Maria Vlachopoulou

Title: Maria Vlachopoulou: Inventor in Machine Learning Anomaly Detection.

Introduction

Maria Vlachopoulou is an inventor based in Bellevue, Washington. Her patent work addresses machine-learning methods for improving anomaly detection by adjusting reconstruction losses according to feature importance.

Latest Patents

Vlachopoulou is associated with 1 patent, titled “Anomaly Detection Performance Enhancement Using Gradient-Based Feature Importance.” The invention describes techniques that adjust the reconstruction loss of a model, such as principal component analysis, based on the importance of individual features. During production inference, the system weights each feature’s reconstruction loss by its importance, then evaluates the combined weighted losses to determine whether an input is anomalous. This approach is designed to compensate for concept drift.

Career Highlights

Vlachopoulou is associated with Oracle International Corporation. Her patent reflects work at the intersection of machine learning, feature importance, and anomaly detection.

Collaborations

The patent information names Saeid Allahdadian and Yuting Sun as coworkers associated with Vlachopoulou’s work.

Conclusion

Maria Vlachopoulou’s patent describes a feature-weighted approach to reconstruction-based anomaly detection, with the goal of improving performance as data patterns change.