Add additional information (e.g. data packages or surrogates) to problems

[miguelgondu]

Examples of potential data packages:

• Unlabeled data,
• Labeled data,
• Available pre-trained surrogate models.
• Available pre-trained latent variable models.

[miguelgondu]

How should we handle the data downloading? Idea: persist databases and download them on-the-fly for the user.