There are two deliverables for the final project:
Make sure that all your code is also on github, but you can keep data off github.
During the presentations everyone will fill out a paper feedback form (which will count toward your participation grade). Later on I will anonymize this feedback and send it to each group.
I’ll go through each of these pieces below. When thinking about what to include in your git repo, keep a reproducibility perspective in mind. From your repository, I should be able to reproduce your project and results, subject to random seeds.
Each group will have 12 minutes (+ time for questions and transition to the next group). The best way to make sure you are hitting the right time is to practice.
To make transitions easier, please EMAIL me your slides by midnight the night before your presentation. They must be in PDF format! That way we can use the same computer for all presentations.
In terms of presentation content, you should (very briefly) include all the main components you mentioned in your proposal, as well as future work:
Introduce your topic and goal in a creative or visual way. Whenever you give a presentation, there will be those in the audience less interested in the topic than you are, who might question the “point” of your topic or thesis. Give them a reason to pay attention. Often this involves placing your topic in a larger context, using an image the audience can relate to, telling a personal story, or posing a question you’ll answer later in the talk.
Explain your dataset and chosen methods. Try to pick one detail or aspect that you found interesting or challenging. If you are using methods we’ve talked about in class, you could expand on how you prepared the data. (Make sure to include details about the features/labels.) If you are implementing or using a new method, tie it to our class material and then explain how it is different or novel. Overall, try to give the project a narrative; explain your thought-process throughout the project.
Display your results in a visual way. Negative results are results too, and can definitely be included. How did you evaluate and interpret your results? If they did not match your expectations, what might be going on?
In a few words, what were your main takeaways from the project? What would you do if you had 6 months to work on this project instead of a few weeks? What aspects would you change or extend further?
Speak loudly, to the back of the class.
Avoid text-heavy slides, try to use images and diagrams to convey information.
Include citations for any figures/info you use that you did not create (on the slide where you use it).
You do not need to include a full list of references in the slides.
Make sure group members get roughly equal time to present.
As an audience member, be respectful to the other presenters. Be on time and give them your full attention (and also fill out the feedback form: this counts toward your participation grade).
Each person should ask at least one question to another group (one question total, not one per group). I know we have a large class and a long time block - sometimes keeping a question in mind is a good way to stay engaged.
The writeup should be in the style of a research paper, but it is okay to keep the text brief. Include the parts of the presentation mentioned above as separate sections (you may wish to have separate sections for Data and Methods) and use full sentences. For your results section you should have at least three figures and these should be described and interpreted in the text. In the conclusions and future work section, include thoughts about how to extend your work as well as difficulties encountered and how you overcame them.
For the final project you are allowed to use Github Copilot, ChatGPT, or another code-generating tool of your choice. If you decide to use one of these tools, please reflect on the process in your writeup. Did it speed up your code development process? Were its suggestions generally helpful or unhelpful?
Except for external software and large datasets, include all code that was necessary to obtain your final results. Keep your code organized and commented. You can include some small example datasets, but avoid putting large data files on git since this can cause problems. Err on the side of including more results though (output files, figures, etc).
In the README include a brief description of the environment setup you used, as well as command lines needed to reproduce your analysis and results. Think about the standard of reproducibility when creating your README.
During the presentations everyone will fill out feedback forms for all the other groups. This will consist of two short questions: