structured extraction
The process of automatically extracting structured data (entities, relationships, facts) from unstructured input (text, speech) using NLP or LLM techniques, converting raw input into a standard format like subject-verb-object triples
As defined in Citation Needed: Provenance for LLM-Built Knowledge Graphs — Daniel Chalef, Zep AI
Where it comes up · 1
Most important claims · 3 of 3Search all claims for “structured extraction” →
- In Graffiti and Zep, structured extraction works by extracting entities, relationships between them, and candidate facts as materialized triples (subject-verb-object structured as 'fact').Citation Needed: Provenance for LLM-Built Knowledge Graphs — Daniel Chalef, Zep AI
- The add_episode method in Graffiti runs a complicated pipeline on episode ingestion with multiple high-level steps: structured extraction (entities and facts), deduplication and deconfliction of entities and facts, and mutation tracking (invalidating facts when contradicted by new data).Citation Needed: Provenance for LLM-Built Knowledge Graphs — Daniel Chalef, Zep AI
- After structured extraction, Graffiti runs a deduplication and deconfliction process that deduplicates entities and deconflicts facts—identifying when new facts invalidate or mutate existing facts in the graph.Citation Needed: Provenance for LLM-Built Knowledge Graphs — Daniel Chalef, Zep AI