MUMBAI, India, Feb. 6 -- Intellectual Property India has published a patent application (202641003326 A) filed by Justyukti Services Llp, Bengaluru, Karnataka, on Jan. 12, for 'system and method for constraint-controlled probabilistic ai inference in regulated-domain decisioning.'
Inventor(s) include Pooja Kadambi; and Divya Ajitsaria.
The application for the patent was published on Feb. 6, under issue no. 06/2026.
According to the abstract released by the Intellectual Property India: "A computer-implemented system 100 is disclosed for controlling probabilistic artificial-intelligence inference in regulated-domain decisioning. The system 100 enforces hard regulatory constraints using a deterministic rule-enforcement module 136 to derive a constraint-defined decision space prior to any probabilistic inference. Execution of a probabilistic inference module 140 is conditionally gated such that outputs violating the constraints are prevented. Regulatory text segments relevant to the constraint-defined decision space are retrieved prior to generation, each segment being version-identified. An output decision is generated solely from the retrieved regulatory text segments and is blocked when no constraint-compliant segment is available. The system 100 further attaches an auditable evidence bundle to the output decision, including regulatory clause references and corresponding version identifiers, thereby enabling traceable, reproducible, and compliant AI-assisted decisioning in regulated environments."
Disclaimer: Curated by HT Syndication.
