MUMBAI, India, Aug. 24 -- Intellectual Property India has published a patent application (202621081475 A) filed by Symbiosis International Deemed University on July 01, 2026, for Ai-Powered Legal Document Summarization System Using Domain-Specific Transformer Model With Rule-Based Post-Processing Validation.
Inventors include Nidhi Mangtani; Vedant Rajekar; Yash Parate; and Pratik Agrawal.
The application for the patent was published on August 21, 2026, under issue no. 34/2026.
Abstract: ABSTRACT AI-POWERED LEGAL DOCUMENT SUMMARIZATION SYSTEM USING DOMAIN-SPECIFIC TRANSFORMER MODEL WITH RULE-BASED POST-PROCESSING VALIDATION The present invention provides an AI-powered legal document summarization system (100) comprising a user interface module (110) built using the Gradio framework, a text preprocessing module (120) for cleaning and tokenizing legal text, a domain-specific transformer summarization engine (130) utilizing the nsi319/legal-pegasus model integrated through the HuggingFace pipeline, a summary length configuration module (135) providing short, medium, and full summary formats, a post-processing validation module (140) that checks generated summaries for the presence of essential legal elements including appeal rights, sentencing details, and statutory sections and automatically appends standardized legal statements where such elements are missing, and an output module (150) for displaying the validated summary. The system (100) employs a hybrid approach combining AI-based abstractive summarization with rule-based validation to ensure legal completeness, accuracy, and reliability of the generated summaries. [
Disclaimer: Curated by HT Syndication.
