and a description of the content. Alternatively, you can include your finished research proposal in this email (rather than just the title and a description) to save a step later on. Wait for the director to confirm your committee selection.
4. Schedule the examination:
   1. If you did not include your research proposal in your email to the Director of Graduate Studies, send them an email with a copy of your proposal.
      • If the body of your research proposal is longer than 20 pages, you must first send an email to your committee members asking for their consent. Every committee member must confirm that they are willing to read a research proposal that exceeds the suggested length.
   2. Send an email to your committee—your supervisor and the two other CS faculty members—with a copy of your research proposal and a proposed date and time for your examination. This date must be at least two weeks from the time that you send the email. Wait for all committee members to confirm the date and time. Reschedule as necessary.
   3. Book a room for the examination. (Hint: you probably want to book DC 2310 or DC 2314.)
   4. Send an email to your committee with the final date, time, and room number.
5. Make a presentation for your research proposal. The presentation should be approximately 30 minutes long.
6. Complete the examination:
   1. Print a PhD Comprehensive-II Examination Report form. Fill in the name, student ID, and date.
   2. Give your presentation to the committee at the scheduled time and place. Consult the presentation tips.
   3. The committee members will ask you questions about your research proposal.
      • The question period typically lasts between one and three hours; 1.5 hours is a typical length. In extreme cases, the question period may last longer.
      • If you wrote about something in the research proposal, expect to be asked about it during the question period.
      • The committee members may also ask you questions about underlying background information that is relevant to your topic (e.g., a student presenting a new authentication scheme for Android might be asked general questions about Android's security model, and a student presenting a new pairing-based cryptographic protocol might be asked questions about elliptic curves and group theory).
   4. Have your committee members fill out and sign the examination report form.
   5. Bring the completed form to the CS Graduate Office.

While all of these steps must be completed to fulfill the requirement, there is some leniency in the ordering. You can assemble an Advisory Committee before your research proposal is finished, but do not wait too long between assembling the commitee and sending them the proposal. You can (and should) make the presentation while scheduling the examination time.

Writing the Research Proposal

The purpose of the research proposal is to convince the committee that the chosen research area is suitable and that you possess an appropriate breadth of knowledge in the chosen area. Specifically, a research proposal must demonstrate two things:

1. Your chosen topic can produce a reasonable Ph.D. thesis. The topic is appropriately sized, and the topic includes sufficiently novel research.
2. You are capable of conducting the research that you are proposing. Your proposal should demonstrate that you have sufficient knowledge of the subject matter. If you have already completed some work related to the topic, describe it in the research proposal. Note that this prior work does not have to be part of your proposed thesis. For example, if you published papers during your Master's program in a similar research area, you should include these as evidence that you can conduct the proposed research.

The format and content of research proposals vary significantly, and there is no established template. The structure of a proposal is heavily influenced by the topic you have selected, and the research you have already conducted. However, good research proposals typically include the following elements:

• An introduction that specifies the purpose of the research, including:
  • Motivation for the research. Why is this topic important?
  • Problems to be overcome. What is wrong with the current state of affairs? If you are following the common approach of publishing several papers that you will eventually combine into a Ph.D. thesis, then these problems will be the ones that your papers will solve.
  • A high-level description of your proposed research. What do you plan to do to address the aforementioned problems?
  • An overview of the structure of the document.
• A problem statement. This is a one- or two-sentence description of the problem(s) that your research will overcome.
  • It should be obvious to the reader that these sentences are your problem statement; consider using distinct paragraph formatting and explicitly mentioning that this is your problem statement.
  • A good way to think about the problem statement is in terms of your (eventual) thesis statement. When your degree is complete, the questions in your problem statement should be directly answered by your thesis statement. Your problem statement might even be a thesis statement in the form of a question. For example, the problem statement "Can we build a web browser with high quality of service while protecting users' metadata from mass surveillance?" is answered by the thesis statement "We can efficiently and practically protect the privacy of web browsing metadata while maintaining a reasonable quality of service".
  • Depending on your topic, the problem statement may appear in your introduction, or much later in the document. In any case, the reader should clearly understand the full meaning of the problem statement when they encounter it.
• Background information. This generally makes up the bulk of the document—possibly even the majority of your pages. The background information is what primarily demonstrates your breadth of knowledge in the chosen area, and its dependence on the research topic is what causes the high variability in the structure of research proposals. You should completely describe each of the problems that you will be researching. Include substantial surveys of related work. Most of what you write here will eventually be copied, and likely expanded, to form the "background" chapter of your Ph.D. thesis.
• Your previous related work. If you have already published results that are related to your topic but are not part of the proposed research (e.g., related work from your Master's thesis), describe those results. These details will help to convince the committee that you are prepared to conduct your proposed research.
• A description of your proposed research. This content should make up a substantial portion of the document. You should answer the following questions:
  • Given the problems that you have identified, what research do you intend to conduct?
  • How do you plan to conduct this research? Provide details of your proposed approach.
  • What work has already been done? You do not need to have new, publishable results at this stage, but you should include any results that you do already have.
• A timeline for your proposed research. Specify when you intend to complete milestones, such as publishing results that answer aspects of your problem statement or completing remaining degree requirements. Since your Comp-II requirement is normally completed within the first two years, it is typical to expect to defend your thesis roughly two years after the research proposal.

You will need to choose a way to structure your research proposal to include all of this information with respect to your chosen topic area. Consider writing all of the section headings in your proposal before writing the actual text; doing so will allow you to plan out the flow of the document while rearranging the structure remains trivial. There are two general approaches to structuring the document that work well, depending on the nature of your proposed research. In the first approach, the sections in the document are structured as they appear in the above list. This approach causes the document to resemble the structure of an academic paper. In the second approach, each major section in the body of the document focuses on a specific project or problem. Each of these sections includes its own subsections describing background information, previous work, and proposed work for the specific topic. This approach causes each of the sections in the document to resemble a miniature academic paper. The second approach works well when your proposed topics are relatively independent (although they should be connected by the overall theme established in your problem statement).

There are some general rules of thumb to follow when writing the document:

• The body of your proposal (not including title pages, the table of contents, or the bibliography) should be exactly 20 pages.
• Use a double column format with 10pt "Times" font and 1-inch margins to achieve a suitable content density.
• When you are describing things that you will do, use the active first person voice (e.g., "I will solve this problem as part of the proposed research"), since the research proposal is a declaration of your individual research intentions to the committee. Note that this is very different from the way that academic papers are normally written, even when they have a sole author.
• Do not be afraid to reuse text that you have previously used in other documents; the content of your research proposal does not need to be new. However, be sure to reference the source of duplicated content to avoid self-plagiarism.

The CS Graduate Office in DC 2599 has a binder containing examples of previous research proposals. This wiki also contains examples of research proposals written by CrySP students . You may be able to find additional examples by asking others in the lab.