17994530. UNIFY95: META-LEARNING CONTAMINATION THRESHOLDS FROM UNIFIED ANOMALY SCORES simplified abstract (Oracle International Corporation)

UNIFY95: META-LEARNING CONTAMINATION THRESHOLDS FROM UNIFIED ANOMALY SCORES

Organization Name

Oracle International Corporation

Inventor(s)

Yasha Pushak of Vancouver (CA)

Hesam Fathi Moghadam of Sunnyvale CA (US)

Anatoly Yakovlev of Hayward CA (US)

Robert David Hopkins, Ii of Foster City CA (US)

UNIFY95: META-LEARNING CONTAMINATION THRESHOLDS FROM UNIFIED ANOMALY SCORES - A simplified explanation of the abstract

This abstract first appeared for US patent application 17994530 titled 'UNIFY95: META-LEARNING CONTAMINATION THRESHOLDS FROM UNIFIED ANOMALY SCORES

Simplified Explanation

The patent application describes a method for setting a universal anomaly threshold based on anomaly scores from multiple anomaly detection algorithms and labeled datasets. This threshold is used to detect anomalies in data.

• Anomaly detection algorithms are trained on labeled datasets to generate anomaly scores for data points.
• Anomalies are identified based on the anomaly scores exceeding the calculated anomaly threshold.
• The universal anomaly threshold is determined as the average of individual anomaly thresholds from different anomaly detection algorithms.

Potential Applications

This technology can be applied in various fields such as cybersecurity, fraud detection, and quality control in manufacturing.

Problems Solved

This technology addresses the challenge of setting a consistent anomaly threshold across different anomaly detection algorithms and datasets, improving the accuracy of anomaly detection.

Benefits

- Improved anomaly detection accuracy - Consistent anomaly threshold across different algorithms and datasets - Enhanced data security and fraud prevention

Potential Commercial Applications

The technology can be utilized in industries such as finance, healthcare, and e-commerce for real-time anomaly detection and prevention.

Possible Prior Art

One possible prior art could be the use of ensemble methods in anomaly detection to combine multiple anomaly detection algorithms for improved performance.

Unanswered Questions

How does this technology handle imbalanced datasets in anomaly detection?

The article does not provide information on how the universal anomaly threshold is adjusted for imbalanced datasets where anomalies are rare.

What computational resources are required to implement this technology at scale?

The article does not mention the computational resources needed to train multiple anomaly detection algorithms and calculate the universal anomaly threshold.

Original Abstract Submitted

Herein is a universal anomaly threshold based on several labeled datasets and transformation of anomaly scores from one or more anomaly detectors. In an embodiment, a computer meta-learns from each anomaly detection algorithm and each labeled dataset as follows. A respective anomaly detector based on the anomaly detection algorithm is trained based on the dataset. The anomaly detector infers respective anomaly scores for tuples in the dataset. The following are ensured in the anomaly scores from the anomaly detector: i) regularity that an anomaly score of zero cannot indicate an anomaly and ii) normality that an inclusive range of zero to one contains the anomaly scores from the anomaly detector. A respective anomaly threshold is calculated for the anomaly scores from the anomaly detector. After all meta-learning, a universal anomaly threshold is calculated as an average of the anomaly thresholds. An anomaly is detected based on the universal anomaly threshold.