This inventor holds 1 USPTO granted patent. Top assignee: Hewlett Packard Enterprise Development LP. Active years: 2026.
Data source: USPTO (United States Patent and Trademark Office) public patent records. Weekly synchronization. How IDiyas builds this profile
Company Filing History:
Years Active: 2026
Title: Amir Idar: Inventor in Machine Learning Data Integrity
Introduction
Amir Idar is an inventor based in Herzliya, Israel. His patent work addresses the integrity of training data used in machine learning models.
Latest Patents
Idar is associated with 1 patent, titled “Training Data Poisoning Detection.” The invention describes a system that evaluates changes to features in training samples, calculates an integrity score for a training data set, and uses that score to detect potential data poisoning.
Career Highlights
Idar is associated with Hewlett Packard Enterprise Development LP. His patent reflects work on identifying compromised training data, an important concern for the reliability of machine learning systems.
Collaborations
Idar’s listed coworker is Omer Uretzky.
Conclusion
Amir Idar’s patent focuses on detecting training data poisoning through feature-change analysis and integrity scoring, contributing to efforts to improve the trustworthiness of machine learning data.
Data source: USPTO (United States Patent and Trademark Office) public patent records. Weekly synchronization. How IDiyas builds this profile