Amirhossein Yazdavar

Title: Amirhossein Yazdavar: Inventor in Machine Learning and Personalized Care

Introduction

Amirhossein Yazdavar is an inventor based in Scottsdale, Arizona. His work includes machine-learning methods for evaluating graph data and systems that use patient information to generate personalized care paths. He is named on 2 patents.

Latest Patents

Yazdavar’s latest patents address applications of graph-based machine learning in data evaluation and healthcare.

“Graph-Based Machine Learning Model with Feature-Agnostic Training and Holistic Feature Metrics” describes techniques for evaluating a knowledge graph. A target graph model generates a predictive representation for a graph node, and a feature prediction model uses that representation to estimate feature values. Those predictions support a data evaluation score, which is combined with the model’s predictive output to produce an evaluation output.

“Systems and Methods for Generating Personalized Care Paths for Patients” describes a method for determining an individual care path from demographic and medical data gathered from multiple sources. A graph convolutional neural network-based model is trained using patient-bucket-procedure graphs representing care paths, and information about the resulting path is provided to a user’s device.

Career Highlights

Yazdavar has worked with Optum, Inc. and Optum Services (Ireland) Limited. His listed inventions reflect work at the intersection of data science, graph-based modeling, and healthcare applications.

Collaborations

Brian Carter is identified as a coworker. The available information does not specify the nature of their collaboration.

Conclusion

Amirhossein Yazdavar’s two listed patents apply graph-based machine learning to knowledge-graph evaluation and personalized patient care, connecting computational methods with practical data-driven systems.