SQL4NN: Validation and expressive querying of models as data
This addresses the need for efficient model analysis in machine learning, though it appears incremental by applying existing database methods to a new context.
The paper tackles the problem of analyzing machine learning models as a form of data by proposing to use relational database systems and SQL for validation and expressive querying, demonstrating their suitability for such tasks.
We consider machine learning models, learned from data, to be an important, intensional, kind of data in themselves. As such, various analysis tasks on models can be thought of as queries over this intensional data, often combined with extensional data such as data for training or validation. We demonstrate that relational database systems and SQL can actually be well suited for many such tasks.