US 11,991,050 B2
Drift detection in edge devices via multi-algorithmic deltas
Andrew C. M. Hicks, Highland, NY (US); and Michael Terrence Cohoon, Fishkill, NY (US)
Assigned to INTERNATIONAL BUSINESS MACHINES CORPORATION, Armonk, NY (US)
Filed by International Business Machines Corporation, Armonk, NY (US)
Filed on Oct. 25, 2022, as Appl. No. 18/049,362.
Prior Publication US 2024/0137286 A1, Apr. 25, 2024
Int. Cl. H04L 41/16 (2022.01); H04L 41/0631 (2022.01)
CPC H04L 41/16 (2013.01) [H04L 41/0631 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented system, comprising:
a memory that stores computer executable components; and
a processor that executes the computer executable components stored in the memory, wherein the computer executable components comprise:
a verification component that verifies an accuracy of a first model and an accuracy of a second model, wherein the second model is trained to have a lower accuracy than the accuracy of the first model;
a computation component that computes a first ratio based on verification of respective accuracies of the first model and the second model using test data, and computes a second ratio based on verification of the respective accuracies of the first model and the second model using real-time data generated by an edge device using one or more sensors; and
an analysis component that uses a difference between the first ratio and the second ratio to detect data drift associated with the edge device and to determine whether performance degradation of at least one of the first model or the second model is a function of the data drift, wherein the edge device is deployed without network connectivity.