The goal of the project in this course is for you to independently explore data privacy, find a topic that excites you, and dive deeper into it. Through this process, you will learn how to conduct a comprehensive review, manage your time effectively, and write detailed reports.
You are free to use any format for your project deliverables. Whether it is in LaTeX, Google Docs, or any other medium, the key is to convey your findings effectively. There are three types of projects you can choose from:
If you choose a research project and invest a lot of effort into organizing and analyzing existing methods—even if your final result is a minor new contribution—that is completely acceptable for this course. Similarly, finding new insights during an evaluation project is highly valued. Please make sure to clearly outline your contributions from both perspectives.
You can work individually or in a group of up to two people, but note that the expected amount of work will double for a group project. Group members will receive the same grade. There is no preference from my side regarding project format, topic, or whether you work alone or in a group.
Some ideas (and you are very welcome to come up with your own idea):
A successful project will include a well-thought-out proposal (your topic may evolve, but the more planning you do early on, the better), a comprehensive survey of related work, and thorough execution. There are three key milestones/deliverables for the project: Proposal, Poster, and Final Report.
In a company setting, you often need to propose your ideas before receiving resources (e.g., budget, headcounts); in a research setting, writing a proposal (e.g., internal or NSF proposals) is key for your research agenda. A well-structured proposal convinces stakeholders that your idea is worth pursuing.
Objective
Expectations
Whether at a tech/academic conference or internal company review, posters are a common way to present mid-project progress and foster discussion. They allow you to share insights visually and gather feedback for further development. At the end of this class, we will have students present their projects via posters. We will have two rounds, where students are assigned to stay with their posters and present (if two students work together on a project, both students should engage in the presentation), or walk around listening to others.
Objective
Expectations
A final report or publication is the documentation you hand off to a client, the final write-up for a grant-funded project, or the paper you publish after your hard work. It should hopefully adhere to the proposal, but if it deviates, it’s also understandable (plans are always subject to change, but it’s still useful if we think carefully in the proposal phase). In most cases, there will be an evolving draft, then content from a version of the draft is extracted to form the poster, and the draft can be finalized after incorporating feedbacks from the poster session and final touches.
Objective
Expectations
The overall project grade is divided into three major milestones:
Below is a detailed rubric for each component. Students should contact instructor/TAs periodically to ensure the project is related to data privacy. There are also some opportunities to earn bonus points at the end.
Criteria | Points | Description |
---|---|---|
Objectives | 2 | Clearly state the project goal(s) with convincing motivations (background, big picture, significance). |
Related Work | 2 | Literature review (cite key papers). |
Proposed Work | 3 | Outline of proposed tasks (typically there are two or three aspects under one theme) and methods. |
Preliminary Work | 3 | Show evidence of initial exploration (this can go before the actual proposed tasks and serve as a motivation). |
Scope | 1 | The problems are not too broad nor too narrow for the course timeline. |
Evaluation Plan | 3 | Describe how you will measure success (datasets, baselines for comparisons, metrics). |
Risk Mitigation Plan | 2 | Identify potential issues (e.g., data/resource availability, technical hurdles) and strategies to address them. |
Timeline & Approach | 2 | Provide a clear work plan or schedule, highlighting major milestones and how each step will be executed. |
Criteria | Points | Description |
---|---|---|
Content Coverage | 3 | Cover project background, motivation, and methods; present key findings or progress made. |
Clarity & Logical Flow | 2 | While covering key aspects, the poster should also be concise (don’t include much text/figures), with a logical structure, headings, and transitions. |
Poster Visuals | 3 | Effective and self-explanable diagrams, charts, and color schemes. |
Delivery | 2 | Clear and engaging communication (voice, pace, eye contact), minimal reading from poster or transcript. |
Handling Questions | 2 | Answer audience questions and foster discussion where appropriate. |
Criteria | Points | Description |
---|---|---|
Writing | 3 | Clearly state the project goal(s) with convincing motivations (background, big picture, significance). The flow is logical and easy to follow. Minimizes grammatical or stylistic errors, ensuring readability throughout the document. You can reuse materials from the proposal. |
Execution | 3 | Demonstrates a careful and methodical approach— properly setting up experiments and comprehensive experiments/comparisons with sufficient ablation studies. |
Results Presentation | 1 | Clearly displays findings using tables, charts, or figures (should be a superset of those from the poster). |
Results Interpretation | 1 | Discusses observed patterns, significance, or unexpected outcomes. |
Discussion | 1 | Reflects on key insights and challenges. Evaluates how well the project met its stated objectives and why. |
Citation & References | 1 | Uses consistent citation style, as required in the proposal. All external sources are properly referenced and included in the bibliography. |
These bonus points apply only within the project category. That is, your total project score cannot exceed the maximum 40% allocated to the project in the course.
LLMs Bonus (+2 points)
If your project is related to large language models (LLMs), you can earn 2 bonus points.
Note: This area is challenging because LLM research evolves rapidly and often requires substantial computational resources.
Popular Poster Vote (+2 points)
During the poster session, each student will select two of their favorite posters (they cannot vote for themselves). The two posters that receive the most votes will each earn 2 bonus points. Votes can be based on any criteria—novelty, clarity, or overall impression.
Distinguished Proposal and Project Awards (+2 points)
The instructor will select a few Distinguished Posters and Distinguished Projects based on overall quality, innovation, and impact.
Research, Evaluation, or Application
If you feel confused about the distinction about these three types, you’re on the right track! Indeed, they have a big overlap and are closely connected. A good project always starts from a comprehensive literature review, and some initial benchmarking. Later, if you could find a new idea (sometimes a new idea naturally appears if the benchmarking is comprehensive), you could propose and ‘sell’ this idea (becomes a ‘research’ project). If not, you could make the benchmark more comprehensive (‘evaluation’) or apply these methods to a new domain (‘application’).