The Caltech Mathathon is a math hackathon run by a registered student organization at Caltech, with faculty advisors Sergei Gukov, Terence Tao, and Tony Yue Yu. It is centered on one question: how can AI augment human understanding of mathematics? Participants pick a problem with an unintuitive solution (examples include the four color theorem, the ABC conjecture, and the Navier-Stokes problem) and develop an alternative proof or exposition. The organizers also encourage generalizations, new notations and techniques, and connections to other problems.
Format:
- Round 1 (onsite, 40 hours): 100 teams make preliminary progress, then give live talks before judges and other contestants.
- Round 2 (remote, 2 months): teams produce an online explainer (a draft paper, video, interactive game, or other form) and a GitHub repository holding the materials used to make it, including LLM chat histories and code.
Participants may take part in Round 1 only.
AI policy:
AI is treated as one of many tools. Proprietary models are allowed, and open-source tools are encouraged; the organizers partner with SAIR to provide compute, host workshops, and offer prizes to teams that use open-source models.
Round 1 dates:
| Date | Activity |
|---|---|
| November 13 | Round 1 begins (40-hour onsite round) |
| November 14 | Ongoing work |
| November 15 | Live talks before judges and other contestants |
Location: California Institute of Technology, Pasadena, CA.