Project
PAT 463/563: Music and AI (Fall 2025)
Instructions
- Provide proper citations/references for any external resources you use in your writing and code.
- Submit your work to Gradescope.
- All assignments are due at 11:59pm ET on the date specified.
- No late submissions! Submit your work early and update it later.
Topic
This is an open-ended project. You may work on one of the following topics:
- Building a new AI music tool
- Exploring creative & artistic use of AI music tools
- Analyzing systematically existing AI music tools
Project Pitch
Please give a 10-min pitch on:
- Topic: What do you want to work on?
- Topic: Who are the target audience/users/customers/readers?
- Goals: What are your goals?
- Methodology: How are you going to approach it?
- Methodology: What are the tools (programming languages, platforms, plugins, hardware, etc.) that you’ll be using?
- Expected results: What are the expected deliverables (e.g., an instrument, a plugin, a web/mobile app, a standalone software, an installation, a performance, a composition)?
- Planning: What are the timeline & milestones?
Presentation
Please give a 20-min presentation that summarizes your motivations, methods, results, analysis, and discussions. You may follow any structure that best suits your narrative.
Report
Please turn in a 3 to 4-page (excluding references) report that summarizes your motivations, methods, results, analysis, and discussions. You may use any template for your report.
Rubrics
- Presentation (20pt)
- Attendance (10pt)
- Clarity (5pt)
- Organization & presentation (5pt)
- Report (20pt)
- Writing clarity (5pt)
- Organization & presentation (5pt)
- Results (5pt)
- Discussion (5pt)
Compute
Follow the instructions to learn how to access Great Lakes. You will be provided with ~560 GPU hours for assignments and final project.
Suggestions & Tips
- An active GitHub repository with many open/closed issues is usually a good sign.
- Always look for backup codebase so that you have a plan B. Any online repository is not guaranteed to work.
- If you plan to train or finetune something, think about the “data & model” at the same time. You will need the right dataset and the right model to succeed.