Jan van Gemert

Doing a MSc thesis at Delft University of Technology (TU Delft) in Computer Vision / Deep learning

I try to give MSc students a sense of doing real, cutting edge research and I have published in top venues based on a MSc thesis.
I exclusively supervise MSc thesis topics on visual data (images, video).

First meeting.
For our first meeting I want you to prepare the following, which we will discuss in turn:

  • Information: I will need your first name, last name, email address, formal thesis starting date, expected graduation date, and student number.
  • Create a git project named 'mscYourName' with 3 folders: 'code' (for your code), 'thesis' (for your latex files), 'notes' (for our agreements). Send an email to eip-ewi at tudelft dot nl to let them create a TU Delft gitlab repository for you, and mention that you are my MSc thesis student.
  • You are resposible for making sure that you graduate on time. Add a text file to the 'notes' folder on the git with the name: "planning.txt" with a planning time line (give the expected dates) of when required forms, committee meetings, milestones, green light moments, etc., are due. Note that I cannot keep track of each step for all students, so, please do not depend on me and plan this for yourself.
  • For your thesis committee graduation form we need 3 faculty members. We need at least 1 member outside of the PRB lab and one more member that can be from PRB. At our first meeting, please suggest 2 candidate faculty members to be on your committee. You can find potential committee members on the "people" sections of the TUD computer science webpages, or it can be faculty members where you enjoyed their courses or have had contact with before.
  • Please write down in your planning that the final submission of the thesis to the exam committee is 10 working days (2 weeks) before your final defense day.
  • Bring all the required forms to the meeting. (For EWI students: TEP and IEP forms?).
  • Show me that you have good grades for: 1. Deep Learning; and 2. Computer Vision by Deep Learning. Additional Machine Learning or visual data processing courses will be appreciated.
  • After we had our first meeting: send an email to S.Peters at tudelft dot nl to let her know your starting date and that you will start your MSc thesis with me.
  • Read my research guidelines.
  • Write your MSc Thesis in the style of an academic article. This means you will need two parts: 1. The article; preferably in double column CVPR-style Latex format and 2. Supplemental material that provides the background of the thesis. Here are some examples. For the writing: follow my writing guidelines. If you submit something to get feedback, use the "review" format that has line numbers.
  • Are you training deep networks? Unfortunately, significant time goes to tuning hyperparameters, please be prepared to follow A Recipe for Training Neural Networks
  • Schedule yourself for a meeting online.
During your thesis work.
  • You are responsible for scheduling meetings and your own progress.
  • Sign up for our Mattermost and join the public "CV-lab-MSCthesis" channel. You can use this channel to get in contact with other MSc students. Writing a thesis is very individual, so please use this opportunity to meet/discuss with other students in a similar situation.
  • I reserved my agenda on Monday afternoon for open MSc thesis meetings where everybody is welcome to attend. If you find the work of other students interesting, please ask questions and join the meeting. Invite your co-supervisor for each meeting and schedule yourself for a slot online.
  • If you have a co-supervisor: Meet them individually each week.
  • Mandatory student meetings: Every 2 weeks we have scheduled MSc student presentations from the PRB section (schedule). You are required to present a few times (follow my presentation guidelines.) and you have to attend at least 10x, if you cannot make it send an email to Saskia Peters (S.Peters@tudelft.nl). The benefit for you is that you can practice your presentation a few times before your final defense and that you get feedback and insight in what others are doing.
  • Every 3 months we have a few hours of PRB poster presentations where the Bio-Informatics lab, Pattern Recognition lab, and Computer Vision lab members present work to each other, followed by drinks. You will be expected to present your work during the Computer Vision posters. The benefit for you is: independent feedback from experts, insight in all research in the group, experience in presenting your topic to others.
  • If you are interested, feel free to join our "coffee talks": Typically every Tue/Wed/Thu in room Ritchy at the 5th floor someone presents a ML/PR/DL/CV paper in 10 minutes.
  • Do not forget to keep everything in your git up to date
  • Whenever you are stuck: First re-read my guidelines.
Defense.
The formal requirements (forms, timeline, green-light moment, etc.) vary per faculty. You are responsible for managing these requirements.

The procedure during the defense is approximately as follows:
  • You give a presentation of around 20 minutes (follow my presentation guidelines.) The presentation time is short on purpose: we wish to assess how well you can extract the essentials of your work.
  • Some questions from the audience.
  • Detailed questions from the committee members.
  • The committee retreats and decides on a weighted grade, based on this matrix.
  • The committee motivates the grade to the candidate privately.
  • The diploma ceremony proceeds in public.
Some questions you may expect during your defense, and which thus should also be answered in your thesis are:
  • What is "scientific" about your work? (ie: what is interesting about it for other researchers; what have we now learned?). Note that improved accuracy is nice, but the scientific interest is the detailed, well-motivated, well-investigated, answer to "why does it improve?".
  • When will your method fail? What assumptions does it make? In which cases will your approach not work, and in which cases will it work well?
  • Did you look at your results? So not just the numbers, but what do you see when you look at the output on individual samples? For some samples the results improve, and for others the results decrease, are there any explainable patterns for the improving/decreasing samples? Can you quantify this?
  • Why did you use this particular evaluation measure? Its nice that others use this, but is this evaluation measure really suitable for what you want to know?
  • In your article you use the term XXX; what does this mean?
  • Your system has several modules; How well does each module perform? Are all modules needed? Can't we remove modele XXX? (ie: you need to do an ablation study to answer these questions)
  • If you would have to start again, with the knowledge you have now, what would you have done differently?

Valid XHTML 1.0 Transitional
Valid CSS!