Planning and Paperwork for Master's Projects and Master's Theses in the Document and Pattern Recognition Lab (DPRL)

Richard Zanibbi (Jan. 26, 2017)

This document is designed to clarify the process by which I accept students into the DPRL lab, and the Computer Science Department deadlines and deliverables that MS project and thesis students need to know about.

Over the past few years, I have had more and more students contacting me about doing projects and theses in the DPRL, unsure about the process is. For those students that I do advise, a number of them have struggled to make the administrative deadlines associated with their project or thesis because they were uncertain what the deadlines were. This is a significant stressor for students, because it has financial implications. It also often means that documents required for registration (e.g, Independent Study forms, thesis proposals, etc.) are completed in a hurry, making them placeholders rather than the useful planning tools that they are intended to be.

Projects and theses that are planned early, and have enough time to fully develop are the ones that generally lead to good results and research publications. There are many things to learn, and to learn them well requires time.

For myself, the ideal advising model has the student taking one or more courses with me, registering for an Independent Study in Fall, and then completing their project or thesis in Spring. Preferably the student has at most one other course during the Spring semester (particularly for thesis students, who often need to do much more work).

Below are guidelines for planning an MS project or MS thesis in the DPRL.

Getting to Know You, a.k.a Take my Course(s)

Most Computer Science faculty at RIT much prefer to advise students that have taken a course with them in the past. It is very hard to start a student on a research problem before having any sense of their interests, programming skills, theoretical knowledge, analytical skills, working habits, level of motivation, and personality. Mismatching these with advisor expectations is generally a recipe for disaster, or at least significant stress and/or frustration.

If you would like to do a project or thesis in the DPRL, take at least one of my courses.

Independent Studies

Students wishing to work in the DPRL should talk to me about doing an Independent Study near the end of the preceding semester (normally, near the end of a course that I am teaching). Independent studies allow you to make a solid start on your project or thesis, and give you some practice in basic research skills before registering for a project or thesis. During an independent study in the DPRL, you will:

For an independent study to be undertaken with my supervision, a student and myself need to identify a specific research problem of mutual interest. Thesedays I avoid independent studies designed to provide a 'general survey' of the literature on a large topic, as this doesn't often lead to satisfying outcomes. Especially for their first research project, things go better when a student narrows their focus, reading about and working on a specific problem. Normally I'm interested in independent studies that expand upon current and/or previous research in the DPRL, although I am willing to consider other problems. A summary of projects, publications, etc. from the DPRL may be found online (see the lab web page).

To register for an Independent Study, students need to complete a form that describes the purpose and milestones of the study (this form is available from the CS web pages). Students should start drafting this at the end of the previous semester and work with me to refine the form contents. The document must be submitted by Monday of the second week of the semester.

MS Project

Students must register for an MS project by Monday of the second week of term. But students wishing to work in the DPRL should discuss this with me at the end of the previous term, so that there is sufficient time for planning and scheduling on my part. I prefer to advise student projects when the student has completed an Independent Study, for the reasons mentioned above.

The first task a project student needs to complete is defining the three milestones for the project (due at the end of each month), along with the format of the final project report and deliverables. Normally, these include a paper tailored for a specific conference format, code, documentation, emails and lab notebooks, and presentations at bi-weekly lab meetings. A 1-page sketch outlining milestones and deliverables should be prepared by Monday of Week 2 and sent to me for planning purposes.

MS Thesis

Thesis students also must register by Monday of the second week of term, but this requires that: 1) a thesis 'pre-proposal' (1-2 pages) be approved by the Graduate Program Co-ordinator, and 2) that all members of the student's committee have approved the thesis proposal (10-15 pages, normally). A pre-proposal is easy to make, and should be sent around much earlier in the previous semester (during the Independent Study). LaTeX templates for creating thesis sketches (which can act as pre-proposals) along with an MSc thesis template are available online here.

Thesis Commitee. DPRL students have three members of their thesis advising and defense committee - myself, their Reader who will read and critique the document carefully, and their Observer who is there to 'referee' at the defense, and who optionally may also read the report and critique the thesis at the defense. Students should talk with me about who they would like on their committee during the previous semester, and before registering for thesis. Faculty are busy, and it is often the case that the student's first choices are unavailable to serve on their committee.

Once the proposal has been written, normally this is converted directly into an initial rough draft for the thesis (in particular, the related work can be largely re-used).

After experiments are completed and the thesis is written up, the student needs to share their thesis draft with their Reader, before their defense. Once the Reader is satisfied that the student is ready to defend, the student needs to schedule their thesis defense (see the CS web pages for guidelines). As a courtesy to your commmitee, a thesis defense should be scheduled at least two weeks after the defense announcement date.

Summary

Doing good work takes time and planning. Research work is no exception.

I am hopeful that this document will help to clarify the preferred path to completing a project or thesis in the DPRL, and make it easier for students to know what deadlines they will need to meet.