Cross-document intelligent authoring and processing, with arbitration for semantically-annotated documents

A method implemented on a computer system executing instructions for processing documents, the method comprising:

• processing a document set that contains a plurality of documents to identify a collection of chunks in the documents and to generate corresponding annotations for the chunks;
• wherein the chunks comprise ranges of characters, tokens and/or non-text objects from the documents;
• the collection of chunks include chunks that reflect a layout, structure and/or formatting of documents (structure chunks) and chunks that reflect semantic roles played within the documents (semantically-significant chunks);
• and the annotations for the chunks comprise locations of the chunks within the documents, and semantic role labels for the semantic roles played by the semantically-significant chunks within the documents;
• arbitrating among the chunks in the collection to produce hierarchies of chunks representing individual documents, the hierarchies comprising both structure chunks and semantically-significant chunks, wherein arbitrating among the chunks in the collection comprises modifying the collection of chunks based on their semantic role labels to reduce partial overlap of semantically-significant chunks within the hierarchies;
• and
• generating representations of the processed documents in a format comprising the hierarchies of chunks and corresponding annotations for the chunks;
• and
• making the representations in the format available for use by any of a plurality of software applications in downstream processes.