Corpus-scoped annotation and analysis

This invention describes a system for analyzing large collections of documents. It works by automatically creating detailed notes and labels for these documents, then uses these initial notes to decide if more notes are needed for other documents, based on how they relate to each other. All this information is stored in a database, allowing applications to ask complex questions across the entire document collection.