Course Project

The course project is an open-ended, hands-on opportunity to explore an aspect of statistical NLP. Examples include ...

• Summarize your participation in a Kaggle competition or shared task with an upcoming deadline
• Compare the capabilities of several open source generative decoders-only transformers (ex. GPT-style models) on a particular task
• Design and create a synthetic dataset, and demonstrate of how it can help to address a particular gap in capabilities
• Fine-tune a multimodal LLM for a new task (or language) and report its performance

Rubric

Your course project proposal will be assessed on a Superior/Pass/Fail basis.

Superior (3 pts)

Pass criteria + all of the following:

• project is appropriately scoped to the course (i.e., not overly ambitious)
• rough timeline (with dates)

Pass (2 pts)

All of the following:

• pull request submitted to course blog repository by the deadline
• clear outline of the goals of your final project
• project proposal investigates some aspect of statistical NLP
• post includes a link to a course project code repository
• the code repository uses the required assignment template (i.e., the repository is owned by the appropriate GitHub organization)

NOTE: Missing one or more of the above will result in a 1 point deduction.

Fail (0 pts)

Submission is empty or altogether absent