Ph.D. Proposal Presentation

A Powerpoint template is available here: Powerpoint Template

The presentations from past students are available in the SharePoint

General guidelines

• Do a dry-run in front of the lab ideally one week before your presentation. They will give you invaluable feedback about your presentation. This is also a great opportunity to practice answering questions about your project.
• Draft a storyboard first to organize your ideas and ensure a logical flow between the sections. Tell a story with your presentation and guide the audience.
• Be sure each slide has a clear take-home message. What should the audience get from each slide? Everything you show should be there for a reason.
• The Ph.D. proposal presentation is not a lecture on your subject. You do not need to say everything in the presentation. Less is more. Stay focused on what is most relevant to your objectives.
• When discussing the related work, be sure to highlight the holes in the literature, and to contrast them to what you wish to accomplish during your Ph.D. What is not possible before your Ph.D. vs what will be possible after.
• Related work can either be discussed in a section of its own, or per objective. Choose the structure that best fits your objectives.
• Be sure to read the presentations from past students. They can give you ideas on how to better structure your presentation.
• The audience read your Ph.D. proposal document and prepared for its content. Try to not add new material that wasn’t in your document to your presentation.
• Due to department rules, you have to do 10 minutes of your presentation in English. To avoid forgetting it, a good trick is to start in English, and transition when changing sections. It is also acceptable to do the full presentation in English.
• The presentation should be between 30 and 45 minutes. You should aim for 40 minutes. A good rule of thumb is 1 minute per slide, so about 40 slides.
• Don’t remove slides you already made, instead move them at the end of your presentation, in an appendix section. They can be a great visual aid when answering questions.
• After your presentation, there will be two rounds of questions from each person in your committee. While this can be quite intimidating, remember that their questions are to help you. You should be able to answer some of the following questions:
  • Why is your research problem important?
  • If a part of your project doesn’t work, what can you do? Could you still publish something? Do you have a contingency plan?
  • How to mitigate the risks in your project?
  • What would be the main contribution of your Ph.D. project?
  • What are the limitations of the techniques you want to use?
  • If you want to build a new dataset, how do you know when to stop gathering data? What dataset size should be enough for your other projects? Gathering data can be quite time-consuming, and it shouldn’t become a time sink preventing you from doing research.
  • What would be the final demo at the end of your Ph.D.project?

Specific Slides Advice

• Read the section on oral presentations in New Students.
• Use the lab’s PowerPoint templates for building your presentation. Do not use LaTeX Beamer (use at your own risk).
• Be sure that the page number is visible on each slide.
• When citing papers, you do not need to put the whole citation. Instead, use the format “Vaswani et al., NeurIPS 2017”. This gives the essential information: who wrote the paper, what conference or journal that paper was published in, and is it recent.
• Do not use too much text on your slide.
• When using bullet points, be sure it makes sense semantically, e.g., do not mix verbs with nouns.
• Aim for approximately 1 minute per slide.
• Do not put too much information in one slide, otherwise it will feel crammed.
• Try to make your slides visually-pleasing. Use PowerPoint tools to center images. Do not just put the images randomly in the slide.
• If you already published a paper, you can add a screenshot of the paper’s title page. It clearly shows you already made a paper. You do not need to present your previously published papers as you’d do at a conference. Give the high level and the main contributions.
• If possible, when presenting the related work, find a way to visually show the hole in the literature you want to fill (e.g., a table, quadrants, etc.)
• Have a slide presenting your research question.
• Have a slide giving an overview of your three objectives.
• Per objective, have a slide giving the main goals.
• When presenting your schedule, be sure to explain the high level of your Gantt (e.g., one project per year).