Machine learning systems and methods for real time anomaly detection and prescriptive feedback
A real-time trained machine learning model addresses the limitations of existing anomaly detection systems by enabling efficient detection of unknown anomalies and accurate analysis of multi-step, multi-source queries, enhancing computing device functionality.
Patent Information
- Authority / Receiving Office
- GB · GB
- Patent Type
- Applications
- Current Assignee / Owner
- ZEBRA TECHNOLOGIES CORP
- Filing Date
- 2025-09-18
- Publication Date
- 2026-04-22
AI Technical Summary
Existing anomaly detection processes lack the ability to perform real-time, on-demand detection of anomalies, fail to convert anomaly detection into a persistent process, and are unable to handle multi-step, multi-source queries, leading to inefficiencies and missed detection of previously unknown anomalies.
A machine learning model trained in real-time is used to analyze multi-step, multi-source queries, generating sets of instructions for identifying anomalous items and executing them in a computing device, ensuring proper query linking and validation for enhanced efficiency and accuracy.
The system efficiently detects previously unknown anomalies in real-time and accurately analyzes multiple linked queries, improving the functionality of computing devices by handling multi-step, multi-source queries with reduced code and increased accuracy.
Smart Images

Figure 00000001_0000 
Figure 00000002_0000 
Figure 00000003_0000
Abstract
Citation Information
Patent Citations
A New Retail QR Code Scanning Data Anomaly Detection Method Based on Cloud POS Machine
CN109408268B
Machine learning-based techniques for detecting payroll fraud
US20210004915A1
Generating models for detection of anomalous patterns
US20240098106A1
Systems and methods of controlling retail product allocation and retail market variations based on customized insight
WO2023219949A1