MUMBAI, India, Aug. 10 -- Intellectual Property India has published a patent application (202641089570 A) filed by Dr Ravichandra Bandi; Dr P N E Naveen; Dr S Sambhu Prasad; Mr K Chaitanya; Dr D Sirisha; and Mr. E. Prabhakar on July 22, 2026, for A System And Method For Contextual Logical Error Detection And Auto-Correction In Source Code Using Intent-Inference Machine Learning.
Inventors include Dr Ravichandra Bandi; Dr P N E Naveen; Dr S Sambhu Prasad; Mr K Chaitanya; Dr D Sirisha; and Mr. E. Prabhakar.
The application for the patent was published on July 31, 2026, under issue no. 31/2026.
Abstract: A system and method that analyze programming logic for source code logical errors by comparing a programmer's intent (extracted from natural language artifacts using NLP) to the actual logic being implemented (represented as a language-agnostic semantic graph). A discrepancy is identified when the alignment score between the two falls below a predefined threshold. The discrepancy may represent a logic error such as an incorrect boundary condition or an inversion of operands. The system includes a Correction Candidate Engine that proposes one or more code modifications for fixing the discrepancy. The discrepancy detection and correction suggestion features can operate at the time of development as a real-time coding assistant or as a batch process for auditing legacy code. The system provides several advantages over standard syntax- and pattern-based linters/formatters, including enhanced discoverability of logical errors due to its ability to reason
Disclaimer: Curated by HT Syndication.
