MUMBAI, India, Jan. 23 -- Intellectual Property India has published a patent application (202521105476 A) filed by Persistent Systems, Pune, Maharashtra, on Oct. 31, 2025, for 'system and method for ontology extraction, schema synthesis, and evolution from dynamically changing data sources.'
Inventor(s) include Mr. Nitish Shrivastava; and Mr. Pradeep Kumar Sharma.
The application for the patent was published on Jan. 23, under issue no. 04/2026.
According to the abstract released by the Intellectual Property India: "The invention provides a system and method for dynamic ontology extraction, evolution, and validation from enterprise data using Triangulated Ontology Synthesis and Adversarial Falsification. The data streams are ingested from structured, semi-structured, and unstructured sources, and candidate ontological elements are generated by fusing distributional, symbolic, and usage evidence. Each candidate is assigned a probabilistic confidence score using Bayesian fusion and is later validated through adversarial falsification including counterexample generation, schema validation, and logical reasoning. The validated elements are promoted to the canonical ontology, while drift monitoring continuously regenerates candidates to maintain alignment with evolving data. The system exports machine-readable ontology artifacts in JSON Schema, SHACL, and OWL formats and maintains governance through version control, rollback, and auditability. The invention enables automated, robust, and continuously evolving semantic modeling, ensuring accuracy, consistency, and machine-actionable ontologies for downstream Artificial Intelligence, analytics, and Knowledge Graph applications."
Disclaimer: Curated by HT Syndication.
