MUMBAI, India, Nov. 2 -- Intellectual Property India has published a patent application (202641125898 A) filed by Koneru Lakshmaiah Education Foundation on October 22, 2026, for Synthlend: Missingness-Aware Behavioral Cadence Fingerprinting For Thin-File Credit Risk..
Inventors include Poli Sai Sreyash Reddy; Leo Stephen Maduri; Ponduri Lakshmi; Majety Charan Sai Sujith; Jampani Abhishek; and Anjana Devi Nandam.
The application for the patent was published on October 30, 2026, under issue no. 44/2026.
Abstract: The present invention provides an integrated computer-implemented system for missingness-aware behavioral fingerprinting and credit-risk assessment for applicants having limited recorded credit histories. The system receives linked applicant profiles and repayment records from one or more financial reporting modalities and processes the data through a computational architecture comprising chronological anomaly correction, independent missingness and repayment-state feature extraction, fusion-based representation learning, and independent reconstruction-based coherence evaluation. The extracted missingness indicators and behavioral cadence features are progressively combined through a specialized fusion autoencoder to generate a compressed behavioral fingerprint. A deterministic correction mechanism adjusts inconsistent installment due-date ordering by shifting corresponding paid dates, rigorously preserving original payment lateness. A distinct coherence-assessment module evaluates latent structural integrity by measuring reconstruction error, generating a coherence score independent of an estimated default probability. The system further incorporates validation-calibrated thresholds to determine applicant-routing pathways, systematically assigning standard, caution-designated, or deferred scoring based on data structural reliability. The risk-estimation architecture maintains a strict separation between representation coherence and direct creditworthiness. The invention thereby provides an integrated computational framework for robust, equitable, and structurally consistent thin-file credit-risk evaluation.
Disclaimer: Curated by HT Syndication.
