This course has one project: a single, collaboratively written review paper on Foundation Models for Biomedical Data.

There are no separate teams and no independent projects. All enrolled students are collaborators on the same paper, which will be written incrementally throughout the semester using Manubot on GitHub.

The project accounts for 30% of your final grade, but it is not graded as a single submission. Instead, credit is distributed across writing contributions, paper-driven presentations, notes, peer review, and professional engagement, as described on the logistics page.

This project is intentionally modeled after how review papers are written in research groups:

  • shared ownership,
  • incremental synthesis,
  • frequent revision,
  • and public-facing standards.

Writing Platform and Infrastructure

  • The paper will be written using Manubot.
  • All source files, citations, issues, and revisions live in a shared GitHub repository.
  • Contributions are tracked via commits, pull requests, and comments.

You are not expected to know Manubot in advance. We will establish a minimal, reproducible workflow early in the semester.

Paper Scope and Outline

The paper follows a shared outline agreed upon by the class:

  1. Outline TBD
  2. Outline TBD
  3. Outline TBD
  4. Outline TBD
  5. Outline TBD

Each student takes primary responsibility for one major section, while contributing to others through discussion, editing, and review.

Primary responsibility means owning coherence, not writing in isolation.

How the Paper Is Built (Important)

The paper is developed through normal class activity, not through separate project checkpoints.

Progress happens via:

  • Assigned paper readings
  • Student-led paper presentations
  • Structured class discussion
  • Shared class notes
  • Section-level drafting and revision

There is no standalone “project presentation” and no final sprint at the end of the semester.

Project Milestones

Dates will be announced in class. These milestones are shared by the whole class.

Roughly, we will have:

Section Ownership and Outline (Early Semester)

Each student will:

  • Claim ownership of one major section
  • Draft a subsection outline for that section
  • Identify the core papers that define its structure

This establishes responsibility while leaving room for refinement.

Section Drafts (Mid-Semester)

Each student will produce a substantive draft of their owned section.

Expectations

  • Real prose, not notes or bullet points
  • Clear synthesis of the literature
  • Accurate technical framing
  • Proper citations using Manubot

Target length will vary by section, but expect roughly 4–6 pages worth of content at this stage.

Drafts are expected to change significantly after feedback.

Final Revised Paper (End of Semester)

The final deliverable is the entire revised paper, incorporating:

  • Peer feedback
  • Instructor feedback
  • Cross-section consistency checks

Each student is expected to:

  • Substantially revise their own section
  • Contribute edits to other sections
  • Help ensure the paper reads as a single coherent document

Evaluation Philosophy

This project is evaluated holistically.

You are assessed on:

  • Quality of synthesis and judgment
  • Clarity and precision of writing
  • Responsiveness to feedback
  • Reliability as a collaborator on a shared research artifact

Originality comes from framing and synthesis, not from building new models or running large experiments.

Authorship and Public Release

The paper is intended to be public-facing.

  • All students who make substantive contributions will be included as authors.
  • Authorship order and naming conventions will be discussed transparently.
  • Students may opt out of public attribution if desired.