Judging Criteria

Groups will give a presentation 5 min + 2 min Q&A of their developmetin the hackathon. It is important to highlight good data practices used throughout the event, among them:

  • Challenges with the data

    • Data cleaning
    • Data pre-processing

  • Molecular Representation

    • Effects and limitations of used representations

  • Model design and understanding

    • Data splitting / Cross-validation
    • Simple baselines
    • Decision process for model selection
    • Hyperparameter tunning

  • Discoveries from the data and modeling

    • General insights from the hackathon
    • Is there enough data?

  • What could be improved

    • How to improve models

       

Winner Selection 

The winning group will selected by our pannel of judges, composed of specialists in data science, and green chemistry. The model from the winning group will be evaluated on a separated test-set gathered by the organisers.