We want an exemplary what-if analysis that uses multiple patches per pipeline variant.
One simple example would be to try out different data corruptions and check if different model variants are more robust to the data quality issues than the current model.
We can build this easily by copying & pasting the existing DataCorruption what-if analysis and adding additional model patches.
Description
DataCorruptionwhat-if analysis and adding additional model patches.