Machine-learning for real-time and secure analysis of digital metrics
US12299093B1 · kind B1 · utility
Assignee
Inventors
Key dates
| Filing date | Aug 23, 2022 |
| Grant date | May 13, 2025 |
| Priority date | — |
| Expiry date | Apr 10, 2043 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG06F21/46
- WIPO fieldComputer technology
- WIPO sectorElectrical engineering
Abstract
Disclosed are example methods, systems, and devices that allow for executing machine-learning models for real-time and secure analysis of digital metrics. The techniques include generating metrics for identity elements stored in digital profiles of users. A subset of profiles can be identified that have metrics that fall below a predetermined thresholds, with which a training dataset can be generated. Machine-learning models can be executed over the training dataset to train an artificial intelligence agent that receives digital profiles as input and outputs translational elements corresponding to identity elements in the digital profiles. After training, additional profiles can be input to the machine-learning models of the artificial intelligence agent to identify a second subset of digital profiles with corresponding metrics. Electronic messages corresponding to the second subset can be generated and transmitted to one or more computing devices identified in the second subset of digital profiles.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.