Project Milestone 4: Final Submission

Now’s the time to finish up the project! All deliverables will be due on May 15 at 11:59PM.

Video Presentation (5 points)

All presentations must be a pre-recorded video – this will allow us to watch the presentations asynchronously while simultaneously leaving you with a nice artifact that you can show off to friends/colleagues/future employers.

• Each video should be about 5 minutes in length.
• Please feel free to use video editing software to cut down and refine your video presentation.
• You are welcome to have multiple team-members trade off speaking in the presentation or have a single presenter.
• This can be on any video hosting platform (YouTube, Vimeo, etc.). Please double check that the video link is visible to the public before submitting. It can be unlisted.

Presentation Grading

Your presentation will be graded on the following criteria.

1. Motivation: The motivation for the project and how it relates to course material is clearly explained.
2. Methods/Demo: Your methods are explained well and in detail / your demo is thorough.
3. Communication: The method and final results are explained in a way the class can understand, without use of acronyms of terms we have not covered in class (unless you define them). The presentation seems polished.
4. Results: The main findings of the project are reported.
5. Lessons Learned: What did you learn from this project and what might you do differently next time? What questions or directions for future work remain?

Demo: Build a novel NLP system

The final deliverable is the demo itself, a video presentation, and a supplemental document. Your demo should be a complete program that can run end-to-end.

Your demo should:

• Be runable by the grader, following the read me.
• Be posted on GitHub (or the repository site of your choice).
• Not be edited after the deadline. We will be grading the last commit pushed before the deadline.

In a separate document called team#-supplement.pdf, provide the

• system diagram, updated to include any changes you made
• LLM Use Statement - Describe exactly how you used LLMs to generate parts of your code (refer to the syllabus for guidance). If you did not use any generative text, please state so in this section.

You no longer need to mention who did what on your team.

What to Submit

Submit the following to Blackboard:

• The link to your repository.
• The link to your video.
• team#-supplement.pdf, the LLM use statement.

Grading

Paper: Investigate an NLP research question

The final deliverable is the paper itself, a video presentation, and a supplemental document. The paper will be a continuation of the same one you’ve been building on.

Here are some example reports from previous classes again to aid you:

• Example Report 1
• Example Report 2
• Example Report 3

Your paper needs to be in a conference template, written in LaTeX.

• You can use Overleaf to write LaTeX papers together. The CSEE department has a subscription to Overleaf now, so you should be able to have multiple collaborators on the same paper.
• If you do not have a conference/workshop in mind, I recommend that you use the template from the Association for Computational Linguistics.

Your paper should include the following sections:

Your paper should have the following components, in order:

1. Title: What are you calling your study?
2. Abstract: A high-level overview of the paper and what people should take away from it if they aren’t reading the whole thing.
3. Introduction: What research question(s) or problem are you trying to solve? (Start numbering sections from here.)
   • Why this it worth researching?
   • What contributions does this work make? (i.e., What are you doing that’s novel?) You might have to compare to other research.
4. Related Work: What previous research has been done on this research question?
   • A narrative of 8-10+ related research papers, including how each of them is relevant to this work.
5. Methods: What are you doing or making to address your research question(s)?
   • If you are using any datasets, does the needed data already exist? If so, how much data is available? What does the data look like? If you are just using a few examples for few-shot prompting, how did you select those?
   • Describe your solution to the problem/research question. What model(s) are you using? If you are using few- or zero-shot prompting, what do your prompts look like?
6. Evaluation: Renamed from Experiments section What experiments will you be running?
   • What are you measuring and why? What evaluation metrics do you plan to use to answer your research question(s)?
   • What are your proposed baselines? (That is, what are you comparing your method against?)
   • How do these experiments answer your research questions?
7. Results: What did you find from your evaluation?
   • This should be more than just a list of raw numbers. You should explain to the reader how to interpret any graphs or tables you have and describe any trends that you see.
8. Discussion: What insights can you take away from the results? What do the results mean for your research question?
9. Conclusion: In CS, this is often just a brief paragraph reminding people what the highlights of your paper were.
10. References: Papers are cited correctly in the conference’s format and are references in the main text. (Do not number this section.)

In a separate document called team#-supplement.pdf, provide the LLM Use Statement: Describe exactly how you used LLMs to generate parts of your paper (refer to the syllabus for guidance). If you did not use any generative text, please state so in this section.

What to Submit

Submit the following to Blackboard:

• team#-finaldraft.pdf which is your final paper.
• A link to your presentation video.
• team#-supplement.pdf, the LLM use statement.

Grading