Planning:
| Event | From | To | Where |
| Hackathon and partenerships presentation | 01/24/2025 - 6 pm | 01/24/2025 - 7 pm | TBD |
| Hackathon - Submission are opened ! Permanaces with ML experts are held. |
01/24/2025 - 7 pm | 01/26/2025 - 3 pm | PC TDB |
| Watson plateform presentation | 01/25/2025 - 9:30 am | 01/25/2025 - 10:30 am | TBD |
| First judging session - shortlist selection | 01/26/2025 - 3 pm | 01/26/2025 - 4:30 pm | TBD |
| Shortlist 10-min presentations | 01/26/2025 - 4:30 pm | 01/26/2025 - 5:30 pm | TBD |
| Final judging session | 01/26/2025 - 5:30 pm | 01/26/2025 - 6:00 pm | TBD |
| Winners announced | 01/26/2025 - 6:00 pm | 01/26/2025 - 6:30 pm | TBD |
Eligibility:
Students selected by the 'Binet IA'
Project and Submission Requirements
Each submission must include:
- Predictions for test data set 1 in csv format
- Predictions for test data set 2 in csv format
- Code from which submissions originate (jupyter / python / bash)
- Description of the method used (any format)
Prizes
- First place: 1000€ total gift cards (250€ / team member)
- Second place: 400€ total gift cards (100€ / team member)
- Third place: 200€ total gift cards (50€ / team member)
Judging Criteria and Winner Selection
- (60%) Prediction performances on tasks 1, 2 and 3 (20% each)
- (20%) Quality and Novelty of the approach (statitical & chem-info point of view)
- (20%) Code clarity and reproducibility
Participants are to use:
- All or some of provided data in any way they see fit.
- Any programming language.
- Any publicly available non-commercial software packages
- Any statistical modeling approach (ML/ not ML)
Participants are not to use:
-
Any hERG inhibition data outside of the dataset provided by the organizers as-is.
-
Any publicly available models explicitly pre-trained on hERG toxicity data in any manner.
-
Any descriptors or features explicitly pre-fit on the hERG data.