What are the baseline pipelines?

[rszeto]

It would be useful to know/document the baseline pipelines that we are being compared against on the leaderboard.

[rszeto]

I found some vaguely useful information:

• Baseline Pipeline: Describes a high-level approach to getting a baseline on a dataset. This might correspond to EXLine (EXpert baseLINE)?
• TA2 Evaluation Baselines: Lists scores obtained by MIT Lincoln Labs (as well as experts and random forest). Baseline scores match the leaderboard for autoMPG and autoPrice. The actual pipelines themselves are not disclosed.
• Leaderboard: Confirms that the baseline column is indeed generated by MIT Lincoln Labs. The actual pipelines themselves are not disclosed.

[rszeto]

Mitar has told me that the code to run the baseline pipelines are in <dataset>/<dataset>_solution, e.g., /z/mid/D3M/datasets/seed_datasets_unsupervised/1491_one_hundred_plants_margin_clust/1491_one_hundred_plants_margin_clust_solution. I can provide a section in the wiki that describes the baseline algorithm for each dataset of interest.