MUMBAI, India, Sept. 22 -- Intellectual Property India has published a patent application (202611110332 A) filed by Chandra Shekar Reddy Avula on September 14, 2026, for An Adaptive Distributed Data Processing System For Automated Data Quality Detection, Feature Reliability Scoring And Model-Ready Dataset Generation.
Inventor includes Chandra Shekar Reddy Avula.
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 automated data-quality detection, feature reliability scoring and generation of a model-ready dataset. Heterogeneous data are partitioned for distributed processing and evaluated using multiple context-aware quality signals. A dependency graph represents relationships among source attributes, transformations, derived features, partitions and dataset versions. A feature reliability scoring engine generates a reliability score and a separate confidence value using direct quality evidence, evidence coverage, temporal recency and propagated dependency reliability. A policy and drift controller distinguishes candidate quality degradation from contextual change, while a repair or quarantine orchestrator selects a corrective action according to expected reliability improvement and transformation risk. Following a defect-state or corrective-action change, only an affected dependency subgraph is incrementally re-evaluated. A model-readiness assembler releases a versioned dataset when feature-specific readiness conditions are satisfied and generates a manifest containing reliability, confidence, corrective- action and provenance information. The architecture reduces redundant distributed recomputation while improving traceability and consistency of model-input data. Accompanied Drawing [FIGS. 1-2].
Disclaimer: Curated by HT Syndication.
