Recommended tags and user_attributes for competition submission

[agitter]

The Competition API submit_predictions takes optional tags and user_attributes. Are there recommended values to use for these like the PyPI classifiers?

I saw some examples in the ASAP Discovery competition .

[cwognum]

Hey @agitter , that's a good question! I'll ping @hmacdope for his thoughts: Are any specific research questions ASAP would like to answer for which we could recommend some tags or user attributes to help classify the submissions?

[hmacdope]

I would say anything you think would be useful to classify your submission. Best way to do this in my opinion would be to describe some general keywords around the model or modelling approach you used, capturing both architecture and specific tooling. For example graph-neural-network, docking, co-folding, alphafold, vina, scikit-learn would be some things I could imagine in a word cloud describing the competition.

Good question though

[agitter]

Thanks for sharing your thoughts. I now have a good idea of what tags and attributes would be appropriate for the submission I have in mind. As a more general suggestion, having examples like this alongside future competitions will continue to guide participants to provide useful submission metadata.