Machine-learning based record processing systems

ChangeBridge: Patent Grants - AI & Computing (G06N)

Published April 7th, 2026

Detected April 7th, 2026

Summary

The USPTO granted Patent US12596963B2 to Capital One Services, LLC covering machine-learning based systems for automatically matching unmatched database records. The patent describes computing devices that use ML models to determine distance values between source and target records and generate matched records to update databases. The patent names Gunther Havel, John Wilson, and Ashwin Assysh Sharma as inventors.

View original document View source feed page

What changed

USPTO granted US Patent 12,596,963 B2 to Capital One Services, LLC on April 7, 2026. The patent covers machine learning techniques for matching source records with target records in databases, using distance values calculated by ML models to identify matching subsets and generate updated database records. The patent contains 20 claims and falls under CPC classifications G06N 20/00, G06N 20/20, G06Q 20/389, G06F 18/2113, and G06F 18/22.

For parties operating in financial services technology or database management, this patent grants Capital One exclusive rights to ML-based database record matching technology in the specified classifications. Competitors developing similar ML systems for financial database matching should review their products for potential infringement risks and consider licensing discussions or design-around strategies.

What to do next

Monitor for updates

Source document (simplified)

← USPTO Patent Grants

Machine-learning based record processing systems

Grant US12596963B2 Kind: B2 Apr 07, 2026

Assignee

Capital One Services, LLC

Inventors

Gunther Havel, John Wilson, Ashwin Assysh Sharma

Abstract

Aspects described herein may allow unmatched records in databases be matched automatically. For example, a computing device may receive source records and target records to be matched together. The computing device may determine, based on a machine learning model and for each of the plurality of target records, a distance value between the respective target record, and a subset of the plurality of source records. A matched record that identifies the subset of the plurality of source records and a selected one or more target records may be generated. The matched record may be configured to update records in a database.