MUMBAI, India, Aug. 17 -- Intellectual Property India has published a patent application (202641095077 A) filed by Cmr Institute Of Technology; Cmr Technical Campus; and Cmr College Of Engineering & Technology on August 06, 2026, for A System And Method For Bias-Aware Data Imputation Using Causal Inference In Sparse Datasets.
Inventors include Divya Lingisetti; Borusu Ramani; Y Ashok Kumar; Dr Harikrishna P; Mr. H. Swaraj Bharath; and Ms. R. Dhana Lakshmi.
The application for the patent was published on August 14, 2026, under issue no. 33/2026.
Abstract: A system for bias-aware data imputation using causal inference in sparse datasets is disclosed. The system receives datasets containing missing values and preprocesses them to identify sparsity patterns and missingness types. A causal structure learning module constructs a causal graph representing relationships among variables, enabling identification of confounding factors and bias sources. A bias detection module quantifies distributional and fairness-related biases, while a counterfactual inference module estimates missing values using structural causal models. An imputation module generates bias-mitigated data by leveraging counterfactual estimations to reduce propagation of existing biases. A validation and refinement module evaluates the imputed dataset using statistical consistency and fairness metrics and iteratively improves performance. The disclosed system enhances accuracy and fairness of imputation in sparse datasets, making it suitable for applications involving sensitive or incomplete data.
Disclaimer: Curated by HT Syndication.
