PhD Proposal

What goes into a proposal?

After a couple of semesters of cs395, you are ready to start putting together your PhD proposal. You should have some solid first results already, and a good idea about what your dissertation might look like. The proposal should present work that, (1) when completed, would be a significant advance, and (2) you can complete within a reasonable time (e.g. 1-3 years).

The proposal document is about 1/3 background, motivation, and related work, 1/3 about what you've done already, and 1/3 about what you will do to complete it. Take a look at the suggested proposal outline, well as the Heilmeier questions, and general suggestions on scientific writing. It is also a good idea to read some of the earlier proposal documents by people in the group; they are all linked from the nn calendar.

As usual, we will start with the outline, and make progress in chunks, i.e. you finish writing a few sections that go together and I'll read and make suggestions. After a few such iterations, you will distributed the proposal to your committee (2 weeks before the proposal talk at least).

Selecting a committee

In addition to your advisor, you need 1-2 people from UTCS and 1-2 people outside of UTCS, and at least four altogether. In selecting committee members, you need to balance two things: having people who can and are willing to help you in your work, i.e. have expertise that complements your area. On the other hand, they will probably write you letters when you graduate, so they should also be well known in your area. It is a very good idea to have some well-known person from another university to be on your committee -- someone whose work you have built on, or someone who knows your work well and has interest in it. This latter point is important; his/her recommendation letter will only be useful if s/he really knows what you are doing and likes it, and is willing to say so in his/her letter. A superficial letter from a well-known person does not help you much.

Proposal talk

Proposal talks at UT are announced UTCS-wide and it may be a good idea to go to one of them to see what the process is like. It is sometimes difficult to find time for the entire committee to participate, but it has become easier now that some of them can participate remotely. The remote participation rules keep changing, but at the moment at least the candidate and the advisor need to be there in person--others can be remote. The proposal talk should be about 40 minutes long; it is perhaps more similar to a JobTalk than to a conference presentation in that you have a wider audience, and you have to spend more time on motivation and significance and (especially) future work. You should also prepare some extra slides, anticipating questions from your audience and committee, and organize them so that you'll find them easily when you are nervous. We will usually schedule a practice presentation at the NN meeting a few days/weeks before the talk; it is usually very useful way to find out which parts of the talk are working well and which need more work.

Last modified: Sun Sep 25 11:37:49 PDT 2022