Abstract: A sheet music search and discovery system is disclosed that uses specific mathematical rules to analyze and characterize sheet music and provides functionality for users to identify sheet music based on those characterizations. The system stores sheet music data and metadata characterizing each composition, provides a graphical user interface that provides functionality for users to search the sheet music data for compositions, and generates search results based at least in part on the metadata characterizing each composition. In one embodiment, metadata describing structured sheet music data is generated using a global vector space that includes semantic representations of elements extracted from a large corpus. In another embodiment, metadata describing unstructured sheet music data is generated using machine learning-based pattern recognition.

Abstract: A sheet music search and discovery system is disclosed that uses specific mathematical rules to analyze and characterize sheet music and provides functionality for users to identify sheet music based on those characterizations. The system stores sheet music data and metadata characterizing each composition, provides a graphical user interface that provides functionality for users to search the sheet music data for compositions, and generates search results based at least in part on the metadata characterizing each composition. In one embodiment, metadata describing structured sheet music data is generated using a global vector space that includes semantic representations of elements extracted from a large corpus. In another embodiment, metadata describing unstructured sheet music data is generated using machine learning-based pattern recognition.