MUMBAI, India, April 17 -- Intellectual Property India has published a patent application (202641043189 A) filed by Dayananda Sagar College Of Engineering, Bangalore, Karnataka, on April 4, for 'a system and method for pre-execution non-termination prediction in autonomous ai agent task pipelines.'
Inventor(s) include V. Shriyatha; Gagana. H. R; Dr. Suma V; Prof. Poornima D; Shobha N; Prof. Swathi B V; Nayana Shinde; and Prof. Leelavathi R.
The application for the patent was published on April 17, under issue no. 16/2026.
According to the abstract released by the Intellectual Property India: "A system and method for predicting, prior to execution, whether an autonomous AI agent task pipeline will terminate within bounded resources or enter a non-terminating state. The system constructs a semantic graph of the proposed task, performs forward simulation using a lightweight shadow agent, and applies three halting failure detectors cyclic semantic similarity, monotonic resource trajectory, and self-referential goal modification to the simulated graph. Upon detection of a non-termination signature, the system generates a bounded reformulation of the task incorporating explicit termination conditions, and presents this to the operator for confirmation. The shadow agent operates at approximately one percent of the primary agent's compute cost. The system is model-agnostic and integrates with any agent framework through the task specification interface."
Disclaimer: Curated by HT Syndication.
