Presentations

Optional feedback round: 4/24

Slides due: 5/2

Goal: Present to a public audience what you’ve done this semester.

Logistics

1. Presenting during final exam period: 5/2 2pm – 5 pm in LSRC A156
2. Turn your slides in as a deliverable to Gradescope by 5/2
3. This presentation will be advertised to the department via email. The email will include your title, name, and abstract.
4. You are welcome to invite anyone to the presentation.
5. Your presentation should be 15-20 minutes long.
6. There will be 5 minutes for Q&A.
7. The presentation order will be determined by the wheel of names on 4/22.

Requirements

The following is the information you must include in your presentation. It does not have to be in this order (except the first and last slide).

1. Title slide with names – This should be your first slide.
2. Motivation + Research Questions
3. Background information
   1. Not necessarily all of the related work cited in your research paper.
   2. The goal is to familiarize your audience with what they need to know to understand the rest of your work. You may assume your audience is at least a 3rd-year computer science major and therefore knows terms such as introductory programming, office hours, generative AI, etc.
4. Context and Data
5. Methods
6. Results
7. (optional) Limitations
8. Conclusions or Key Takeaway slide – This should be your last slide. Do not end with a slide with only the text, “Questions?” Give your audience a slide with key information on it to help them remember your talk as they think of questions to ask.

The deliverable to Gradescope should be a pdf of your slides or a pdf of the slides and speaker notes (so Prof. Stephens-Martinez also has a sense of what you plan to say).

Optional: Feedback Round due Thursday, 4/24

If you wish to receive feedback on your slides before the presentation, you must submit a pdf of the slides on Gradescope by Thursday, 4/24. If you are using some online platform that has comment capabilities and would rather receive feedback that way, do the following:

1. Share it with ksm@cs.duke.edu, or make it world comment-able.
2. Put the URL on your title slide.
3. State on your title slide that you want feedback on the platform and not in Gradescope.

Note: This is the day after the report is due. That is by design to force you to do a very rough slide deck without spending a lot of time on it before you get feedback. This way you will also get feedback by the end of Friday 4/25.

Recommendations

1. Your last slide should have useful information to remind your audience of the key information in your talk. Do not have it only say “Questions?”
2. Your motivation, research questions, and related work are sometimes mixed together into a single set of slides because they are what your talk needs to argue why your project was interesting, reasonable to do, and novel.
3. While it is tempting, your talk is not a copy and paste of your paper. Instead, start with the key results you want to convey and work backward:
   1. What methods does the audience need to know to understand and believe the results?
   2. What context and data do they need to know to understand and believe the results?
   3. What background and related work does the audience need to:
      1. Understand what I did and why I did it?
      2. Understand my context and the data I have?
4. Do not have distracting subtle movement on your slides.
   1. Animations are fine, especially if they help draw the eye to what you want the audience to focus on.
   2. A GIF as the background where some graphic is “blowing in the wind or something” is not fine. It is distracting to a general audience and especially to neurodivergent audience members.
5. Unless you have a good sense of color theory, it is okay to pick a simple color palette/slide template and just use it. You will be graded on legibility, not the prettiness of your slides.

Grading

• Exemplary – Above and beyond Satisfactory, such as a talk that flows well with clear logical progression.
• Satisfactory – The talk fulfills all of the above criteria.
• Not Yet – The talk is missing one part.
• Unassessable – There is a talk, but it does not fulfill the Not Yet criteria.