How to discover relevant metadata for data management?
Science / Computer Science
Metadata is descriptive data about datasets. They can be simple, such as attribute names and statistics, or complex, such as semantic restrictions between different records. Data management and science provide compelling applications to users, but these applications depend on complex metadata. Most datasets manipulated by scientists include, at most, simple metadata, which limits the potential delivered by applications. Manually identifying this descriptive data is an error-prone task, made difficult by the complexity of modern data and applications. The objective of this project is to develop automated solutions for metadata discovery. To deal with the gigantic space of possible results, we propose to characterize and explore the synergy between data and applications to converge on the space that best assists users in applications of interest.
