Omer Uretzky

Title: Omer Uretzky: Inventor in Machine Learning and Virtualization

Introduction

Omer Uretzky is an inventor based in Herzliya, Israel. His patent work includes technologies for detecting poisoned machine-learning training data and securing data replication in virtualized environments. He is associated with Hewlett Packard Enterprise Development LP and is named on 7 patents.

Latest Patents

“Training Data Poisoning Detection” describes a system that evaluates the integrity of a machine-learning training data set. It measures changes to features across training samples, calculates an integrity score from those measurements, and uses the score to detect potential poisoning.

“Authentication for Network-Based Virtual Machine Replication” concerns secure data replication in virtualized environments. A data change filter installed in a hypervisor intercepts operations from a virtual machine and uses certificates managed by the hypervisor to establish an authenticated connection with a replication processing service. The intercepted changes are then sent over that connection.

Career Highlights

Uretzky’s listed inventions address data integrity and security in computing systems. His work on training-data poisoning detection focuses on identifying potentially compromised machine-learning inputs, while his virtualization patent describes certificate-based authentication for replication traffic.

Collaborations

Uretzky is associated with Hewlett Packard Enterprise Development LP, the company identified in connection with his patent work.

Conclusion

Omer Uretzky’s patents reflect an emphasis on protecting data and strengthening trust in machine-learning and virtualized systems.