MUMBAI, India, Sept. 22 -- Intellectual Property India has published a patent application (202611110418 A) filed by Ravikumar Chennupati on September 15, 2026, for An Adaptive Distributed Data Processing System With Context-Aware Data Quality Detection, Feature Reliability Scoring And Model-Ready Dataset Generation.
Inventor includes Ravikumar Chennupati.
The application for the patent was published on September 18, 2026, under issue no. 38/2026.
Abstract: An adaptive distributed data processing system is disclosed for generating model-ready datasets from heterogeneous and changing data sources. Distributed processing nodes profile respective data partitions and generate normalized quality-event descriptors carrying contextual and lineage information. A global quality-state engine aggregates local evidence to distinguish partition-specific conditions from source-wide conditions. A feature reliability engine determines feature reliability states from source integrity, completeness, distribution stability, semantic conformity, cross-partition consistency, lineage confidence and remediation burden, with context- sensitive weighting according to a target model contract. An adaptive processing-plan generator selectively constructs a directed transformation graph according to reliability deficits, expected reliability improvement, processing cost and semantic risk. Post-transformation profiling verifies whether selected operations improved reliability and can trigger alternative processing, quarantine or exclusion. A model-ready dataset assembler admits, weights or excludes features in accordance with verified reliability and model-contract requirements, while a provenance store generates a reproducible manifest containing source, quality, transformation and reliability evidence. The system thereby controls distributed processing using an evolving reliability state rather than a fixed preprocessing sequence. Accompanied Drawing [FIGS. 1-2].
Disclaimer: Curated by HT Syndication.
