Architectural Pattern Recognition using Convolutional Neural Networks

Identifies repeated structural patterns in text across six hierarchical levels: phrase, line, paragraph/list/table/title, chunk, section, and document. Higher levels are built from recurring lower-level patterns, revealing the structural DNA of the content. Outputs results in YAML or JSON for analysis, templating, and validation.