MUMBAI, India, Aug. 31 -- Intellectual Property India has published a patent application (202611007562 A) filed by Netaji Subhas University Of Technology,dwarka,delhi; and Vijyant Agarwal on January 27, 2026, for Distributed Hierarchical Ai-Assisted Agent-Governed Platform With Edge– Central Orchestration, Ai Mentor Agents, Closed-Loop Policy Traceability, And Interoperable Service Generation For Public Universities.
Inventor includes Vijyant Agarwal.
The application for the patent was published on August 28, 2026, under issue no. 35/2026.
Abstract: The invention relates to a computer-implemented method and system for hierarchical AI-assisted, agent-governed development of interoperable information services for public universities. Workflow specifications are decomposed into hierarchical tasks executed by specialized agents under governance approval gates. A policy corpus is ingested and versioned with effective dates, and a traceability knowledge graph links requirements and data elements to authoritative policy clauses with evidence spans. An AI-assisted relevance engine proposes candidate mappings with confidence values and explanation metadata, while a governance agent restricts acceptance through escalation and approvals to prevent autonomous policy decisions. Interoperability artifacts including versioned APIs, events, schemas, and adapters are generated, together with verification artifacts and runtime validators. AI mentor agents instantiated as machine learning models including transformer-based LLMs fine-tuned on domain-specific data provide role-bound interactive guidance, with latency-sensitive inference deployed at edge nodes to reduce interactive latency, including illustrative minimum time inference targets on supported mobile GPUs under optimized model settings. An edge–central orchestration engine optimizes placement of mentor inference, generation, verification, and monitoring tasks across edge and central compute based on telemetry and governance constraints. A policy-timeline synchronization engine aligns policy effective dates with deployed rule configurations and evidence snapshots, enabling temporal correctness and auditable reconstruction. A closed-loop feedback layer uses runtime enforcement logs, audit outcomes, and user corrections to trigger controlled regeneration, recalibration, and optional model updates under governance approvals. A responsible audit layer records mappings, approvals, explanations, synchronization parameters, orchestration decisions, verification results, and deployment provenance. The system provides technical effects of improved interoperability, traceability, reproducible deployment, runtime conformance enforcement, policy-aligned temporal correctness, low- latency interactive guidance via edge inference, bandwidth efficiency via distributed processing, and controlled continual improvement via closed-loop feedback.
Disclaimer: Curated by HT Syndication.
