VIPP Guidelines

The following are a brief summary of the guidelines to VIPP student at the music intelligence lab. The vertically integrated Project Program (VIPP) is not a 'typical' course; it's more of an applied research position. Before you get started, here's the general approach you should keep in mind:

  • Learn by building. Start from the data and the task, and learn the theory and tools as you go.
  • Work in public (within the lab). Share progress, links (on Slack), and code (on github) so others can learn with you.
  • Document everything. Good notes today become your portfolio and our collective memory tomorrow.

Here are some more tips and requirements for the VIPP course, regardless of your level.

‍

Weekly rhythm & deliverables

You'll have weekly meetings and working sessions: Monday afternoons in the lab. Before the meeting (due Mondays by 12:00 PM): post a 1-page PDF progress brief in the VIPP Slack channel AND in your project's channel.

Your 1-pager should include:

  1. Done: What you actually built/tested/read (results, screenshots, links).
  2. Learned: Concepts/tools you picked up; what surprised you.
  3. Blocked: Issues you hit; what you tried.
  4. Next: Your plan for the coming week (clear, small, testable steps).
  5. Links: GitHub commits/branches, papers, datasets, notebooks, demos.

After the meeting (same day), Post your finalized to-dos for the week on #vipp. During the week:

  • Work your plan, commit code, and share notable finds (papers, repos, demos) in Slack. It's your responsibility to keep the channel active and your team informed.
  • Check Slack once each morning and reply within 24 hours when someone tags you.

‍

Time expectations (credits vs. learning)

  • A VIPP course typically implies at least 3 hours/credit/week of focused project work. It's preferable if you do it in the lab, to get help from your supervisors.
  • Plan extra time for background learning (e.g., Python/ML/music theory) so your project hours stay productive.
  • Learn while building: start from the end goal, prototype something small, and deepen your understanding as you iterate.

‍

Using AI tools (responsibly)

We encourage using language models to accelerate learning, but you must understand what you submit.

  • Probe & verify: Ask the model why, show the math, cite sources, explain like I'm new to X. You're responsible for what you build and aren't exempt from understanding it.
  • Validate: Run the code, write minimal tests, and sanity-check outputs.
  • Be explainable: You should be able to walk us through your code, decisions, and results.
  • Credit tools: If AI assisted, note that in your 1-pager ("Assistance: ChatGPT + prompts…"). If you find an effective way to work with AI, share it with others. Be explicit when you aren't why something works or doesn't.

‍

Tools & onboarding checklist

Core tools

  • Python (or language relevant to your project), Jupyter/VS Code, Git & GitHub.
  • DAWs / music tools as needed (e.g., Ableton). We’ll announce access/workshops and confirm availability.
  • Slack (VIPP channel); make sure you're on Slack ASAP, and turn on notifications (potentially email digests)

Make sure to do the following for onboarding:

  1. Join Slack; set notifications for the VIPP channel.
  2. Set up Git/GitHub; create a private repo under lab guidance, with a clear README.md. Or join existing lab project repo.
  3. Verify you have Python set up and working, it can run notebooks, and can install dependencies.
  4. Gather background materials (papers, docs, datasets) specific to your project domain (e.g., Maqam theory if relevant).
  5. Push a hello-world prototype (even tiny): a toy dataset load, a plot, a stub class, a small evaluation. Anything concrete.

‍

Documentation standards (the "three D’s")

  • Daily: Keep lightweight notes (what you tried, commands used, links).
  • Dev: Commit often, with meaningful messages; include a simple ``How to run in README.md``.
  • Digest: Your Monday 1-page PDF is the weekly digest. Keep it visual and link-rich.

For your weekly notes and updates, here's what "good documentation" looks like

  • Short, skimmable, and reproducible.
  • Clear inputs/outputs, versions, and settings.
  • Links to papers, repos, datasets, and commits.
  • If you share a link in Slack, add context: what does it matter? who can benefit from it?.

‍

Working style (engineering mindset)

  • Start small. Reduce scope; prove a tiny piece works today.
  • Test early. Can you run that repo? Can you replicate a figure on a toy dataset?
  • Iterate. Build → test → reflect → refine.
  • Make it visible. Screenshots, short clips, plotted metrics—evidence beats claims.

‍

Collaboration & culture

  • We’re interdisciplinary. Expect to collaborate with musicians, artists, dancers, and peers from other fields.
  • Communicate generously. Ask concise questions; show what you tried; be responsive on Slack and by email.
  • Be a good citizen. Review pull requests (when you are required to) kindly, share credit, cite sources, and respect licenses/data privacy.
  • Have fun. Curiosity and play are the core part of the craft.

‍

Deliverables & milestones

  • Weekly: 1-page PDF progress brief + Slack to-dos + active repo.
  • Mid-term check-in: short demo/presentation showing your progress on a simplified task towards the middle of the semester.
  • Final deliverable: working prototype/demonstration and a concise write-up (or slide deck) that others can reproduce.
  • Optional but recommended: lightning talk in a lab showcase whenever you have interesting findings to share with the group.

Evaluation focuses on consistency, clarity of documentation, quality of collaboration, technical depth appropriate to your credit load, and a working demo by the end.

‍

Getting started: your first task

  1. Pick a project thread (or affirm the one assigned).
  2. Draft a one-paragraph goal for the next two weeks (problem, small target, evidence of success).
  3. Produce a tiny prototype this week and report it in your first 1-pager on Monday.

‍

Finally, this is our first VIPP iteration; we’ll improve as we go. If you have any feedback on your experience in the lab, do not hesitate to let us know. Any updates will be posted in Slack and reflected here. If you’re unsure about anything, ask early. We’re here to help you learn and produce meaningful work. Enjoy the process!