MUMBAI, India, Sept. 22 -- Intellectual Property India has published a patent application (202641109007 A) filed by Mlr Institute Of Technology; and Marri Laxman Reddy Institute Of Technology And Management on September 10, 2026, for Autonomous Multi-Agent System For Natural Language Driven Data Analysis Using Local Large Language Models.
Inventors include Mrs. P. Swathi; Mr. Mittapalli Dileep; Mr. Munagala Sandeep Kumar; Mr. Kota Harsha; and Mr. E. Y. S. V. S Abhay.
The application for the patent was published on September 18, 2026, under issue no. 38/2026.
Abstract: The use of data analysis plays a vital role in the decision-making process, and for many users, conducting complex analyses often poses barriers due to their lack of programming or statistical expertise. Traditional tools typically require manual coding and domain-related knowledge, which can limit access and generate inefficiencies across all industries. Recent progress has been made with the introduction of Large Language Model (LLM) capabilities, facilitating greater automation of the data analysis workflow and providing users with additional means of interacting with data analysis in a user-friendly manner. Unfortunately, most current systems continue to limit advanced data analysis capabilities by providing little in the way of autonomy, flexibility and an adequate error correction methodology during the execution phase of the analysis. This paper introduces the Autonomous Multi-Agent System to address the limitations of most existing systems. The system automatically generates executable Python or SQL code based on a plain-language query, executes the code within a sandbox environment, verifies the outcome of the code execution and iteratively corrects any errors detected through automated fault tolerance mechanisms. The Autonomous Multi-Agent System architecture consists of several modules, including the Planner, which breaks the task into smaller components; the Code Synthesizer for generating programmatic code, Executor for executing the programmatic code with controlled execution; and Verifier for confirming the results generated by the comprehensive data analysis process. The combined functionality of intelligent code generation, runtime feedback, and iterative verification loops allows the system to support domain-independent, consistent, reliable, repeatable analytics. As a result, non-technical users can accurately conduct clear, transparent analyses using the system, freeing up the user from acting as the manual input of data into the analysis process. 5 Claims and 1 Figures
Disclaimer: Curated by HT Syndication.
