MUMBAI, India, Sept. 14 -- Intellectual Property India has published a patent application (202641107343 A) filed by J. Hari Krishna; Maguluri V Lavanya; L L S Maneesha; A Anil Kumar Reddy; Muntimadugu Vijaya Kanth; Ms. Chakala Navya; Kommala Aparna; and Srirangam Anitha on September 07, 2026, for Ethical Data Science Pipeline Analysis For Bias Detection And Algorithmic Fairness Optimizations.
Inventors include J. Hari Krishna; Maguluri V Lavanya; L L S Maneesha; A Anil Kumar Reddy; Muntimadugu Vijaya Kanth; Ms. Chakala Navya; Kommala Aparna; and Srirangam Anitha.
The application for the patent was published on September 11, 2026, under issue no. 37/2026.
Abstract: The present work introduces an ethical data science pipeline that systematically analyzes datasets and models for bias while optimizing algorithmic fairness throughout the machine-learning lifecycle. Sensitive attributes and protected-group indicators are examined at every stage from data ingestion through feature engineering, model training, and post-deployment monitoring. Quantitative bias metrics are computed in parallel with fairness constraints, and an adaptive optimization layer continuously adjusts sampling weights, feature representations, or decision thresholds so that predictive performance remains high while group disparities are reduced. The pipeline maintains a transparent audit trail of fairness interventions, enabling reproducible ethical assessments. Experimental validation on representative classification and regression tasks demonstrates measurable reductions in disparate impact and equalized-odds gaps without substantial loss of overall accuracy, establishing a practical framework for responsible data-science practice.
Disclaimer: Curated by HT Syndication.
