#### 1. COMPETITION SPONSOR  
This challenge is organized by the [ML+X](https://hub.datascience.wisc.edu/communities/mlx/) community at UW–Madison with support from its sponsors.

#### 2. PRIZES  
This is a non-monetary educational challenge. There are no cash or material prizes.  
Top teams may be invited to present at the Machine Learning Marathon (MLM25) showcase or contribute to open source outputs.

#### 3. TEAM LIMITS  
- Maximum team size: 5 participants  
- Team mergers are allowed as long as the merged team’s total number of submissions remains within the allowed limits (5 submissions/day × days active), and you do not exceed the maximum team size.

#### 4. SUBMISSION INTEGRITY  
- Manual lookup is not allowed for submissions. Many questions could be answered by reading the source papers directly, and it’s fine to sanity-check your pipeline on a smaller scale, but you may not use manual lookup as your “solution.” The point is to design systems that scale across the full benchmark.  
- Submissions must be produced with code. Do not rely on manually prompting ChatGPT (or similar tools) to generate answers. It’s fine to experiment interactively during development, but this cannot be your final approach. The challenge is to build automated systems. If you are not yet familiar with APIs or open-source libraries, use this opportunity to learn — many options exist on Hugging Face and elsewhere.

#### 5. SUBMISSION LIMITS  
- Up to 5 submissions per day  
- At the close of the competition, only your 3 most recent submissions will be carried forward to the private leaderboard. Of those 2, the highest-scoring submission will automatically count as your final result.

#### 6. COMPETITION DATA  
The dataset is for non-commercial use only. It may be used for competition participation and academic research. It may not be redistributed or republished.

#### 7. LICENSE  
Winning teams must release their full solution under an open source license that permits reuse.