ingestor

Duplicate content detection computes a SHA-256 hash for each input document and checks it against existing entries. This hashing mechanism prevents redundant material from entering the cognitive ingestion pipeline and cluttering the vault.

Cognitive ingestion noise filtering evaluates incoming documents by estimating their content gain and risk. It classifies material as Cognitive Content, New Skill, or Hybrid to ensure only high-value knowledge passes through the pipeline.

An ingestion decision matrix is a generated file that reports the estimated gain, risk, and classification of a document. You need it when processing cognitive content to evaluate whether material should be approved for repository entry.

Yes, upon approval at the human gate, the pipeline performs an automatic handoff to Stage 2 (Librarian) and Stage 3 (Scientist). This ensures approved documents flow seamlessly into downstream cognitive processing.

Yes, cognitive ingestion requires a human gate. After generating the decision matrix with value deltas and classifications, the pipeline awaits explicit human approval before allowing content into the repository.

Low-value documents are blocked to prevent noise and redundancy in the vault. The ingestor estimates a content gain delta and filters out material that does not provide a meaningful knowledge addition to the repository.