Dataset discovery is increasingly supported or even led by AI tools. However, recommendations are lacking on what metadata fields should be prioritized to best support this discovery method. Additionally, guidelines are also missing on what to include in those fields (e.g., the description field). These recommendations and guidelines should be based on research into what changes in a dataset (or other) metadata record increase the likelihood of a given dataset being suggested by an AI tool, and across multiple AI tools of various types. The NASA science data repositories in Heliophysics are interested in this work.
Dataset discovery is increasingly supported or even led by AI tools. However, recommendations are lacking on what metadata fields should be prioritized to best support this discovery method. Additionally, guidelines are also missing on what to include in those fields (e.g., the description field). These recommendations and guidelines should be based on research into what changes in a dataset (or other) metadata record increase the likelihood of a given dataset being suggested by an AI tool, and across multiple AI tools of various types. The NASA science data repositories in Heliophysics are interested in this work.