I am often asked if I have a list of topics. I do not, although I do have various things I think ought to be done. I generally think it is better if you suggest a topic in which you are interested, since then you will be more likely to devote the necessary energy to it. Since an MS thesis is not trying to push the frontiers of knowledge, it is not crucial that it be an active research area of mine, though I should have some knowledge of it. (Of course, if it's completely outside my area, you should be looking for a different advisor!) Beyond that, there is room for work in all areas where I'm active; see my regularly updated research issues pages (CS access only) for some ideas.
For committee members, I suggest: (1) faculty who are knowledgable in your research area, be they in this department or others; (2) CS faculty who you know, from classes or in other ways; (3) other CS faculty. We will discuss potential committee members, and then you can contact them to see if they are willing to serve on your committee.
There must be two committee members besides me. We can have additional committee members, but generally the bigger the committee, the harder it is to schedule your defense.
Read Strunk and White's Elements of Style (available at the UTK bookstore and virtually all other bookstores) for sensible recommendations on writing style; the book is short and you can read it in an evening. For writing problems peculiar to academia, consult Mary-Claire Van Leunen's Handbook for Scholars (available in the library and bookstores); I do not agree with all her opinions, but you won't go far wrong by following them.
LaTeX macros for producing theses in a format acceptable to the school are available in ~cs219/thesis. I want all my students to use LaTeX unless there are compelling reasons to use some other word processor. If at all possible the entire thesis should be machine-readable so that it can be made available electronically.
You can give me outlines and drafts of chapters to review as you write them. Usually other committee members will not want to see your thesis until it is in almost finished form, but you should ask them what they would like. You should ensure that each committee member has a nearly final copy of your thesis at least one week before your defense. However, you should ask individual committee members if they want it earlier, since they may be traveling or have other demands on their time (such as reading other theses).
You will have to give a presentation (generally with tranparencies, and demo, if appropriate) lasting perhaps 1/2 hour. This is not much time! You will have to move quickly to present necessary background as well as your results. It's OK to answer information questions along the way, but try not to get bogged down; I will also help things along if it looks like they have got stuck.
After that you will be expected to answer questions from your committee and other attendees. Some questions are for information (i.e., to understand better what you have said); others are to test your knowledge of your topic and how it relates to other areas. Expect to be asked questions such as "What is the significance of this work?" or "What have you learned from this project?" They are not hostile, but are intended to see if you understand your work in a wider context. Therefore, since you have been buried for some time in the minutia of your research and thesis writing, you should spend some time before your defense to consider your work in its broader context, and to discuss it with me and others.
The whole thing should be over in an hour, though they sometimes go on a little longer (which may indicate an interesting topic rather than a problem). I am generally responsible for seeing that it doesn't get bogged down in tangential questions. The committee will ask you to leave while we discuss the results, after which you will be called back in to tell you the results and sign the exam forms. Not infrequently the committee may ask you to make some minor changes in the thesis before it's turned in. (Of course, more extensive changes may be necessary, but we try to get those things fixed before the defense - another good reason to ensure your committee has enough time to read your thesis.)
The purpose of the afternoon exam, so far as I am concerned, is to demonstrate to me that you have an adequate background in your specialty area (e.g., neural information processing, synthetic ethology, etc.). Therefore, if you did not take an afternoon exam in an area relevant to your research (e.g., you changed specialties after passing the qualifier), I may ask you to take an informal specialty exam to demonstrate to me that you have adequate background in the area. I would anticipate that this informal exam would be the same format as the official Afternoon Qualifying Exams.
After we discuss potential committee members, you can contact them to see if they are willing to serve on your committee. A committee has to have three faculty members besides me, at least one of whom must be from another department (not usually a problem in my research areas). (In addition to me, at least two other committee members must be approved by the Graduate Council to direct doctoral research.) There can be additional committee members if that would contribute to your program, but generally, the bigger the committee, the harder it is to schedule committee meetings. I have even been on one doctoral committee where the student had to do his defense twice because there was no time during the semester when all the committee members were in town!
The idea of the Research Proposal is straighforward if you keep in mind the purpose of the Prelim. Typically, in addition to an introduction to the research topic, the proposal has two major parts, a literature review and a research plan.
The literature review proves that you have done your homework; it shows that you are intimately familiar with previous and current research in the area, that you are aware of the significant open problems, and that you have a good critical understanding of the state of the art in their solution. This background work is what enables you to answer later questions such as, "Didn't Smith solve this problem in 1965?" "How is your approach different from Jones', which she reported at last year's neural net conference?" "Didn't Minsky and Papert prove that what you are trying to do is impossible?" Etc.
Depending on your research topic, the literature review might include topics that are not strictly CS. Remember that the purpose is to determine if you have adequate background knowledge to carry out your research competently. So, for example, you might also be expected to show some knowledge in statistics, cognitive science, linguistics, neuroscience, etc., as required for your work.
The literature review shows that you have a good problem; but you don't have a good research proposal unless you have a credible plan for attacking the problem. This should include a definition of the precise problem you plan to attack, the way you plan to attack it, and the way you intend to test your solution. As with any project proposal, you should describe how the problem is broken into smaller problems, and give a tentative schedule. You should also define critical points in the research, where you can decide if you are going in the right direction, and have contingency plans, if things do not work out as expected. Of course, you cannot plan for every possibility; if you could, it wouldn't be research! The point is that you have a reasonably well-thought-out plan.
The Research Proposal will give you practical experience that will be useful later in writing grant proposals.
The oral part of the Prelim, your Proposal Defense, gives your committee a chance to explore these same issues (background knowledge and research plan) in an interactive fashion. Therefore, you can expect technical questions over the areas that your committee believes to be relevant for your research and practical questions about how you plan to conduct your research. I prefer to think of the oral part not as an EXAM (with its adversarial connotations), but as an early project review intended to guide you in your research. Your committee, with its (presumably) greater research experience, or at least greater distance from your project, will try to steer you from unproductive approaches and help you to become a better researcher. Try to think of your committee as advisors, with your ultimate success as their goal, rather than inquisitors bent on your destruction. (And we will try to fulfill your expectations!)
I recall from my own grad student days that the prevailing wisdom was to put off your Prelim as long as possible, after most of the research was already done; "that way you won't promise anything you can't deliver." It still seems to be the prevailing wisdom. However, it's good advice only if you take the view that the Prelim is "yet one more hurtle to be overcome"; then you would certainly want to put it off until you are confident you can pass it. On the other hand, if you view it as a project review with a constructive goal, then it makes sense to do it earlier in the process; then problems can be corrected before you have wasted a lot of time and effort. (In my experience, one common conclusion of the Prelim is that the student is being too ambitious, that is, biting off too big a problem. Your committee may want you to do less work!) So, my recommendation for scheduling your Prelim is: Do it as soon as you're ready, but no sooner.
As usual, the best way to find out what's required is by reading the proposals of students who have successfully passed the Prelim, and by attending public oral examinations.
When in doubt, talk to your committee, especially me!
Send mail to Bruce MacLennan / MacLennan@cs.utk.edu