MUMBAI, India, Sept. 14 -- Intellectual Property India has published a patent application (202611088054 A) filed by Rupesh Kumar on July 19, 2026, for Biochem_evaluatorllm™: An Explainable Ai-Based Rubric-Integrated Automated Assessment System For Descriptive Answer Evaluation.
Inventors include Dr. Samir Kumar Singh; and Dr Sairoz.
The application for the patent was published on September 11, 2026, under issue no. 37/2026.
Abstract: The present invention relates to a computer-implemented artificial intelligence (AI) system and method for automated evaluation of descriptive answers using large language models (LLMs). The invention provides an explainable, rubric-integrated assessment framework that combines examiner-defined marking rubrics, model answers, structured prompt engineering, evidence extraction, criterion-wise scoring, multi-pass inference, statistical score aggregation, and personalized feedback generation within a unified platform. Student responses are processed by an LLM configured with predefined evaluation criteria to identify relevant content corresponding to individual rubric components, assign partial or complete marks, generate transparent justifications for awarded scores, and provide actionable learning recommendations. To improve reliability and reduce inference variability, the system performs multiple independent evaluations under controlled stochastic parameters and computes a consensus score using statistical aggregation. The platform further incorporates confidence estimation, audit trails, configurable rubrics, retrieval- augmented knowledge integration, learning analytics, and competency-based educational assessment. Unlike conventional automated grading systems that rely primarily on keyword matching, semantic similarity, or domain-specific model training, the disclosed invention enables explainable, transparent, scalable, and domain- adaptable assessment without retraining the underlying language model. The invention substantially reduces evaluator workload while maintaining expert-level scoring consistency and providing detailed formative feedback suitable for medical, scientific, engineering, and higher education assessments. The disclosed system is applicable to on-premise, cloud-based, and learning management system (LMS) environments for large-scale automated educational evaluation.
Disclaimer: Curated by HT Syndication.
