USPTO Patent Grant: Anomaly Detection Thresholds and Explanations

ChangeBridge: Patent Grants - AI & Computing (G06N)

Published March 24th, 2026

Detected March 24th, 2026

Summary

The USPTO has granted a patent to Dell Products L.P. for a framework related to anomaly detection models. This framework aims to provide quantitative explanations for identified anomalies by mapping feature-outlier scores to feature-value ranges.

View original document View source feed page

What changed

The United States Patent and Trademark Office (USPTO) has issued patent US12585532B2 to Dell Products L.P. for a "Framework for selecting thresholds for anomaly detection models and generating quantitative explanations." The patent covers a method involving computing outlier scores for a dataset, extracting feature-specific thresholds to identify abnormal entries, and generating feature-value ranges as explanations for the detected anomalies. This framework is intended to provide quantitative metrics to explain the behavior of anomaly detection models.

This patent grant is primarily of interest to technology companies and researchers in the field of artificial intelligence and machine learning. While patents do not impose direct regulatory obligations on other entities, they establish intellectual property rights that can influence product development and market competition. Compliance officers should be aware of this patent as it pertains to technologies used in fraud detection, cybersecurity, and data integrity within their organizations or those of their vendors.

Source document (simplified)

← USPTO Patent Grants

Framework for selecting thresholds for anomaly detection models and generating quantitative explanations

Grant US12585532B2 Kind: B2 Mar 24, 2026

Assignee

Dell Products L.P.

Inventors

Adriana Bechara Prado, Alexander Eulalio Robles Robles, Eduarda Tatiane Caetano Chagas, Helen Cristina de Mattos Senefonte, Jonathan Mendes De Almeida, Karen Stéfany Martins

Abstract

A framework for extracting quantitative and comparable explanations in terms of feature-value ranges for anomaly detection models based on outlier scores is disclosed. In a first phase, outlier scores are computed for a data set. In a second phase, thresholds per feature, which are used to identify abnormal entries or records in the data set, are extracted. In a third phase, a map between feature-outlier scores and feature-value ranges is generated. The feature-value ranges represent explanations. The explanations may include extracting quantitative metrics to explain a model.

CPC Classifications

G06F 11/0793 G06F 11/079 G06F 11/0772 G06F 11/0709 G06F 11/0751 G06F 18/15 G06F 16/2246 G06F 16/2465 G06F 18/22 G06F 17/16 G06F 11/3409 G06F 11/08 G06F 18/217 G06F 18/214 G06F 11/3452 G06F 11/3476 G06F 18/251 G06F 17/18 G06F 18/2113 G06F 16/1824 G06F 16/13 G06F 40/169 G06F 40/40 G06F 40/30 G06F 40/35 H04L 63/1425 H04L 63/1416 H04L 41/069 H04L 41/145 H04L 43/50 H04L 41/065 H04L 43/08 H04W 4/70 G06N 20/10 G06N 20/00