Implement Knowledge Mining
& Information Extraction
Knowledge Mining covers finding, extracting, and indexing information from large unstructured datasets. This part covers Azure AI Search (indexing, skillsets, semantic ranking, vector search), Azure AI Document Intelligence (extracting structured data from forms and documents), and Content Understanding (knowledge graphs, entity extraction at scale).
Knowledge Mining is about making unstructured data searchable, extractable, and useful. Azure AI Search indexes content from any source and adds AI enrichment via skillsets. Azure AI Document Intelligence (formerly Form Recognizer) extracts structured fields from scanned forms, invoices, receipts, and custom documents. Content Understanding builds knowledge graphs and cross-document entity relationships.
Topics in this Part
Study in order — Topic 1 (AI Search) is the foundation that Topics 2 and 3 build on.
Azure's enterprise search platform. Covers indexes, indexers, skillsets (AI enrichment pipeline), semantic ranking, vector search, hybrid search with RRF, and the full indexing pipeline. Includes Python code for creating indexes, uploading documents, and querying with filters.
Formerly Form Recognizer — extracts structured data from scanned documents. Covers prebuilt models (invoice, receipt, ID, business card, tax form), custom models (template and neural), the Read model for OCR, and the Layout model for tables and structure. Python code with every line explained.
Azure AI Content Understanding — the newest addition to the knowledge mining domain. Covers multimodal content extraction, knowledge graph concepts, entity linking across documents, and how AI Search skillsets connect to Content Understanding for end-to-end knowledge mining pipelines.